From 7fd939c06bfc50fbe6b3316070ffe79e73ca2cee Mon Sep 17 00:00:00 2001 From: Andrew Flowers Date: Wed, 4 Nov 2015 17:07:42 -0500 Subject: [PATCH] add study drugs nb --- .../pierson_study_drugs-checkpoint.ipynb | 1409 ++++++++++++++++ study-drugs/README.md | 13 + study-drugs/pierson_study_drugs.ipynb | 1416 +++++++++++++++++ 3 files changed, 2838 insertions(+) create mode 100644 study-drugs/.ipynb_checkpoints/pierson_study_drugs-checkpoint.ipynb create mode 100644 study-drugs/README.md create mode 100644 study-drugs/pierson_study_drugs.ipynb diff --git a/study-drugs/.ipynb_checkpoints/pierson_study_drugs-checkpoint.ipynb b/study-drugs/.ipynb_checkpoints/pierson_study_drugs-checkpoint.ipynb new file mode 100644 index 0000000..98ea371 --- /dev/null +++ b/study-drugs/.ipynb_checkpoints/pierson_study_drugs-checkpoint.ipynb @@ -0,0 +1,1409 @@ +{ + "cells": [ + { + "cell_type": "code", + "execution_count": 18, + "metadata": { + "collapsed": false + }, + "outputs": [], + "source": [ + "%matplotlib inline " + ] + }, + { + "cell_type": "code", + "execution_count": 2, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "data": { + "text/html": [ + " " + ], + "text/plain": [ + "" + ] + }, + "execution_count": 2, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "from IPython.core.display import HTML\n", + "def css_styling():\n", + " styles = \" \"\n", + " return HTML(styles)\n", + "css_styling()\n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "# First load data" + ] + }, + { + "cell_type": "code", + "execution_count": 3, + "metadata": { + "collapsed": false + }, + "outputs": [], + "source": [ + "from pylab import *\n", + "import pandas as pd\n", + "import numpy as np\n", + "import os\n", + "import statsmodels.api as sm\n", + "from statsmodels.iolib.summary2 import summary_col\n", + "\n", + "#load in NSDUH data\n", + "nsduh_data = pd.read_csv('ICPSR_35509/DS0001/35509-0001-Data.tsv', sep = '\\t')\n", + "#load in Niche data\n", + "niche_data = pd.read_csv('niche/SchoolPoll.csv')\n", + "#load in IPEDS data\n", + "def getIPEDSData():\n", + " files = ['ipeds_data/CSV_7312015-162.csv', 'ipeds_data/CSV_7312015-826.csv']\n", + " \n", + " for i, f in enumerate(files):\n", + " d = pd.read_csv(f)\n", + " d.index = d.unitid \n", + " del d['unitid']\n", + " if i == 0:\n", + " all_d = d\n", + " else:\n", + " all_d = pd.concat([all_d, d])\n", + " ipeds_d = {}\n", + " for c in all_d.columns:\n", + " new_name = c.replace(' - ', '_').replace('/', '_').replace(',', '_').replace('.', '_').replace(' ', '_')\n", + " new_name = new_name.replace(\"Percent_of_total_enrollment_that_are_\", '')\n", + " ipeds_d[new_name] = dict(zip(all_d.index, all_d[c]))\n", + " return ipeds_d\n", + "ipeds_data = getIPEDSData()\n", + "#load in Google data\n", + "google_filenames = ['ADHD_Indexed.csv','Adderall_Indexed.csv']\n", + "google_data = {}\n", + "for filename in google_filenames:\n", + " d = pd.read_csv(os.path.join('google_data', filename))\n", + " d.index = d.Month\n", + " del d['Month']\n", + " d = d.transpose()\n", + " google_data[filename.replace('_Indexed.csv', '')] = d\n", + " \n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "###### \"When I looked at more recent data (the 2013 National Survey on Drug Use and Health (NSDUH), an annual government survey that includes more than 55,000 Americans) the difference turned out to be closer to 1.3x, not 2x.\"" + ] + }, + { + "cell_type": "code", + "execution_count": 5, + "metadata": { + "collapsed": false + }, + "outputs": [], + "source": [ + "#People who respond 1 have used Adderall non-medically; no one else has. \n", + "nsduh_data['ADDERALL'] = (nsduh_data['ADDERALL'] == 1)*1.\n", + "nsduh_data['RITALIN'] = (nsduh_data['RITMPHEN'] == 1)*1." + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "COLLENR levels:\n", + "1: FT college students age 18 - 22; 2: other age 18 - 22; 3: other\n", + "Category COLLENR\n", + "Mean value of ADDERALL for level 1 is 0.149; unweighted, 0.140 (4310 values)\n", + "Mean value of ADDERALL for level 2 is 0.115; unweighted, 0.114 (6934 values)\n", + "Mean value of ADDERALL for level 3 is 0.034; unweighted, 0.048 (43914 values)\n", + "Ratio between maximum value and minimum value 4.45754687876\n", + "Category COLLENR\n", + "Mean value of RITALIN for level 1 is 0.048; unweighted, 0.048 (4310 values)\n", + "Mean value of RITALIN for level 2 is 0.050; unweighted, 0.049 (6934 values)\n", + "Mean value of RITALIN for level 3 is 0.022; unweighted, 0.026 (43914 values)\n", + "Ratio between maximum value and minimum value 2.2055874138\n" + ] + } + ], + "source": [ + "\n", + "def stratifyByCat(nsduh_data, cat, drug_to_measure = 'ADDERALL', sortByVal = False):\n", + " \"\"\"\n", + " Checked. Computes the weighted and unweighted mean values of Adderall, stratifying by \n", + " the levels in cat. \n", + " \"\"\"\n", + " levels = sorted(list(set(nsduh_data[cat].dropna())))\n", + " print 'Category', cat\n", + " table = []\n", + " for l in levels:\n", + " idxs = nsduh_data[cat] == l\n", + " summed_weights = nsduh_data.loc[idxs]['ANALWT_C'].sum()\n", + " mu = (nsduh_data.loc[idxs][drug_to_measure] * nsduh_data.loc[idxs]['ANALWT_C']).sum() / summed_weights\n", + " unweighted_mu = nsduh_data.loc[idxs][drug_to_measure].mean()\n", + " if idxs.sum() > 25:\n", + " table.append([drug_to_measure, l, mu, unweighted_mu, idxs.sum()])\n", + " if sortByVal:\n", + " table = sorted(table, key = lambda x:x[2])[::-1]\n", + " for row in table:\n", + " print 'Mean value of %s for level %s is %2.3f; unweighted, %2.3f (%i values)' % (tuple(row))\n", + " print 'Ratio between maximum value and minimum value', max([a[2] for a in table]) / min([a[2] for a in table])\n", + "print 'COLLENR levels:\\n1: FT college students age 18 - 22; 2: other age 18 - 22; 3: other'\n", + "stratifyByCat(nsduh_data, 'COLLENR') \n", + "stratifyByCat(nsduh_data, 'COLLENR', drug_to_measure = 'RITALIN') \n", + "\n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "###### \"This is far smaller than the difference between white 18 - 22 year olds and black 18 - 22 year olds (6x, 18% vs 3%) or the difference between 18 - 22 year olds whose families do not receive food stamps and those whose do (1.6x, 14% vs 9%).\"" + ] + }, + { + "cell_type": "code", + "execution_count": 7, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Categories for race:\n", + "1: white, 2: black, 3: Native Am; 4: Native HI/Other Pac Isl; 5: Asian; 6: Multiracial; 7: Hispanic\n", + "Category NEWRACE2\n", + "Mean value of ADDERALL for level 1 is 0.181; unweighted, 0.171 (6194 values)\n", + "Mean value of ADDERALL for level 2 is 0.031; unweighted, 0.034 (1602 values)\n", + "Mean value of ADDERALL for level 3 is 0.052; unweighted, 0.086 (163 values)\n", + "Mean value of ADDERALL for level 4 is 0.084; unweighted, 0.078 (64 values)\n", + "Mean value of ADDERALL for level 5 is 0.084; unweighted, 0.086 (532 values)\n", + "Mean value of ADDERALL for level 6 is 0.113; unweighted, 0.155 (491 values)\n", + "Mean value of ADDERALL for level 7 is 0.076; unweighted, 0.065 (2198 values)\n", + "Ratio between maximum value and minimum value 5.84203431864\n", + "Category NEWRACE2\n", + "Mean value of RITALIN for level 1 is 0.071; unweighted, 0.069 (6194 values)\n", + "Mean value of RITALIN for level 2 is 0.007; unweighted, 0.009 (1602 values)\n", + "Mean value of RITALIN for level 3 is 0.026; unweighted, 0.037 (163 values)\n", + "Mean value of RITALIN for level 4 is 0.013; unweighted, 0.016 (64 values)\n", + "Mean value of RITALIN for level 5 is 0.029; unweighted, 0.028 (532 values)\n", + "Mean value of RITALIN for level 6 is 0.054; unweighted, 0.067 (491 values)\n", + "Mean value of RITALIN for level 7 is 0.026; unweighted, 0.022 (2198 values)\n", + "Ratio between maximum value and minimum value 9.9793813036\n", + "Categories for food stamp: 1: respondent or family member received food stamp; 2: did not\n", + "Category IRFSTAMP\n", + "Mean value of ADDERALL for level 1 is 0.088; unweighted, 0.089 (2564 values)\n", + "Mean value of ADDERALL for level 2 is 0.140; unweighted, 0.135 (8680 values)\n", + "Ratio between maximum value and minimum value 1.59577871899\n", + "Category IRFSTAMP\n", + "Mean value of RITALIN for level 1 is 0.041; unweighted, 0.043 (2564 values)\n", + "Mean value of RITALIN for level 2 is 0.051; unweighted, 0.050 (8680 values)\n", + "Ratio between maximum value and minimum value 1.24093935037\n" + ] + } + ], + "source": [ + "young_people_idxs = nsduh_data['COLLENR'].map(lambda x:x in [1, 2])\n", + "print '\\nCategories for race:\\n1: white, 2: black, 3: Native Am; 4: Native HI/Other Pac Isl; 5: Asian; 6: Multiracial; 7: Hispanic'\n", + "stratifyByCat(nsduh_data.loc[young_people_idxs], 'NEWRACE2') \n", + "stratifyByCat(nsduh_data.loc[young_people_idxs], 'NEWRACE2', drug_to_measure = 'RITALIN')\n", + "\n", + "print 'Categories for food stamp: 1: respondent or family member received food stamp; 2: did not'\n", + "stratifyByCat(nsduh_data.loc[young_people_idxs], 'IRFSTAMP') \n", + "stratifyByCat(nsduh_data.loc[young_people_idxs], 'IRFSTAMP', drug_to_measure = 'RITALIN') \n", + " \n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### Graphs on Adderall usage by age" + ] + }, + { + "cell_type": "code", + "execution_count": 8, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Category COLLENR\n", + "Mean value of ADDERALL for level 1 is 0.149; unweighted, 0.140 (4310 values)\n", + "Mean value of ADDERALL for level 2 is 0.115; unweighted, 0.114 (6934 values)\n", + "Mean value of ADDERALL for level 3 is 0.034; unweighted, 0.048 (43914 values)\n", + "Ratio between maximum value and minimum value 4.45754687876\n", + "Category age_cat_to_plot\n", + "Mean value of ADDERALL for level 12 - 14 is 0.004; unweighted, 0.005 (8689 values)\n", + "Mean value of ADDERALL for level 15 - 17 is 0.043; unweighted, 0.043 (9047 values)\n", + "Mean value of ADDERALL for level 23 - 25 is 0.145; unweighted, 0.140 (6896 values)\n", + "Mean value of ADDERALL for level 26 - 34 is 0.102; unweighted, 0.089 (5446 values)\n", + "Mean value of ADDERALL for level 35+ is 0.011; unweighted, 0.015 (13836 values)\n", + "Ratio between maximum value and minimum value 33.1820207109\n", + "Age\tPercent Using Adderall\n", + "12-14\t0.4\n", + "15-17\t4.3\n", + "18-22 (Not College)\t11.5\n", + "18-22 (In College)\t14.9\n", + "23-25\t14.5\n", + "26-34\t10.2\n", + "35+\t1.1\n" + ] + } + ], + "source": [ + "def remapAges(x):\n", + " if x <= 3:\n", + " return '12 - 14'\n", + " elif x <= 6:\n", + " return '15 - 17'\n", + " elif x <= 10:\n", + " return 'college age'\n", + " elif x <= 12:\n", + " return '23 - 25'\n", + " elif x <= 14:\n", + " return '26 - 34'\n", + " return '35+'\n", + "nsduh_data['age_cat_to_plot'] = nsduh_data['AGE2'].map(remapAges)\n", + "nsduh_data['age_cat_to_plot'].loc[nsduh_data['COLLENR'] != 3] = np.nan\n", + "\n", + "#This computes the values for Adderall usage stratified by age and college enrollment. \n", + "stratifyByCat(nsduh_data, 'COLLENR') \n", + "stratifyByCat(nsduh_data, 'age_cat_to_plot') \n", + "\n", + "#This prints out the actual data in an easy form for making graphs. Kind of hacky. \n", + "col_vals = [0.4, 4.3, 11.5, 14.9, 14.5, 10.2, 1.1]\n", + "col_labels = ['12-14', '15-17', '18-22 (Not College)', '18-22 (In College)', '23-25', '26-34', '35+']\n", + "print 'Age\\tPercent Using Adderall'\n", + "for i in range(len(col_vals)):\n", + " print '%s\\t%2.1f' % (col_labels[i], col_vals[i])" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "###### \"Study drugs were most frequently used in New England schools\"" + ] + }, + { + "cell_type": "code", + "execution_count": 9, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + " OLS Regression Results \n", + "==============================================================================\n", + "Dep. Variable: PctOfTotal R-squared: 0.126\n", + "Model: OLS Adj. R-squared: 0.110\n", + "Method: Least Squares F-statistic: 7.700\n", + "Date: Tue, 03 Nov 2015 Prob (F-statistic): 1.18e-09\n", + "Time: 14:09:22 Log-Likelihood: 350.49\n", + "No. Observations: 436 AIC: -683.0\n", + "Df Residuals: 427 BIC: -646.3\n", + "Df Model: 8 \n", + "Covariance Type: nonrobust \n", + "=============================================================================================================================================\n", + " coef std err t P>|t| [95.0% Conf. Int.]\n", + "---------------------------------------------------------------------------------------------------------------------------------------------\n", + "Intercept 0.2824 0.016 17.881 0.000 0.251 0.313\n", + "HD2013_Geographic_region[T.Great Lakes IL IN MI OH WI] 0.0604 0.021 2.901 0.004 0.019 0.101\n", + "HD2013_Geographic_region[T.Mid East DE DC MD NJ NY PA] 0.1021 0.019 5.330 0.000 0.064 0.140\n", + "HD2013_Geographic_region[T.New England CT ME MA NH RI VT] 0.1152 0.022 5.283 0.000 0.072 0.158\n", + "HD2013_Geographic_region[T.Outlying areas AS FM GU MH MP PR PW VI] 0.1047 0.111 0.946 0.344 -0.113 0.322\n", + "HD2013_Geographic_region[T.Plains IA KS MN MO NE ND SD] 0.0775 0.027 2.871 0.004 0.024 0.131\n", + "HD2013_Geographic_region[T.Rocky Mountains CO ID MT UT WY] -0.0432 0.030 -1.457 0.146 -0.102 0.015\n", + "HD2013_Geographic_region[T.Southeast AL AR FL GA KY LA MS NC SC TN VA WV] 0.0891 0.019 4.579 0.000 0.051 0.127\n", + "HD2013_Geographic_region[T.Southwest AZ NM OK TX] 0.0558 0.025 2.191 0.029 0.006 0.106\n", + "==============================================================================\n", + "Omnibus: 10.358 Durbin-Watson: 1.920\n", + "Prob(Omnibus): 0.006 Jarque-Bera (JB): 10.752\n", + "Skew: 0.384 Prob(JB): 0.00463\n", + "Kurtosis: 2.941 Cond. No. 22.7\n", + "==============================================================================\n", + "\n", + "Warnings:\n", + "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n" + ] + }, + { + "data": { + "text/plain": [ + "" + ] + }, + "execution_count": 9, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAEWCAYAAACAOivfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBwEzsxL7/0ZGrp7wszgNAAAAAElFTkSuQmCC\n", + "text/plain": [ + "" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "#Join Niche data with IPEDS data. \n", + "niche_adderall_idxs = (niche_data['Response'] == 'Prescription \"study drugs\" (Adderall, Ritalin) ')\n", + "niche_adderall_fracs = niche_data.loc[niche_adderall_idxs]\n", + "at_least_ten_responses = niche_adderall_fracs['ResponseCt'] >= 10#filter out schools with very few responses. \n", + "niche_adderall_fracs = niche_adderall_fracs.loc[at_least_ten_responses]\n", + "for school_characteristic in ipeds_data.keys():\n", + " niche_adderall_fracs[school_characteristic] = niche_adderall_fracs['IPEDS_Id'].map(lambda x:ipeds_data[school_characteristic][x]\n", + " if x in ipeds_data[school_characteristic]\n", + " else None)\n", + " \n", + "#Make geographic plot and run regression to confirm discrepancies are significant. \n", + "region_vals = []\n", + "region_names = []\n", + "for region in list(set(ipeds_data['HD2013_Geographic_region'].values())):\n", + " region_idxs = niche_adderall_fracs['HD2013_Geographic_region'] == region\n", + " if region_idxs.sum() >= 10:\n", + " mean_val = niche_adderall_fracs.loc[region_idxs]['PctOfTotal'].mean()\n", + " if np.isnan(mean_val):\n", + " continue\n", + "\n", + " region_vals.append(mean_val)\n", + " region_names.append(' '.join([a for a in region.split() if len(a) > 2]))\n", + "model = sm.OLS.from_formula('PctOfTotal ~ HD2013_Geographic_region', data = niche_adderall_fracs).fit()\n", + "print model.summary()\n", + "region_vals = np.array(region_vals)\n", + "region_names = np.array(region_names)\n", + "sorted_idxs = np.argsort(region_vals)\n", + "barh(range(len(region_vals)), region_vals[sorted_idxs])\n", + "yticks(np.arange(len(region_vals)) + .5, region_names[sorted_idxs])\n", + "subplots_adjust(left = .3)\n", + "xlabel('Proportion of students reporting that prescription study drugs\\n are among the most popular on campus')\n", + "\n", + " " + ] + }, + { + "cell_type": "code", + "execution_count": 10, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Region\tProportion\n", + "New England\t39.8\n", + "Mid East\t38.5\n", + "Southeast\t37.2\n", + "Plains\t36.0\n", + "Great Lakes\t34.3\n", + "Southwest\t33.8\n", + "Far West\t28.2\n", + "Rocky Mountains\t23.9\n" + ] + } + ], + "source": [ + "#Create data for chart (this chart did not end up being included).\n", + "print 'Region\\tProportion'\n", + "for i in range(len(region_names)):\n", + " print '%s\\t%2.1f' % (region_names[sorted_idxs[::-1][i]], region_vals[sorted_idxs[::-1][i]] * 100)\n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "###### \"Study drugs were also more frequently used at colleges that were more selective or had higher median SAT or ACT scores. In the graph below, each point represents one school; the horizontal axis is the school’s 75th percentile ACT score, and the vertical axis is the fraction of students responding that study drugs are popular. The correlation is highly statistically significant\"" + ] + }, + { + "cell_type": "code", + "execution_count": 11, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Regressing percentage of students at each college saying study drugs are popular on various measures of college selectivity. \n", + "For each measure of selectivity (first column) we run two regressions: \n", + "a simple model -- study_drugs_popular ~ selectivity_measure\n", + "a more complex model -- study_drugs_popular ~ selectivity_measure + college_covs\n", + "where college_covs are college_racial_breakdown, college_gender_breakdown, college_type, and college_estimated_enrollment\n", + "\n", + "below the columns are selectivity_measure, simple_selectivity_measure_beta, simple_selectivity_measure_p, complex_selectivity_measure_beta, complex_selectivity_measure_p\n", + "\n", + "IC2013_SAT_Critical_Reading_75th_percentile_score 0.00068 1.352e-13 0.00057 9.374e-07\n", + "IC2013_SAT_Writing_75th_percentile_score 0.00059 3.890e-10 0.00062 2.139e-06\n", + "IC2013_SAT_Math_75th_percentile_score 0.00062 1.922e-13 0.00074 4.848e-10\n", + "IC2013_ACT_Composite_75th_percentile_score 0.01531 5.626e-18 0.01551 1.232e-09\n", + "DRVIC2013_Percent_admitted_total -0.00129 2.996e-06 -0.00141 3.155e-06\n" + ] + }, + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAERCAYAAAB7FtAjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nccgAAAAASUVORK5CYII=\n", + "text/plain": [ + "" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "achievement_columns = ['IC2013_SAT_Critical_Reading_75th_percentile_score', 'IC2013_SAT_Writing_75th_percentile_score', \n", + " 'IC2013_SAT_Math_75th_percentile_score', 'IC2013_ACT_Composite_75th_percentile_score', 'DRVIC2013_Percent_admitted_total']\n", + "control_columns = ['DRVEF2013_White', 'DRVEF2013_women', 'IC2013_Institutional_control_or_affiliation', \n", + " 'DRVEF2013_Asian', 'DRVEF2013_American_Indian_or_Alaska_Native', \n", + " 'DRVEF2013_Black_or_African_American', 'EFEST2013_Estimated_enrollment__total', \n", + " 'DRVEF2013_Hispanic_Latino', 'DRVEF2013_Native_Hawaiian_or_Other_Pacific_Islander']\n", + "print \"\"\"\n", + "Regressing percentage of students at each college saying study drugs are popular on various measures of college selectivity. \n", + "For each measure of selectivity (first column) we run two regressions: \n", + "a simple model -- study_drugs_popular ~ selectivity_measure\n", + "a more complex model -- study_drugs_popular ~ selectivity_measure + college_covs\n", + "where college_covs are college_racial_breakdown, college_gender_breakdown, college_type, and college_estimated_enrollment\n", + "\n", + "below the columns are selectivity_measure, simple_selectivity_measure_beta, simple_selectivity_measure_p, complex_selectivity_measure_beta, complex_selectivity_measure_p\n", + "\"\"\"\n", + "\n", + "for coef in achievement_columns:\n", + " model = sm.OLS.from_formula('PctOfTotal ~ %s' % coef, data = niche_adderall_fracs).fit()\n", + " model_controlling_for_covariates = sm.OLS.from_formula('PctOfTotal ~ %s + %s' % (coef, '+'.join(control_columns)), data = niche_adderall_fracs).fit()\n", + " print '%-60s %3.5f %2.3e %3.5f %2.3e' % (coef, model.params[coef], model.pvalues[coef], \n", + " model_controlling_for_covariates.params[coef], model_controlling_for_covariates.pvalues[coef])\n", + "scatter(niche_adderall_fracs['IC2013_ACT_Composite_75th_percentile_score'], niche_adderall_fracs['PctOfTotal'])\n", + "xlabel('75th Percentile ACT Score')\n", + "ylabel('Fraction of Students Reporting Adderall Popular')\n", + "xlim([20, 36])\n", + "ylim([0, .7])\n", + "\n", + "show()\n", + " \n", + " " + ] + }, + { + "cell_type": "code", + "execution_count": 14, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "ACT\tAdderall\tTrendline\n", + "27.000\t31.480\t35.021\n", + "29.000\t42.110\t38.082\n", + "30.000\t47.960\t39.613\n", + "23.000\t23.260\t28.898\n", + "28.000\t36.960\t36.552\n", + "28.000\t33.910\t36.552\n", + "23.000\t27.120\t28.898\n", + "30.000\t60.000\t39.613\n", + "27.000\t53.130\t35.021\n", + "24.000\t30.890\t30.429\n", + "29.000\t34.040\t38.082\n", + "29.000\t28.330\t38.082\n", + "30.000\t45.160\t39.613\n", + "30.000\t43.750\t39.613\n", + "23.000\t28.260\t28.898\n", + "26.000\t29.070\t33.490\n", + "33.000\t64.710\t44.205\n", + "30.000\t48.760\t39.613\n", + "33.000\t68.970\t44.205\n", + "24.000\t31.330\t30.429\n", + "28.000\t51.350\t36.552\n", + "32.000\t28.570\t42.674\n", + "31.000\t7.720\t41.144\n", + "25.000\t7.180\t31.959\n", + "34.000\t29.170\t45.736\n", + "27.000\t62.160\t35.021\n", + "32.000\t50.000\t42.674\n", + "30.000\t34.690\t39.613\n", + "26.000\t40.000\t33.490\n", + "30.000\t43.180\t39.613\n", + "26.000\t15.180\t33.490\n", + "24.000\t29.070\t30.429\n", + "22.000\t10.680\t27.367\n", + "24.000\t12.350\t30.429\n", + "24.000\t22.360\t30.429\n", + "22.000\t19.650\t27.367\n", + "22.000\t13.890\t27.367\n", + "20.000\t11.460\t24.306\n", + "23.000\t28.000\t28.898\n", + "34.000\t41.940\t45.736\n", + "26.000\t23.260\t33.490\n", + "27.000\t35.560\t35.021\n", + "33.000\t37.140\t44.205\n", + "27.000\t40.540\t35.021\n", + "25.000\t21.150\t31.959\n", + "24.000\t34.650\t30.429\n", + "23.000\t26.090\t28.898\n", + "29.000\t34.180\t38.082\n", + "28.000\t25.000\t36.552\n", + "31.000\t54.130\t41.144\n", + "25.000\t20.310\t31.959\n", + "23.000\t37.680\t28.898\n", + "32.000\t48.000\t42.674\n", + "32.000\t47.730\t42.674\n", + "28.000\t55.130\t36.552\n", + "23.000\t40.000\t28.898\n", + "32.000\t40.630\t42.674\n", + "30.000\t33.850\t39.613\n", + "32.000\t39.390\t42.674\n", + "31.000\t29.410\t41.144\n", + "27.000\t32.370\t35.021\n", + "34.000\t40.380\t45.736\n", + "29.000\t37.930\t38.082\n", + "34.000\t48.810\t45.736\n", + "30.000\t37.210\t39.613\n", + "34.000\t43.590\t45.736\n", + "31.000\t68.750\t41.144\n", + "28.000\t30.000\t36.552\n", + "31.000\t54.170\t41.144\n", + "30.000\t31.860\t39.613\n", + "34.000\t42.550\t45.736\n", + "27.000\t40.820\t35.021\n", + "24.000\t48.480\t30.429\n", + "22.000\t36.360\t27.367\n", + "24.000\t30.770\t30.429\n", + "24.000\t28.260\t30.429\n", + "24.000\t25.420\t30.429\n", + "24.000\t14.130\t30.429\n", + "23.000\t28.070\t28.898\n", + "28.000\t39.470\t36.552\n", + "22.000\t29.550\t27.367\n", + "26.000\t50.000\t33.490\n", + "30.000\t25.640\t39.613\n", + "32.000\t40.910\t42.674\n", + "25.000\t53.330\t31.959\n", + "24.000\t38.570\t30.429\n", + "25.000\t38.890\t31.959\n", + "24.000\t25.560\t30.429\n", + "24.000\t30.610\t30.429\n", + "27.000\t25.900\t35.021\n", + "27.000\t28.570\t35.021\n", + "29.000\t49.810\t38.082\n", + "30.000\t40.540\t39.613\n", + "25.000\t50.000\t31.959\n", + "28.000\t25.810\t36.552\n", + "31.000\t60.240\t41.144\n", + "33.000\t47.690\t44.205\n", + "26.000\t32.650\t33.490\n", + "32.000\t44.590\t42.674\n", + "25.000\t40.560\t31.959\n", + "25.000\t26.200\t31.959\n", + "26.000\t37.230\t33.490\n", + "32.000\t38.460\t42.674\n", + "23.000\t15.940\t28.898\n", + "27.000\t55.000\t35.021\n", + "35.000\t30.560\t47.267\n", + "26.000\t47.460\t33.490\n", + "28.000\t32.840\t36.552\n", + "29.000\t39.020\t38.082\n", + "26.000\t15.280\t33.490\n", + "24.000\t13.510\t30.429\n", + "26.000\t35.560\t33.490\n", + "22.000\t23.610\t27.367\n", + "30.000\t51.530\t39.613\n", + "22.000\t28.710\t27.367\n", + "24.000\t19.640\t30.429\n", + "25.000\t24.620\t31.959\n", + "28.000\t30.220\t36.552\n", + "27.000\t47.320\t35.021\n", + "27.000\t50.000\t35.021\n", + "34.000\t52.830\t45.736\n", + "24.000\t28.240\t30.429\n", + "25.000\t34.420\t31.959\n", + "22.000\t27.660\t27.367\n", + "25.000\t33.330\t31.959\n", + "31.000\t46.430\t41.144\n", + "26.000\t48.280\t33.490\n", + "32.000\t57.500\t42.674\n", + "26.000\t7.270\t33.490\n", + "23.000\t41.670\t28.898\n", + "28.000\t52.890\t36.552\n", + "29.000\t39.130\t38.082\n", + "29.000\t37.000\t38.082\n", + "29.000\t44.440\t38.082\n", + "28.000\t33.330\t36.552\n", + "27.000\t41.670\t35.021\n", + "29.000\t47.140\t38.082\n", + "24.000\t22.810\t30.429\n", + "26.000\t40.740\t33.490\n", + "26.000\t40.910\t33.490\n", + "27.000\t50.000\t35.021\n", + "28.000\t30.000\t36.552\n", + "23.000\t13.890\t28.898\n", + "30.000\t57.380\t39.613\n", + "28.000\t41.880\t36.552\n", + "24.000\t31.780\t30.429\n", + "33.000\t57.690\t44.205\n", + "24.000\t29.270\t30.429\n", + "24.000\t40.540\t30.429\n", + "28.000\t34.380\t36.552\n", + "26.000\t28.570\t33.490\n", + "31.000\t33.330\t41.144\n", + "25.000\t52.380\t31.959\n", + "25.000\t35.480\t31.959\n", + "27.000\t23.290\t35.021\n", + "23.000\t28.570\t28.898\n", + "24.000\t16.380\t30.429\n", + "32.000\t48.190\t42.674\n", + "30.000\t38.390\t39.613\n", + "26.000\t27.270\t33.490\n", + "33.000\t40.000\t44.205\n", + "25.000\t22.950\t31.959\n", + "24.000\t27.000\t30.429\n", + "25.000\t16.220\t31.959\n", + "25.000\t33.330\t31.959\n", + "24.000\t32.260\t30.429\n", + "34.000\t32.760\t45.736\n", + "26.000\t30.770\t33.490\n", + "31.000\t37.380\t41.144\n", + "26.000\t42.640\t33.490\n", + "27.000\t57.140\t35.021\n", + "28.000\t37.840\t36.552\n", + "23.000\t30.970\t28.898\n", + "27.000\t35.510\t35.021\n", + "25.000\t36.230\t31.959\n", + "25.000\t61.760\t31.959\n", + "29.000\t36.840\t38.082\n", + "23.000\t35.140\t28.898\n", + "23.000\t33.330\t28.898\n", + "32.000\t43.480\t42.674\n", + "28.000\t43.590\t36.552\n", + "30.000\t36.130\t39.613\n", + "27.000\t53.230\t35.021\n", + "22.000\t41.670\t27.367\n", + "26.000\t35.290\t33.490\n", + "31.000\t38.710\t41.144\n", + "22.000\t35.290\t27.367\n", + "31.000\t41.380\t41.144\n", + "25.000\t24.390\t31.959\n", + "31.000\t21.880\t41.144\n", + "29.000\t50.980\t38.082\n", + "24.000\t25.860\t30.429\n", + "26.000\t30.230\t33.490\n", + "30.000\t34.210\t39.613\n", + "28.000\t41.670\t36.552\n", + "27.000\t27.910\t35.021\n", + "24.000\t36.960\t30.429\n", + "26.000\t42.000\t33.490\n", + "26.000\t41.380\t33.490\n", + "23.000\t27.420\t28.898\n", + "26.000\t42.940\t33.490\n", + "24.000\t20.740\t30.429\n", + "25.000\t27.160\t31.959\n", + "31.000\t35.480\t41.144\n", + "27.000\t20.800\t35.021\n", + "28.000\t33.330\t36.552\n", + "27.000\t31.250\t35.021\n", + "30.000\t66.670\t39.613\n", + "22.000\t38.100\t27.367\n", + "27.000\t44.120\t35.021\n", + "30.000\t32.350\t39.613\n", + "23.000\t32.690\t28.898\n", + "25.000\t26.190\t31.959\n", + "24.000\t30.000\t30.429\n", + "22.000\t22.640\t27.367\n", + "25.000\t20.000\t31.959\n", + "25.000\t16.950\t31.959\n", + "31.000\t58.460\t41.144\n", + "26.000\t18.570\t33.490\n", + "24.000\t28.950\t30.429\n", + "26.000\t38.100\t33.490\n", + "27.000\t50.000\t35.021\n", + "27.000\t27.350\t35.021\n", + "30.000\t42.310\t39.613\n", + "24.000\t29.270\t30.429\n", + "34.000\t20.750\t45.736\n", + "24.000\t28.410\t30.429\n", + "28.000\t21.430\t36.552\n", + "30.000\t33.650\t39.613\n", + "25.000\t41.030\t31.959\n", + "26.000\t35.710\t33.490\n", + "25.000\t40.480\t31.959\n", + "26.000\t25.580\t33.490\n", + "29.000\t44.000\t38.082\n", + "27.000\t22.220\t35.021\n", + "26.000\t42.590\t33.490\n", + "24.000\t34.210\t30.429\n", + "26.000\t32.500\t33.490\n", + "27.000\t35.710\t35.021\n", + "28.000\t47.480\t36.552\n", + "23.000\t23.210\t28.898\n", + "27.000\t31.070\t35.021\n", + "26.000\t28.210\t33.490\n", + "29.000\t40.840\t38.082\n", + "23.000\t30.770\t28.898\n", + "29.000\t63.460\t38.082\n", + "25.000\t43.200\t31.959\n", + "27.000\t47.100\t35.021\n", + "24.000\t39.220\t30.429\n", + "25.000\t34.910\t31.959\n", + "29.000\t51.520\t38.082\n", + "24.000\t27.270\t30.429\n", + "30.000\t29.730\t39.613\n", + "33.000\t31.580\t44.205\n", + "32.000\t64.620\t42.674\n", + "31.000\t43.330\t41.144\n", + "26.000\t50.000\t33.490\n", + "28.000\t40.210\t36.552\n", + "25.000\t27.030\t31.959\n", + "30.000\t47.930\t39.613\n", + "28.000\t28.570\t36.552\n", + "27.000\t38.100\t35.021\n", + "28.000\t49.170\t36.552\n", + "33.000\t35.810\t44.205\n", + "29.000\t17.050\t38.082\n", + "27.000\t16.550\t35.021\n", + "31.000\t35.100\t41.144\n", + "24.000\t14.930\t30.429\n", + "25.000\t23.530\t31.959\n", + "31.000\t33.330\t41.144\n", + "29.000\t47.620\t38.082\n", + "26.000\t17.200\t33.490\n", + "26.000\t37.500\t33.490\n", + "28.000\t30.090\t36.552\n", + "35.000\t38.100\t47.267\n", + "28.000\t35.430\t36.552\n", + "29.000\t38.380\t38.082\n", + "25.000\t18.750\t31.959\n", + "30.000\t51.970\t39.613\n", + "29.000\t45.830\t38.082\n", + "29.000\t44.700\t38.082\n", + "30.000\t37.500\t39.613\n", + "31.000\t43.010\t41.144\n", + "30.000\t44.390\t39.613\n", + "26.000\t35.090\t33.490\n", + "27.000\t24.730\t35.021\n", + "27.000\t26.940\t35.021\n", + "26.000\t31.030\t33.490\n", + "31.000\t40.000\t41.144\n", + "28.000\t55.280\t36.552\n", + "28.000\t47.790\t36.552\n", + "28.000\t50.440\t36.552\n", + "25.000\t32.690\t31.959\n", + "28.000\t25.300\t36.552\n", + "26.000\t31.340\t33.490\n", + "29.000\t31.940\t38.082\n", + "29.000\t38.130\t38.082\n", + "25.000\t40.000\t31.959\n", + "28.000\t28.890\t36.552\n", + "25.000\t17.500\t31.959\n", + "32.000\t46.360\t42.674\n", + "32.000\t41.780\t42.674\n", + "27.000\t22.920\t35.021\n", + "26.000\t60.530\t33.490\n", + "30.000\t31.030\t39.613\n", + "27.000\t55.560\t35.021\n", + "28.000\t52.800\t36.552\n", + "26.000\t30.000\t33.490\n", + "28.000\t27.470\t36.552\n", + "24.000\t11.590\t30.429\n", + "26.000\t27.340\t33.490\n", + "26.000\t30.770\t33.490\n", + "25.000\t23.380\t31.959\n", + "25.000\t19.260\t31.959\n", + "25.000\t32.260\t31.959\n", + "25.000\t23.260\t31.959\n", + "26.000\t45.900\t33.490\n", + "32.000\t45.120\t42.674\n", + "25.000\t40.830\t31.959\n", + "26.000\t31.820\t33.490\n", + "26.000\t28.890\t33.490\n", + "26.000\t23.300\t33.490\n", + "24.000\t28.240\t30.429\n", + "24.000\t33.330\t30.429\n", + "34.000\t19.300\t45.736\n", + "29.000\t44.860\t38.082\n", + "27.000\t42.860\t35.021\n", + "34.000\t38.100\t45.736\n", + "30.000\t35.970\t39.613\n", + "30.000\t35.480\t39.613\n", + "26.000\t42.570\t33.490\n", + "31.000\t54.290\t41.144\n", + "32.000\t31.250\t42.674\n", + "30.000\t46.250\t39.613\n", + "27.000\t34.090\t35.021\n", + "27.000\t40.000\t35.021\n", + "26.000\t21.740\t33.490\n", + "29.000\t47.480\t38.082\n", + "26.000\t34.380\t33.490\n", + "28.000\t30.120\t36.552\n", + "33.000\t30.970\t44.205\n", + "27.000\t23.810\t35.021\n", + "28.000\t40.000\t36.552\n", + "26.000\t39.290\t33.490\n", + "29.000\t51.430\t38.082\n", + "25.000\t44.230\t31.959\n", + "25.000\t25.640\t31.959\n", + "26.000\t22.790\t33.490\n", + "31.000\t44.700\t41.144\n", + "31.000\t18.920\t41.144\n", + "21.000\t19.790\t25.837\n", + "25.000\t25.000\t31.959\n", + "25.000\t22.730\t31.959\n", + "32.000\t47.830\t42.674\n", + "27.000\t23.130\t35.021\n", + "29.000\t49.300\t38.082\n", + "33.000\t40.660\t44.205\n", + "30.000\t28.160\t39.613\n", + "26.000\t28.380\t33.490\n", + "22.000\t34.020\t27.367\n", + "30.000\t52.590\t39.613\n", + "26.000\t25.490\t33.490\n", + "26.000\t24.070\t33.490\n", + "24.000\t34.860\t30.429\n", + "24.000\t32.760\t30.429\n", + "25.000\t25.640\t31.959\n", + "24.000\t31.580\t30.429\n", + "27.000\t22.080\t35.021\n", + "27.000\t9.870\t35.021\n", + "24.000\t30.160\t30.429\n", + "34.000\t52.080\t45.736\n", + "33.000\t34.380\t44.205\n", + "31.000\t50.850\t41.144\n", + "26.000\t25.630\t33.490\n", + "27.000\t43.480\t35.021\n", + "33.000\t62.500\t44.205\n", + "29.000\t41.380\t38.082\n", + "25.000\t28.480\t31.959\n", + "34.000\t27.140\t45.736\n", + "26.000\t21.650\t33.490\n", + "33.000\t36.670\t44.205\n", + "26.000\t51.580\t33.490\n", + "24.000\t32.610\t30.429\n", + "23.000\t21.670\t28.898\n", + "25.000\t31.430\t31.959\n", + "25.000\t42.150\t31.959\n", + "28.000\t27.030\t36.552\n", + "22.000\t45.450\t27.367\n", + "28.000\t27.500\t36.552\n", + "25.000\t37.040\t31.959\n", + "29.000\t25.640\t38.082\n", + "25.000\t32.350\t31.959\n", + "25.000\t22.730\t31.959\n", + "27.000\t48.840\t35.021\n", + "35.000\t27.910\t47.267\n", + "25.000\t34.620\t31.959\n" + ] + } + ], + "source": [ + "#Print out data for chart. \n", + "model = sm.OLS.from_formula('PctOfTotal ~ IC2013_ACT_Composite_75th_percentile_score', data = niche_adderall_fracs).fit()\n", + "ones_to_plot = niche_adderall_fracs[['IC2013_ACT_Composite_75th_percentile_score', 'PctOfTotal']].dropna()\n", + "ones_to_plot['trendline'] = niche_adderall_fracs['IC2013_ACT_Composite_75th_percentile_score'] * model.params['IC2013_ACT_Composite_75th_percentile_score'] + model.params['Intercept']\n", + "ones_to_plot['trendline'] = ones_to_plot['trendline'] * 100\n", + "print 'ACT\\tAdderall\\tTrendline'\n", + "for i in range(len(ones_to_plot)):\n", + " print '%2.3f\\t%2.3f\\t%2.3f' % (ones_to_plot['IC2013_ACT_Composite_75th_percentile_score'].iloc[i],\n", + " ones_to_plot['PctOfTotal'].iloc[i] * 100, \n", + " ones_to_plot['trendline'].iloc[i])" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "###### \"When I looked at the NSDUH data, I found that college students who had used Adderall non-medically reported significantly higher levels of depression and were more likely to have considered suicide.\"" + ] + }, + { + "cell_type": "code", + "execution_count": 15, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "\n", + "Looking at depression levels in college students\n", + "Category seriously_thought_about_killing_oneself_last_year\n", + "Mean value of ADDERALL for level 0.0 is 0.143; unweighted, 0.136 (3944 values)\n", + "Mean value of ADDERALL for level 1.0 is 0.213; unweighted, 0.191 (366 values)\n", + "Ratio between maximum value and minimum value 1.48723023239\n", + "Optimization terminated successfully.\n", + " Current function value: 0.404718\n", + " Iterations 6\n", + "Optimization terminated successfully.\n", + " Current function value: 0.389079\n", + " Iterations 7\n", + "Optimization terminated successfully.\n", + " Current function value: 0.403017\n", + " Iterations 6\n", + "Optimization terminated successfully.\n", + " Current function value: 0.386725\n", + " Iterations 7\n", + "\n", + "=========================================================================================\n", + " 0 1 2 3 \n", + "-----------------------------------------------------------------------------------------\n", + "seriously_thought_about_killing_oneself_last_year 0.410*** 0.435*** \n", + " (0.141) (0.144) \n", + "worst_K6_score_past_year 0.033*** 0.039*** \n", + " (0.007) (0.007) \n", + "C(IRFSTAMP, Sum)[S.1] -0.059 -0.076 \n", + " (0.089) (0.089) \n", + "C(NEWRACE2, Sum)[S.1] 0.746*** 0.754*** \n", + " (0.186) (0.186) \n", + "C(NEWRACE2, Sum)[S.2] -0.693*** -0.676***\n", + " (0.248) (0.249) \n", + "C(NEWRACE2, Sum)[S.3] 0.152 0.209 \n", + " (0.549) (0.552) \n", + "C(NEWRACE2, Sum)[S.4] -0.499 -0.554 \n", + " (0.892) (0.893) \n", + "C(NEWRACE2, Sum)[S.5] -0.038 -0.056 \n", + " (0.245) (0.245) \n", + "C(NEWRACE2, Sum)[S.6] 0.562** 0.533** \n", + " (0.249) (0.250) \n", + "IRSEX -0.373*** -0.439***\n", + " (0.090) (0.091) \n", + "Intercept -1.852*** -1.781*** -2.058*** -1.948***\n", + " (0.047) (0.236) (0.070) (0.240) \n", + "N 4310 4310 4310 4310 \n", + "=========================================================================================\n", + "Standard errors in parentheses.\n", + "* p<.1, ** p<.05, ***p<.01\n", + "\n", + "\n", + "Looking at depression levels in adults\n", + "Category seriously_thought_about_killing_oneself_last_year\n", + "Mean value of ADDERALL for level 0.0 is 0.026; unweighted, 0.034 (18526 values)\n", + "Mean value of ADDERALL for level 1.0 is 0.067; unweighted, 0.093 (756 values)\n", + "Ratio between maximum value and minimum value 2.54581324356\n", + "Optimization terminated successfully.\n", + " Current function value: 0.153638\n", + " Iterations 7\n", + "Optimization terminated successfully.\n", + " Current function value: 0.148860\n", + " Iterations 8\n", + "Optimization terminated successfully.\n", + " Current function value: 0.147275\n", + " Iterations 8\n", + "Optimization terminated successfully.\n", + " Current function value: 0.142961\n", + " Iterations 8\n", + "\n", + "=========================================================================================\n", + " 0 1 2 3 \n", + "-----------------------------------------------------------------------------------------\n", + "seriously_thought_about_killing_oneself_last_year 1.074*** 0.908*** \n", + " (0.132) (0.135) \n", + "worst_K6_score_past_year 0.098*** 0.097*** \n", + " (0.005) (0.006) \n", + "C(IRFSTAMP, Sum)[S.1] 0.292*** 0.157*** \n", + " (0.046) (0.048) \n", + "C(NEWRACE2, Sum)[S.1] 0.565*** 0.529*** \n", + " (0.111) (0.112) \n", + "C(NEWRACE2, Sum)[S.2] -0.848*** -0.825***\n", + " (0.190) (0.191) \n", + "C(NEWRACE2, Sum)[S.3] 0.490* 0.433 \n", + " (0.263) (0.267) \n", + "C(NEWRACE2, Sum)[S.4] 0.225 0.309 \n", + " (0.446) (0.448) \n", + "C(NEWRACE2, Sum)[S.5] -0.461* -0.432 \n", + " (0.268) (0.270) \n", + "C(NEWRACE2, Sum)[S.6] 0.750*** 0.668*** \n", + " (0.199) (0.201) \n", + "IRSEX -0.262*** -0.415***\n", + " (0.078) (0.080) \n", + "Intercept -3.357*** -3.090*** -3.948*** -3.505***\n", + " (0.041) (0.158) (0.061) (0.163) \n", + "N 19282 19282 19282 19282 \n", + "=========================================================================================\n", + "Standard errors in parentheses.\n", + "* p<.1, ** p<.05, ***p<.01\n" + ] + } + ], + "source": [ + "def lookAtDepressionLevels(nsduh_data, idxs):\n", + " nsduh_data['seriously_thought_about_killing_oneself_last_year'] = 1.* (nsduh_data['MHSUITHK'] == 1)\n", + " nsduh_data['worst_K6_score_past_year'] = 1.*nsduh_data['K6SCMAX']\n", + " data_to_use = nsduh_data.loc[idxs]\n", + " weights = data_to_use['ANALWT_C']\n", + " stratifyByCat(data_to_use, 'seriously_thought_about_killing_oneself_last_year') \n", + "\n", + " model1 = sm.Logit.from_formula('ADDERALL ~ seriously_thought_about_killing_oneself_last_year', \n", + " weights = weights, \n", + " data = data_to_use).fit()\n", + " model2 = sm.Logit.from_formula('ADDERALL ~ seriously_thought_about_killing_oneself_last_year + IRSEX + C(IRFSTAMP, Sum) + C(NEWRACE2, Sum)', \n", + " weights = weights, \n", + " data = data_to_use).fit()\n", + " model3 = sm.Logit.from_formula('ADDERALL ~ worst_K6_score_past_year', \n", + " weights = weights, \n", + " data = data_to_use).fit()\n", + " model4 = sm.Logit.from_formula('ADDERALL ~ worst_K6_score_past_year + IRSEX + C(IRFSTAMP, Sum) + C(NEWRACE2, Sum)', \n", + " weights = weights, \n", + " data = data_to_use).fit()\n", + "\n", + "\n", + " models = [model1, model2, model3, model4]\n", + " print summary_col(models, model_names = range(len(models)), stars = True,\n", + " regressor_order = ['seriously_thought_about_killing_oneself_last_year', 'worst_K6_score_past_year'],\n", + " float_format='%0.3f',\n", + " info_dict={'N':lambda x: \"{0:d}\".format(int(x.nobs))})\n", + "college_idxs = nsduh_data['COLLENR'] == 1\n", + "adult_idxs = nsduh_data['CATAG6'] >= 3\n", + "\n", + "print '\\n\\nLooking at depression levels in college students'\n", + "lookAtDepressionLevels(nsduh_data, college_idxs)\n", + "print '\\n\\nLooking at depression levels in adults'\n", + "lookAtDepressionLevels(nsduh_data, adult_idxs)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "###### \"I found that Google searches for Adderall in college towns spiked during exam months and drop during summer months.\"" + ] + }, + { + "cell_type": "code", + "execution_count": 16, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "0 towns processed out of 2084\n", + "Hamilton, New York not in index\n", + "Farmville, Virginia not in index\n", + "10 towns processed out of 2084\n", + "Canton, New York not in index\n", + "Orono, Maine not in index\n", + "20 towns processed out of 2084\n", + "Menomonie, Wisconsin not in index\n", + "Lexington, Virginia not in index\n", + "30 towns processed out of 2084\n", + "40 towns processed out of 2084\n", + "Geneseo, New York not in index\n", + "Kutztown, Pennsylvania not in index\n", + "50 towns processed out of 2084\n", + "60 towns processed out of 2084\n", + "Frostburg, Maryland not in index\n", + "70 towns processed out of 2084\n", + "town 4355\n", + "ADHD 4355\n", + "ADHD_zero_meaned_by_town 4355\n", + "month 4355\n", + "ADHD_normalized_by_town 4355\n", + "year 4355\n", + "Adderall 4355\n", + "Adderall_normalized_by_town 4355\n", + "Adderall_zero_meaned_by_town 4355\n" + ] + }, + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVMAAAFSCAYAAABPFzzRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkqkxxrjAkqkxxrjAkqkxxrjg/wF8PwJFsLeceAAAAABJRU5ErkJggg==\n", + "text/plain": [ + "" + ] + }, + "metadata": {}, + "output_type": "display_data" + }, + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVMAAAFSCAYAAABPFzzRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n", + "text/plain": [ + "" + ] + }, + "metadata": {}, + "output_type": "display_data" + }, + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVMAAAFSCAYAAABPFzzRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", + "text/plain": [ + "" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "college_towns_to_percentages = {}\n", + "#read in college towns. \n", + "for line in open('college_town_percentages'):\n", + " line = ' '.join(line.split()[1:])\n", + " city_name = line.split('(')[0].strip()\n", + " percentage = float(line.split('(')[1].replace('%)', ''))\n", + " if city_name in d.index:\n", + " college_towns_to_percentages[city_name] = percentage\n", + "#reprocess google data into easier-to-analyze format. \n", + "def createGoogleSearchDataframe(all_searches, college_towns_to_percentages):\n", + " df = {'month':[], 'year':[], 'town':[]}\n", + " for k in all_searches:\n", + " df[k] = []\n", + " df[k + '_normalized_by_town'] = []\n", + " df[k + '_zero_meaned_by_town'] = []\n", + " \n", + " for i, town_name in enumerate(college_towns_to_percentages.keys()):\n", + " if i % 10 == 0:\n", + " print i,'towns processed out of', len(d.index)\n", + " \n", + " mean_vals_by_town = {}\n", + " bad_town = False\n", + " for k in all_searches:\n", + " mean_vals_by_town[k] = {}\n", + " if town_name not in all_searches[k].index:\n", + " print town_name, 'not in index'\n", + " bad_town = True\n", + " if town_name in all_searches[k].index:\n", + " mean_vals_by_town[k]['mu'] = all_searches[k].loc[town_name].mean()\n", + " mean_vals_by_town[k]['sigma'] = all_searches[k].loc[town_name].std()\n", + " if bad_town:\n", + " continue\n", + " for date in all_searches['Adderall'].columns:\n", + " month = date.split('-')[1]\n", + " year = date.split('-')[0]\n", + " df['month'].append(month)\n", + " df['year'].append(year)\n", + " df['town'].append(town_name)\n", + " for k in all_searches.keys():\n", + " search_volume = float(all_searches[k].loc[town_name][date])\n", + " df[k].append(search_volume)\n", + " df[k + '_normalized_by_town'].append((search_volume - mean_vals_by_town[k]['mu']) / mean_vals_by_town[k]['sigma'])\n", + " df[k + '_zero_meaned_by_town'].append((search_volume - mean_vals_by_town[k]['mu']))\n", + "\n", + " for a in df.keys():\n", + " print a, len(df[a])\n", + " df = pd.DataFrame(df)\n", + " return df\n", + "google_search_dataframe = createGoogleSearchDataframe(google_data, college_towns_to_percentages)\n", + "\n", + "#Do regressions for various adjustments to the Google data to confirm we see the same pattern. \n", + "#Eg, if we subtract off the mean for each town, do we still see it? What if we divide by each town's standard deviation...? Pattern is robust to all these things. \n", + "for adjustment_to_use in ['', '_normalized_by_town', '_zero_meaned_by_town']:\n", + " figure(figsize = [5, 5])\n", + " for k in ['Adderall', 'ADHD']:\n", + " model = sm.OLS.from_formula('%s ~ C(month, Sum) + C(year, Sum)' \n", + " % (k + adjustment_to_use), data = google_search_dataframe).fit()\n", + " month_coefs = []\n", + " month_ticks = []\n", + " for param in model.params.index:\n", + " if 'month' in param:\n", + " month_coefs.append(model.params[param])\n", + " month_ticks.append(param.split('S.')[1].replace(']', ''))\n", + " month_coefs.append(- np.mean(month_coefs))\n", + " month_ticks.append('12')\n", + " plot(range(len(month_coefs)), month_coefs, label = k)\n", + " xticks(range(len(month_coefs)), month_ticks, rotation = 90)\n", + " ylim([-1, 1])\n", + " legend(loc = 3, fontsize = 8)\n", + "\n", + " xlabel('Month')\n", + " ylabel('Normalized Search Rate in College Towns')\n", + " title('Adjustment:' +adjustment_to_use)\n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "###### \"In fact, NSDUH data showed that adults whose family incomes were below 10,000 had the highest rates of nonmedical Adderall use, and those whose family incomes were greater than 75,000 had the lowest.\"" + ] + }, + { + "cell_type": "code", + "execution_count": 17, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Examining Adderall use frequency in adults as a function of family income\n", + "Ritalin results are qualitatively similar although the effects seem to be weaker\n", + "\n", + "Categories for total family income:\n", + "1: <10k, 2: 10-20, 3: 20-30; 4: 30-40; 5: 40-50; 6: 50-75; 7: 75\n", + "Category IRFAMIN3\n", + "Mean value of ADDERALL for level 1 is 0.037; unweighted, 0.053 (1132 values)\n", + "Mean value of ADDERALL for level 2 is 0.025; unweighted, 0.036 (2308 values)\n", + "Mean value of ADDERALL for level 3 is 0.032; unweighted, 0.042 (2102 values)\n", + "Mean value of ADDERALL for level 4 is 0.032; unweighted, 0.040 (2100 values)\n", + "Mean value of ADDERALL for level 5 is 0.029; unweighted, 0.036 (2085 values)\n", + "Mean value of ADDERALL for level 6 is 0.028; unweighted, 0.033 (3418 values)\n", + "Mean value of ADDERALL for level 7 is 0.024; unweighted, 0.031 (6137 values)\n", + "Ratio between maximum value and minimum value 1.52577800068\n", + "Category IRFAMIN3\n", + "Mean value of RITALIN for level 1 is 0.030; unweighted, 0.038 (1132 values)\n", + "Mean value of RITALIN for level 2 is 0.018; unweighted, 0.025 (2308 values)\n", + "Mean value of RITALIN for level 3 is 0.024; unweighted, 0.031 (2102 values)\n", + "Mean value of RITALIN for level 4 is 0.026; unweighted, 0.031 (2100 values)\n", + "Mean value of RITALIN for level 5 is 0.018; unweighted, 0.022 (2085 values)\n", + "Mean value of RITALIN for level 6 is 0.022; unweighted, 0.026 (3418 values)\n", + "Mean value of RITALIN for level 7 is 0.020; unweighted, 0.025 (6137 values)\n", + "Ratio between maximum value and minimum value 1.67773828665\n", + "Optimization terminated successfully.\n", + " Current function value: 0.154529\n", + " Iterations 7\n", + "Optimization terminated successfully.\n", + " Current function value: 0.154656\n", + " Iterations 7\n", + "Optimization terminated successfully.\n", + " Current function value: 0.132304\n", + " Iterations 11\n", + "Optimization terminated successfully.\n", + " Current function value: 0.124139\n", + " Iterations 8\n", + "Optimization terminated successfully.\n", + " Current function value: 0.124316\n", + " Iterations 8\n", + "Optimization terminated successfully.\n", + " Current function value: 0.111418\n", + " Iterations 10\n", + "\n", + "==========================================================================================\n", + " ADDERALL I ADDERALL II ADDERALL III RITALIN I RITALIN II RITALIN III\n", + "------------------------------------------------------------------------------------------\n", + "C(CATAG6, Sum)[S.3] 2.208*** 1.373*** \n", + " (0.186) (0.104) \n", + "C(CATAG6, Sum)[S.4] 0.742*** 0.467*** \n", + " (0.191) (0.110) \n", + "C(CATAG6, Sum)[S.5] -0.150 -0.244* \n", + " (0.215) (0.139) \n", + "C(IRFAMIN3)[T.2] -0.406** -0.387** -0.426** -0.426** \n", + " (0.173) (0.181) (0.205) (0.210) \n", + "C(IRFAMIN3)[T.3] -0.236 -0.365** -0.213 -0.360* \n", + " (0.171) (0.179) (0.200) (0.206) \n", + "C(IRFAMIN3)[T.4] -0.283 -0.458** -0.196 -0.401* \n", + " (0.173) (0.181) (0.200) (0.205) \n", + "C(IRFAMIN3)[T.5] -0.405** -0.650*** -0.560*** -0.837*** \n", + " (0.177) (0.185) (0.215) (0.221) \n", + "C(IRFAMIN3)[T.6] -0.484*** -0.825*** -0.390** -0.785*** \n", + " (0.163) (0.172) (0.189) (0.195) \n", + "C(IRFAMIN3)[T.7] -0.572*** -0.834*** -0.421** -0.811*** \n", + " (0.152) (0.161) (0.175) (0.182) \n", + "C(NEWRACE2, Sum)[S.1] 0.795*** 1.001*** \n", + " (0.114) (0.139) \n", + "C(NEWRACE2, Sum)[S.2] -0.798*** -1.727*** \n", + " (0.191) (0.346) \n", + "C(NEWRACE2, Sum)[S.3] 0.424 0.441 \n", + " (0.269) (0.334) \n", + "C(NEWRACE2, Sum)[S.4] 0.156 0.411 \n", + " (0.451) (0.516) \n", + "C(NEWRACE2, Sum)[S.5] -0.606** -0.480 \n", + " (0.270) (0.329) \n", + "C(NEWRACE2, Sum)[S.6] 0.858*** 1.182*** \n", + " (0.203) (0.226) \n", + "IRFAMIN3 -0.068*** -0.042* \n", + " (0.019) (0.022) \n", + "IRSEX -0.264*** -0.326*** \n", + " (0.080) (0.090) \n", + "Intercept -2.883*** -2.963*** -3.964*** -3.232*** -3.379*** -3.748*** \n", + " (0.133) (0.096) (0.274) (0.155) (0.112) (0.256) \n", + "N 19282 19282 19282 19282 19282 19282 \n", + "==========================================================================================\n", + "Standard errors in parentheses.\n", + "* p<.1, ** p<.05, ***p<.01\n" + ] + } + ], + "source": [ + "#run regressions + stratify by income levels to confirm that we observe income effects.\n", + "adult_idxs = nsduh_data['CATAG6'] >= 3\n", + "\n", + "data_to_use = nsduh_data.loc[adult_idxs]\n", + "weights = data_to_use['ANALWT_C']\n", + "print 'Examining Adderall use frequency in adults as a function of family income'\n", + "print 'Ritalin results are qualitatively similar although the effects seem to be weaker'\n", + "print '\\nCategories for total family income:\\n1: <10k, 2: 10-20, 3: 20-30; 4: 30-40; 5: 40-50; 6: 50-75; 7: 75'\n", + "stratifyByCat(data_to_use, 'IRFAMIN3') \n", + "stratifyByCat(data_to_use, 'IRFAMIN3', drug_to_measure = 'RITALIN')\n", + "\n", + "\n", + "model1 = sm.Logit.from_formula('ADDERALL ~ C(IRFAMIN3)', data = data_to_use, weights = weights).fit()#treat income as categorical variable\n", + "model2 = sm.Logit.from_formula('ADDERALL ~ IRFAMIN3', data = data_to_use, weights = weights).fit()#treat as continuous variable (bad, bad, bad)\n", + "model3 = sm.Logit.from_formula('ADDERALL ~ IRSEX + C(CATAG6, Sum) + C(NEWRACE2, Sum) + C(IRFAMIN3)', data = data_to_use, weights = weights).fit()#control for other covariates: age, race, sex. \n", + "model4 = sm.Logit.from_formula('RITALIN ~ C(IRFAMIN3)', data = data_to_use, weights = weights).fit()\n", + "model5 = sm.Logit.from_formula('RITALIN ~ IRFAMIN3', data = data_to_use, weights = weights).fit()\n", + "model6 = sm.Logit.from_formula('RITALIN ~ IRSEX + C(CATAG6, Sum) + C(NEWRACE2, Sum) + C(IRFAMIN3)', data = data_to_use, weights = weights).fit()\n", + "\n", + "models = [model1, model2, model3, model4, model5, model6]\n", + "print summary_col(models, model_names = ['ADDERALL', 'ADDERALL', 'ADDERALL', 'RITALIN', 'RITALIN', 'RITALIN'], stars = True,\n", + " float_format='%0.3f',\n", + " info_dict={'N':lambda x: \"{0:d}\".format(int(x.nobs))})" + ] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 2", + "language": "python", + "name": "python2" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 2 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython2", + "version": "2.7.10" + } + }, + "nbformat": 4, + "nbformat_minor": 0 +} diff --git a/study-drugs/README.md b/study-drugs/README.md new file mode 100644 index 0000000..9eb447b --- /dev/null +++ b/study-drugs/README.md @@ -0,0 +1,13 @@ +### Study Drugs + +This directory contains an IPython notebook of analysis for the story [College Students Aren’t The Only Ones Abusing Adderall](http://fivethirtyeight.com/features/college-students-arent-the-only-ones-abusing-adderall). + +The notebook cannot be run because it relies on datasets which are confidential. But to keep the analysis as transparent as possible, the author has provided the output and code. + +The NSDUH data can be accessed at: http://www.icpsr.umich.edu/icpsrweb/NACJD/studies/35509. + +The IPEDS data can be accessed at: https://nces.ed.gov/ipeds/datacenter/InstitutionByName.aspx. + +The Google and Niche datasets are confidential. + +Please contact emmap1@cs.stanford.edu with any comments or questions. \ No newline at end of file diff --git a/study-drugs/pierson_study_drugs.ipynb b/study-drugs/pierson_study_drugs.ipynb new file mode 100644 index 0000000..ffb9430 --- /dev/null +++ b/study-drugs/pierson_study_drugs.ipynb @@ -0,0 +1,1416 @@ +{ + "cells": [ + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## Analysis for the story [College Students Aren’t The Only Ones Abusing Adderall](http://fivethirtyeight.com/features/college-students-arent-the-only-ones-abusing-adderall)" + ] + }, + { + "cell_type": "code", + "execution_count": 18, + "metadata": { + "collapsed": false + }, + "outputs": [], + "source": [ + "%matplotlib inline " + ] + }, + { + "cell_type": "code", + "execution_count": 2, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "data": { + "text/html": [ + " " + ], + "text/plain": [ + "" + ] + }, + "execution_count": 2, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "from IPython.core.display import HTML\n", + "def css_styling():\n", + " styles = \" \"\n", + " return HTML(styles)\n", + "css_styling()\n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "# First load data" + ] + }, + { + "cell_type": "code", + "execution_count": 3, + "metadata": { + "collapsed": false + }, + "outputs": [], + "source": [ + "from pylab import *\n", + "import pandas as pd\n", + "import numpy as np\n", + "import os\n", + "import statsmodels.api as sm\n", + "from statsmodels.iolib.summary2 import summary_col\n", + "\n", + "#load in NSDUH data\n", + "nsduh_data = pd.read_csv('ICPSR_35509/DS0001/35509-0001-Data.tsv', sep = '\\t')\n", + "#load in Niche data\n", + "niche_data = pd.read_csv('niche/SchoolPoll.csv')\n", + "#load in IPEDS data\n", + "def getIPEDSData():\n", + " files = ['ipeds_data/CSV_7312015-162.csv', 'ipeds_data/CSV_7312015-826.csv']\n", + " \n", + " for i, f in enumerate(files):\n", + " d = pd.read_csv(f)\n", + " d.index = d.unitid \n", + " del d['unitid']\n", + " if i == 0:\n", + " all_d = d\n", + " else:\n", + " all_d = pd.concat([all_d, d])\n", + " ipeds_d = {}\n", + " for c in all_d.columns:\n", + " new_name = c.replace(' - ', '_').replace('/', '_').replace(',', '_').replace('.', '_').replace(' ', '_')\n", + " new_name = new_name.replace(\"Percent_of_total_enrollment_that_are_\", '')\n", + " ipeds_d[new_name] = dict(zip(all_d.index, all_d[c]))\n", + " return ipeds_d\n", + "ipeds_data = getIPEDSData()\n", + "#load in Google data\n", + "google_filenames = ['ADHD_Indexed.csv','Adderall_Indexed.csv']\n", + "google_data = {}\n", + "for filename in google_filenames:\n", + " d = pd.read_csv(os.path.join('google_data', filename))\n", + " d.index = d.Month\n", + " del d['Month']\n", + " d = d.transpose()\n", + " google_data[filename.replace('_Indexed.csv', '')] = d\n", + " \n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "###### \"When I looked at more recent data (the 2013 National Survey on Drug Use and Health (NSDUH), an annual government survey that includes more than 55,000 Americans) the difference turned out to be closer to 1.3x, not 2x.\"" + ] + }, + { + "cell_type": "code", + "execution_count": 5, + "metadata": { + "collapsed": false + }, + "outputs": [], + "source": [ + "#People who respond 1 have used Adderall non-medically; no one else has. \n", + "nsduh_data['ADDERALL'] = (nsduh_data['ADDERALL'] == 1)*1.\n", + "nsduh_data['RITALIN'] = (nsduh_data['RITMPHEN'] == 1)*1." + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "COLLENR levels:\n", + "1: FT college students age 18 - 22; 2: other age 18 - 22; 3: other\n", + "Category COLLENR\n", + "Mean value of ADDERALL for level 1 is 0.149; unweighted, 0.140 (4310 values)\n", + "Mean value of ADDERALL for level 2 is 0.115; unweighted, 0.114 (6934 values)\n", + "Mean value of ADDERALL for level 3 is 0.034; unweighted, 0.048 (43914 values)\n", + "Ratio between maximum value and minimum value 4.45754687876\n", + "Category COLLENR\n", + "Mean value of RITALIN for level 1 is 0.048; unweighted, 0.048 (4310 values)\n", + "Mean value of RITALIN for level 2 is 0.050; unweighted, 0.049 (6934 values)\n", + "Mean value of RITALIN for level 3 is 0.022; unweighted, 0.026 (43914 values)\n", + "Ratio between maximum value and minimum value 2.2055874138\n" + ] + } + ], + "source": [ + "\n", + "def stratifyByCat(nsduh_data, cat, drug_to_measure = 'ADDERALL', sortByVal = False):\n", + " \"\"\"\n", + " Checked. Computes the weighted and unweighted mean values of Adderall, stratifying by \n", + " the levels in cat. \n", + " \"\"\"\n", + " levels = sorted(list(set(nsduh_data[cat].dropna())))\n", + " print 'Category', cat\n", + " table = []\n", + " for l in levels:\n", + " idxs = nsduh_data[cat] == l\n", + " summed_weights = nsduh_data.loc[idxs]['ANALWT_C'].sum()\n", + " mu = (nsduh_data.loc[idxs][drug_to_measure] * nsduh_data.loc[idxs]['ANALWT_C']).sum() / summed_weights\n", + " unweighted_mu = nsduh_data.loc[idxs][drug_to_measure].mean()\n", + " if idxs.sum() > 25:\n", + " table.append([drug_to_measure, l, mu, unweighted_mu, idxs.sum()])\n", + " if sortByVal:\n", + " table = sorted(table, key = lambda x:x[2])[::-1]\n", + " for row in table:\n", + " print 'Mean value of %s for level %s is %2.3f; unweighted, %2.3f (%i values)' % (tuple(row))\n", + " print 'Ratio between maximum value and minimum value', max([a[2] for a in table]) / min([a[2] for a in table])\n", + "print 'COLLENR levels:\\n1: FT college students age 18 - 22; 2: other age 18 - 22; 3: other'\n", + "stratifyByCat(nsduh_data, 'COLLENR') \n", + "stratifyByCat(nsduh_data, 'COLLENR', drug_to_measure = 'RITALIN') \n", + "\n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "###### \"This is far smaller than the difference between white 18 - 22 year olds and black 18 - 22 year olds (6x, 18% vs 3%) or the difference between 18 - 22 year olds whose families do not receive food stamps and those whose do (1.6x, 14% vs 9%).\"" + ] + }, + { + "cell_type": "code", + "execution_count": 7, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Categories for race:\n", + "1: white, 2: black, 3: Native Am; 4: Native HI/Other Pac Isl; 5: Asian; 6: Multiracial; 7: Hispanic\n", + "Category NEWRACE2\n", + "Mean value of ADDERALL for level 1 is 0.181; unweighted, 0.171 (6194 values)\n", + "Mean value of ADDERALL for level 2 is 0.031; unweighted, 0.034 (1602 values)\n", + "Mean value of ADDERALL for level 3 is 0.052; unweighted, 0.086 (163 values)\n", + "Mean value of ADDERALL for level 4 is 0.084; unweighted, 0.078 (64 values)\n", + "Mean value of ADDERALL for level 5 is 0.084; unweighted, 0.086 (532 values)\n", + "Mean value of ADDERALL for level 6 is 0.113; unweighted, 0.155 (491 values)\n", + "Mean value of ADDERALL for level 7 is 0.076; unweighted, 0.065 (2198 values)\n", + "Ratio between maximum value and minimum value 5.84203431864\n", + "Category NEWRACE2\n", + "Mean value of RITALIN for level 1 is 0.071; unweighted, 0.069 (6194 values)\n", + "Mean value of RITALIN for level 2 is 0.007; unweighted, 0.009 (1602 values)\n", + "Mean value of RITALIN for level 3 is 0.026; unweighted, 0.037 (163 values)\n", + "Mean value of RITALIN for level 4 is 0.013; unweighted, 0.016 (64 values)\n", + "Mean value of RITALIN for level 5 is 0.029; unweighted, 0.028 (532 values)\n", + "Mean value of RITALIN for level 6 is 0.054; unweighted, 0.067 (491 values)\n", + "Mean value of RITALIN for level 7 is 0.026; unweighted, 0.022 (2198 values)\n", + "Ratio between maximum value and minimum value 9.9793813036\n", + "Categories for food stamp: 1: respondent or family member received food stamp; 2: did not\n", + "Category IRFSTAMP\n", + "Mean value of ADDERALL for level 1 is 0.088; unweighted, 0.089 (2564 values)\n", + "Mean value of ADDERALL for level 2 is 0.140; unweighted, 0.135 (8680 values)\n", + "Ratio between maximum value and minimum value 1.59577871899\n", + "Category IRFSTAMP\n", + "Mean value of RITALIN for level 1 is 0.041; unweighted, 0.043 (2564 values)\n", + "Mean value of RITALIN for level 2 is 0.051; unweighted, 0.050 (8680 values)\n", + "Ratio between maximum value and minimum value 1.24093935037\n" + ] + } + ], + "source": [ + "young_people_idxs = nsduh_data['COLLENR'].map(lambda x:x in [1, 2])\n", + "print '\\nCategories for race:\\n1: white, 2: black, 3: Native Am; 4: Native HI/Other Pac Isl; 5: Asian; 6: Multiracial; 7: Hispanic'\n", + "stratifyByCat(nsduh_data.loc[young_people_idxs], 'NEWRACE2') \n", + "stratifyByCat(nsduh_data.loc[young_people_idxs], 'NEWRACE2', drug_to_measure = 'RITALIN')\n", + "\n", + "print 'Categories for food stamp: 1: respondent or family member received food stamp; 2: did not'\n", + "stratifyByCat(nsduh_data.loc[young_people_idxs], 'IRFSTAMP') \n", + "stratifyByCat(nsduh_data.loc[young_people_idxs], 'IRFSTAMP', drug_to_measure = 'RITALIN') \n", + " \n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### Graphs on Adderall usage by age" + ] + }, + { + "cell_type": "code", + "execution_count": 8, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Category COLLENR\n", + "Mean value of ADDERALL for level 1 is 0.149; unweighted, 0.140 (4310 values)\n", + "Mean value of ADDERALL for level 2 is 0.115; unweighted, 0.114 (6934 values)\n", + "Mean value of ADDERALL for level 3 is 0.034; unweighted, 0.048 (43914 values)\n", + "Ratio between maximum value and minimum value 4.45754687876\n", + "Category age_cat_to_plot\n", + "Mean value of ADDERALL for level 12 - 14 is 0.004; unweighted, 0.005 (8689 values)\n", + "Mean value of ADDERALL for level 15 - 17 is 0.043; unweighted, 0.043 (9047 values)\n", + "Mean value of ADDERALL for level 23 - 25 is 0.145; unweighted, 0.140 (6896 values)\n", + "Mean value of ADDERALL for level 26 - 34 is 0.102; unweighted, 0.089 (5446 values)\n", + "Mean value of ADDERALL for level 35+ is 0.011; unweighted, 0.015 (13836 values)\n", + "Ratio between maximum value and minimum value 33.1820207109\n", + "Age\tPercent Using Adderall\n", + "12-14\t0.4\n", + "15-17\t4.3\n", + "18-22 (Not College)\t11.5\n", + "18-22 (In College)\t14.9\n", + "23-25\t14.5\n", + "26-34\t10.2\n", + "35+\t1.1\n" + ] + } + ], + "source": [ + "def remapAges(x):\n", + " if x <= 3:\n", + " return '12 - 14'\n", + " elif x <= 6:\n", + " return '15 - 17'\n", + " elif x <= 10:\n", + " return 'college age'\n", + " elif x <= 12:\n", + " return '23 - 25'\n", + " elif x <= 14:\n", + " return '26 - 34'\n", + " return '35+'\n", + "nsduh_data['age_cat_to_plot'] = nsduh_data['AGE2'].map(remapAges)\n", + "nsduh_data['age_cat_to_plot'].loc[nsduh_data['COLLENR'] != 3] = np.nan\n", + "\n", + "#This computes the values for Adderall usage stratified by age and college enrollment. \n", + "stratifyByCat(nsduh_data, 'COLLENR') \n", + "stratifyByCat(nsduh_data, 'age_cat_to_plot') \n", + "\n", + "#This prints out the actual data in an easy form for making graphs. Kind of hacky. \n", + "col_vals = [0.4, 4.3, 11.5, 14.9, 14.5, 10.2, 1.1]\n", + "col_labels = ['12-14', '15-17', '18-22 (Not College)', '18-22 (In College)', '23-25', '26-34', '35+']\n", + "print 'Age\\tPercent Using Adderall'\n", + "for i in range(len(col_vals)):\n", + " print '%s\\t%2.1f' % (col_labels[i], col_vals[i])" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "###### \"Study drugs were most frequently used in New England schools\"" + ] + }, + { + "cell_type": "code", + "execution_count": 9, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + " OLS Regression Results \n", + "==============================================================================\n", + "Dep. Variable: PctOfTotal R-squared: 0.126\n", + "Model: OLS Adj. R-squared: 0.110\n", + "Method: Least Squares F-statistic: 7.700\n", + "Date: Tue, 03 Nov 2015 Prob (F-statistic): 1.18e-09\n", + "Time: 14:09:22 Log-Likelihood: 350.49\n", + "No. Observations: 436 AIC: -683.0\n", + "Df Residuals: 427 BIC: -646.3\n", + "Df Model: 8 \n", + "Covariance Type: nonrobust \n", + "=============================================================================================================================================\n", + " coef std err t P>|t| [95.0% Conf. Int.]\n", + "---------------------------------------------------------------------------------------------------------------------------------------------\n", + "Intercept 0.2824 0.016 17.881 0.000 0.251 0.313\n", + "HD2013_Geographic_region[T.Great Lakes IL IN MI OH WI] 0.0604 0.021 2.901 0.004 0.019 0.101\n", + "HD2013_Geographic_region[T.Mid East DE DC MD NJ NY PA] 0.1021 0.019 5.330 0.000 0.064 0.140\n", + "HD2013_Geographic_region[T.New England CT ME MA NH RI VT] 0.1152 0.022 5.283 0.000 0.072 0.158\n", + "HD2013_Geographic_region[T.Outlying areas AS FM GU MH MP PR PW VI] 0.1047 0.111 0.946 0.344 -0.113 0.322\n", + "HD2013_Geographic_region[T.Plains IA KS MN MO NE ND SD] 0.0775 0.027 2.871 0.004 0.024 0.131\n", + "HD2013_Geographic_region[T.Rocky Mountains CO ID MT UT WY] -0.0432 0.030 -1.457 0.146 -0.102 0.015\n", + "HD2013_Geographic_region[T.Southeast AL AR FL GA KY LA MS NC SC TN VA WV] 0.0891 0.019 4.579 0.000 0.051 0.127\n", + "HD2013_Geographic_region[T.Southwest AZ NM OK TX] 0.0558 0.025 2.191 0.029 0.006 0.106\n", + "==============================================================================\n", + "Omnibus: 10.358 Durbin-Watson: 1.920\n", + "Prob(Omnibus): 0.006 Jarque-Bera (JB): 10.752\n", + "Skew: 0.384 Prob(JB): 0.00463\n", + "Kurtosis: 2.941 Cond. No. 22.7\n", + "==============================================================================\n", + "\n", + "Warnings:\n", + "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n" + ] + }, + { + "data": { + "text/plain": [ + "" + ] + }, + "execution_count": 9, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAEWCAYAAACAOivfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBwEzsxL7/0ZGrp7wszgNAAAAAElFTkSuQmCC\n", + "text/plain": [ + "" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "#Join Niche data with IPEDS data. \n", + "niche_adderall_idxs = (niche_data['Response'] == 'Prescription \"study drugs\" (Adderall, Ritalin) ')\n", + "niche_adderall_fracs = niche_data.loc[niche_adderall_idxs]\n", + "at_least_ten_responses = niche_adderall_fracs['ResponseCt'] >= 10#filter out schools with very few responses. \n", + "niche_adderall_fracs = niche_adderall_fracs.loc[at_least_ten_responses]\n", + "for school_characteristic in ipeds_data.keys():\n", + " niche_adderall_fracs[school_characteristic] = niche_adderall_fracs['IPEDS_Id'].map(lambda x:ipeds_data[school_characteristic][x]\n", + " if x in ipeds_data[school_characteristic]\n", + " else None)\n", + " \n", + "#Make geographic plot and run regression to confirm discrepancies are significant. \n", + "region_vals = []\n", + "region_names = []\n", + "for region in list(set(ipeds_data['HD2013_Geographic_region'].values())):\n", + " region_idxs = niche_adderall_fracs['HD2013_Geographic_region'] == region\n", + " if region_idxs.sum() >= 10:\n", + " mean_val = niche_adderall_fracs.loc[region_idxs]['PctOfTotal'].mean()\n", + " if np.isnan(mean_val):\n", + " continue\n", + "\n", + " region_vals.append(mean_val)\n", + " region_names.append(' '.join([a for a in region.split() if len(a) > 2]))\n", + "model = sm.OLS.from_formula('PctOfTotal ~ HD2013_Geographic_region', data = niche_adderall_fracs).fit()\n", + "print model.summary()\n", + "region_vals = np.array(region_vals)\n", + "region_names = np.array(region_names)\n", + "sorted_idxs = np.argsort(region_vals)\n", + "barh(range(len(region_vals)), region_vals[sorted_idxs])\n", + "yticks(np.arange(len(region_vals)) + .5, region_names[sorted_idxs])\n", + "subplots_adjust(left = .3)\n", + "xlabel('Proportion of students reporting that prescription study drugs\\n are among the most popular on campus')\n", + "\n", + " " + ] + }, + { + "cell_type": "code", + "execution_count": 10, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Region\tProportion\n", + "New England\t39.8\n", + "Mid East\t38.5\n", + "Southeast\t37.2\n", + "Plains\t36.0\n", + "Great Lakes\t34.3\n", + "Southwest\t33.8\n", + "Far West\t28.2\n", + "Rocky Mountains\t23.9\n" + ] + } + ], + "source": [ + "#Create data for chart (this chart did not end up being included).\n", + "print 'Region\\tProportion'\n", + "for i in range(len(region_names)):\n", + " print '%s\\t%2.1f' % (region_names[sorted_idxs[::-1][i]], region_vals[sorted_idxs[::-1][i]] * 100)\n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "###### \"Study drugs were also more frequently used at colleges that were more selective or had higher median SAT or ACT scores. In the graph below, each point represents one school; the horizontal axis is the school’s 75th percentile ACT score, and the vertical axis is the fraction of students responding that study drugs are popular. The correlation is highly statistically significant\"" + ] + }, + { + "cell_type": "code", + "execution_count": 11, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Regressing percentage of students at each college saying study drugs are popular on various measures of college selectivity. \n", + "For each measure of selectivity (first column) we run two regressions: \n", + "a simple model -- study_drugs_popular ~ selectivity_measure\n", + "a more complex model -- study_drugs_popular ~ selectivity_measure + college_covs\n", + "where college_covs are college_racial_breakdown, college_gender_breakdown, college_type, and college_estimated_enrollment\n", + "\n", + "below the columns are selectivity_measure, simple_selectivity_measure_beta, simple_selectivity_measure_p, complex_selectivity_measure_beta, complex_selectivity_measure_p\n", + "\n", + "IC2013_SAT_Critical_Reading_75th_percentile_score 0.00068 1.352e-13 0.00057 9.374e-07\n", + "IC2013_SAT_Writing_75th_percentile_score 0.00059 3.890e-10 0.00062 2.139e-06\n", + "IC2013_SAT_Math_75th_percentile_score 0.00062 1.922e-13 0.00074 4.848e-10\n", + "IC2013_ACT_Composite_75th_percentile_score 0.01531 5.626e-18 0.01551 1.232e-09\n", + "DRVIC2013_Percent_admitted_total -0.00129 2.996e-06 -0.00141 3.155e-06\n" + ] + }, + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAERCAYAAAB7FtAjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nccgAAAAASUVORK5CYII=\n", + "text/plain": [ + "" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "achievement_columns = ['IC2013_SAT_Critical_Reading_75th_percentile_score', 'IC2013_SAT_Writing_75th_percentile_score', \n", + " 'IC2013_SAT_Math_75th_percentile_score', 'IC2013_ACT_Composite_75th_percentile_score', 'DRVIC2013_Percent_admitted_total']\n", + "control_columns = ['DRVEF2013_White', 'DRVEF2013_women', 'IC2013_Institutional_control_or_affiliation', \n", + " 'DRVEF2013_Asian', 'DRVEF2013_American_Indian_or_Alaska_Native', \n", + " 'DRVEF2013_Black_or_African_American', 'EFEST2013_Estimated_enrollment__total', \n", + " 'DRVEF2013_Hispanic_Latino', 'DRVEF2013_Native_Hawaiian_or_Other_Pacific_Islander']\n", + "print \"\"\"\n", + "Regressing percentage of students at each college saying study drugs are popular on various measures of college selectivity. \n", + "For each measure of selectivity (first column) we run two regressions: \n", + "a simple model -- study_drugs_popular ~ selectivity_measure\n", + "a more complex model -- study_drugs_popular ~ selectivity_measure + college_covs\n", + "where college_covs are college_racial_breakdown, college_gender_breakdown, college_type, and college_estimated_enrollment\n", + "\n", + "below the columns are selectivity_measure, simple_selectivity_measure_beta, simple_selectivity_measure_p, complex_selectivity_measure_beta, complex_selectivity_measure_p\n", + "\"\"\"\n", + "\n", + "for coef in achievement_columns:\n", + " model = sm.OLS.from_formula('PctOfTotal ~ %s' % coef, data = niche_adderall_fracs).fit()\n", + " model_controlling_for_covariates = sm.OLS.from_formula('PctOfTotal ~ %s + %s' % (coef, '+'.join(control_columns)), data = niche_adderall_fracs).fit()\n", + " print '%-60s %3.5f %2.3e %3.5f %2.3e' % (coef, model.params[coef], model.pvalues[coef], \n", + " model_controlling_for_covariates.params[coef], model_controlling_for_covariates.pvalues[coef])\n", + "scatter(niche_adderall_fracs['IC2013_ACT_Composite_75th_percentile_score'], niche_adderall_fracs['PctOfTotal'])\n", + "xlabel('75th Percentile ACT Score')\n", + "ylabel('Fraction of Students Reporting Adderall Popular')\n", + "xlim([20, 36])\n", + "ylim([0, .7])\n", + "\n", + "show()\n", + " \n", + " " + ] + }, + { + "cell_type": "code", + "execution_count": 14, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "ACT\tAdderall\tTrendline\n", + "27.000\t31.480\t35.021\n", + "29.000\t42.110\t38.082\n", + "30.000\t47.960\t39.613\n", + "23.000\t23.260\t28.898\n", + "28.000\t36.960\t36.552\n", + "28.000\t33.910\t36.552\n", + "23.000\t27.120\t28.898\n", + "30.000\t60.000\t39.613\n", + "27.000\t53.130\t35.021\n", + "24.000\t30.890\t30.429\n", + "29.000\t34.040\t38.082\n", + "29.000\t28.330\t38.082\n", + "30.000\t45.160\t39.613\n", + "30.000\t43.750\t39.613\n", + "23.000\t28.260\t28.898\n", + "26.000\t29.070\t33.490\n", + "33.000\t64.710\t44.205\n", + "30.000\t48.760\t39.613\n", + "33.000\t68.970\t44.205\n", + "24.000\t31.330\t30.429\n", + "28.000\t51.350\t36.552\n", + "32.000\t28.570\t42.674\n", + "31.000\t7.720\t41.144\n", + "25.000\t7.180\t31.959\n", + "34.000\t29.170\t45.736\n", + "27.000\t62.160\t35.021\n", + "32.000\t50.000\t42.674\n", + "30.000\t34.690\t39.613\n", + "26.000\t40.000\t33.490\n", + "30.000\t43.180\t39.613\n", + "26.000\t15.180\t33.490\n", + "24.000\t29.070\t30.429\n", + "22.000\t10.680\t27.367\n", + "24.000\t12.350\t30.429\n", + "24.000\t22.360\t30.429\n", + "22.000\t19.650\t27.367\n", + "22.000\t13.890\t27.367\n", + "20.000\t11.460\t24.306\n", + "23.000\t28.000\t28.898\n", + "34.000\t41.940\t45.736\n", + "26.000\t23.260\t33.490\n", + "27.000\t35.560\t35.021\n", + "33.000\t37.140\t44.205\n", + "27.000\t40.540\t35.021\n", + "25.000\t21.150\t31.959\n", + "24.000\t34.650\t30.429\n", + "23.000\t26.090\t28.898\n", + "29.000\t34.180\t38.082\n", + "28.000\t25.000\t36.552\n", + "31.000\t54.130\t41.144\n", + "25.000\t20.310\t31.959\n", + "23.000\t37.680\t28.898\n", + "32.000\t48.000\t42.674\n", + "32.000\t47.730\t42.674\n", + "28.000\t55.130\t36.552\n", + "23.000\t40.000\t28.898\n", + "32.000\t40.630\t42.674\n", + "30.000\t33.850\t39.613\n", + "32.000\t39.390\t42.674\n", + "31.000\t29.410\t41.144\n", + "27.000\t32.370\t35.021\n", + "34.000\t40.380\t45.736\n", + "29.000\t37.930\t38.082\n", + "34.000\t48.810\t45.736\n", + "30.000\t37.210\t39.613\n", + "34.000\t43.590\t45.736\n", + "31.000\t68.750\t41.144\n", + "28.000\t30.000\t36.552\n", + "31.000\t54.170\t41.144\n", + "30.000\t31.860\t39.613\n", + "34.000\t42.550\t45.736\n", + "27.000\t40.820\t35.021\n", + "24.000\t48.480\t30.429\n", + "22.000\t36.360\t27.367\n", + "24.000\t30.770\t30.429\n", + "24.000\t28.260\t30.429\n", + "24.000\t25.420\t30.429\n", + "24.000\t14.130\t30.429\n", + "23.000\t28.070\t28.898\n", + "28.000\t39.470\t36.552\n", + "22.000\t29.550\t27.367\n", + "26.000\t50.000\t33.490\n", + "30.000\t25.640\t39.613\n", + "32.000\t40.910\t42.674\n", + "25.000\t53.330\t31.959\n", + "24.000\t38.570\t30.429\n", + "25.000\t38.890\t31.959\n", + "24.000\t25.560\t30.429\n", + "24.000\t30.610\t30.429\n", + "27.000\t25.900\t35.021\n", + "27.000\t28.570\t35.021\n", + "29.000\t49.810\t38.082\n", + "30.000\t40.540\t39.613\n", + "25.000\t50.000\t31.959\n", + "28.000\t25.810\t36.552\n", + "31.000\t60.240\t41.144\n", + "33.000\t47.690\t44.205\n", + "26.000\t32.650\t33.490\n", + "32.000\t44.590\t42.674\n", + "25.000\t40.560\t31.959\n", + "25.000\t26.200\t31.959\n", + "26.000\t37.230\t33.490\n", + "32.000\t38.460\t42.674\n", + "23.000\t15.940\t28.898\n", + "27.000\t55.000\t35.021\n", + "35.000\t30.560\t47.267\n", + "26.000\t47.460\t33.490\n", + "28.000\t32.840\t36.552\n", + "29.000\t39.020\t38.082\n", + "26.000\t15.280\t33.490\n", + "24.000\t13.510\t30.429\n", + "26.000\t35.560\t33.490\n", + "22.000\t23.610\t27.367\n", + "30.000\t51.530\t39.613\n", + "22.000\t28.710\t27.367\n", + "24.000\t19.640\t30.429\n", + "25.000\t24.620\t31.959\n", + "28.000\t30.220\t36.552\n", + "27.000\t47.320\t35.021\n", + "27.000\t50.000\t35.021\n", + "34.000\t52.830\t45.736\n", + "24.000\t28.240\t30.429\n", + "25.000\t34.420\t31.959\n", + "22.000\t27.660\t27.367\n", + "25.000\t33.330\t31.959\n", + "31.000\t46.430\t41.144\n", + "26.000\t48.280\t33.490\n", + "32.000\t57.500\t42.674\n", + "26.000\t7.270\t33.490\n", + "23.000\t41.670\t28.898\n", + "28.000\t52.890\t36.552\n", + "29.000\t39.130\t38.082\n", + "29.000\t37.000\t38.082\n", + "29.000\t44.440\t38.082\n", + "28.000\t33.330\t36.552\n", + "27.000\t41.670\t35.021\n", + "29.000\t47.140\t38.082\n", + "24.000\t22.810\t30.429\n", + "26.000\t40.740\t33.490\n", + "26.000\t40.910\t33.490\n", + "27.000\t50.000\t35.021\n", + "28.000\t30.000\t36.552\n", + "23.000\t13.890\t28.898\n", + "30.000\t57.380\t39.613\n", + "28.000\t41.880\t36.552\n", + "24.000\t31.780\t30.429\n", + "33.000\t57.690\t44.205\n", + "24.000\t29.270\t30.429\n", + "24.000\t40.540\t30.429\n", + "28.000\t34.380\t36.552\n", + "26.000\t28.570\t33.490\n", + "31.000\t33.330\t41.144\n", + "25.000\t52.380\t31.959\n", + "25.000\t35.480\t31.959\n", + "27.000\t23.290\t35.021\n", + "23.000\t28.570\t28.898\n", + "24.000\t16.380\t30.429\n", + "32.000\t48.190\t42.674\n", + "30.000\t38.390\t39.613\n", + "26.000\t27.270\t33.490\n", + "33.000\t40.000\t44.205\n", + "25.000\t22.950\t31.959\n", + "24.000\t27.000\t30.429\n", + "25.000\t16.220\t31.959\n", + "25.000\t33.330\t31.959\n", + "24.000\t32.260\t30.429\n", + "34.000\t32.760\t45.736\n", + "26.000\t30.770\t33.490\n", + "31.000\t37.380\t41.144\n", + "26.000\t42.640\t33.490\n", + "27.000\t57.140\t35.021\n", + "28.000\t37.840\t36.552\n", + "23.000\t30.970\t28.898\n", + "27.000\t35.510\t35.021\n", + "25.000\t36.230\t31.959\n", + "25.000\t61.760\t31.959\n", + "29.000\t36.840\t38.082\n", + "23.000\t35.140\t28.898\n", + "23.000\t33.330\t28.898\n", + "32.000\t43.480\t42.674\n", + "28.000\t43.590\t36.552\n", + "30.000\t36.130\t39.613\n", + "27.000\t53.230\t35.021\n", + "22.000\t41.670\t27.367\n", + "26.000\t35.290\t33.490\n", + "31.000\t38.710\t41.144\n", + "22.000\t35.290\t27.367\n", + "31.000\t41.380\t41.144\n", + "25.000\t24.390\t31.959\n", + "31.000\t21.880\t41.144\n", + "29.000\t50.980\t38.082\n", + "24.000\t25.860\t30.429\n", + "26.000\t30.230\t33.490\n", + "30.000\t34.210\t39.613\n", + "28.000\t41.670\t36.552\n", + "27.000\t27.910\t35.021\n", + "24.000\t36.960\t30.429\n", + "26.000\t42.000\t33.490\n", + "26.000\t41.380\t33.490\n", + "23.000\t27.420\t28.898\n", + "26.000\t42.940\t33.490\n", + "24.000\t20.740\t30.429\n", + "25.000\t27.160\t31.959\n", + "31.000\t35.480\t41.144\n", + "27.000\t20.800\t35.021\n", + "28.000\t33.330\t36.552\n", + "27.000\t31.250\t35.021\n", + "30.000\t66.670\t39.613\n", + "22.000\t38.100\t27.367\n", + "27.000\t44.120\t35.021\n", + "30.000\t32.350\t39.613\n", + "23.000\t32.690\t28.898\n", + "25.000\t26.190\t31.959\n", + "24.000\t30.000\t30.429\n", + "22.000\t22.640\t27.367\n", + "25.000\t20.000\t31.959\n", + "25.000\t16.950\t31.959\n", + "31.000\t58.460\t41.144\n", + "26.000\t18.570\t33.490\n", + "24.000\t28.950\t30.429\n", + "26.000\t38.100\t33.490\n", + "27.000\t50.000\t35.021\n", + "27.000\t27.350\t35.021\n", + "30.000\t42.310\t39.613\n", + "24.000\t29.270\t30.429\n", + "34.000\t20.750\t45.736\n", + "24.000\t28.410\t30.429\n", + "28.000\t21.430\t36.552\n", + "30.000\t33.650\t39.613\n", + "25.000\t41.030\t31.959\n", + "26.000\t35.710\t33.490\n", + "25.000\t40.480\t31.959\n", + "26.000\t25.580\t33.490\n", + "29.000\t44.000\t38.082\n", + "27.000\t22.220\t35.021\n", + "26.000\t42.590\t33.490\n", + "24.000\t34.210\t30.429\n", + "26.000\t32.500\t33.490\n", + "27.000\t35.710\t35.021\n", + "28.000\t47.480\t36.552\n", + "23.000\t23.210\t28.898\n", + "27.000\t31.070\t35.021\n", + "26.000\t28.210\t33.490\n", + "29.000\t40.840\t38.082\n", + "23.000\t30.770\t28.898\n", + "29.000\t63.460\t38.082\n", + "25.000\t43.200\t31.959\n", + "27.000\t47.100\t35.021\n", + "24.000\t39.220\t30.429\n", + "25.000\t34.910\t31.959\n", + "29.000\t51.520\t38.082\n", + "24.000\t27.270\t30.429\n", + "30.000\t29.730\t39.613\n", + "33.000\t31.580\t44.205\n", + "32.000\t64.620\t42.674\n", + "31.000\t43.330\t41.144\n", + "26.000\t50.000\t33.490\n", + "28.000\t40.210\t36.552\n", + "25.000\t27.030\t31.959\n", + "30.000\t47.930\t39.613\n", + "28.000\t28.570\t36.552\n", + "27.000\t38.100\t35.021\n", + "28.000\t49.170\t36.552\n", + "33.000\t35.810\t44.205\n", + "29.000\t17.050\t38.082\n", + "27.000\t16.550\t35.021\n", + "31.000\t35.100\t41.144\n", + "24.000\t14.930\t30.429\n", + "25.000\t23.530\t31.959\n", + "31.000\t33.330\t41.144\n", + "29.000\t47.620\t38.082\n", + "26.000\t17.200\t33.490\n", + "26.000\t37.500\t33.490\n", + "28.000\t30.090\t36.552\n", + "35.000\t38.100\t47.267\n", + "28.000\t35.430\t36.552\n", + "29.000\t38.380\t38.082\n", + "25.000\t18.750\t31.959\n", + "30.000\t51.970\t39.613\n", + "29.000\t45.830\t38.082\n", + "29.000\t44.700\t38.082\n", + "30.000\t37.500\t39.613\n", + "31.000\t43.010\t41.144\n", + "30.000\t44.390\t39.613\n", + "26.000\t35.090\t33.490\n", + "27.000\t24.730\t35.021\n", + "27.000\t26.940\t35.021\n", + "26.000\t31.030\t33.490\n", + "31.000\t40.000\t41.144\n", + "28.000\t55.280\t36.552\n", + "28.000\t47.790\t36.552\n", + "28.000\t50.440\t36.552\n", + "25.000\t32.690\t31.959\n", + "28.000\t25.300\t36.552\n", + "26.000\t31.340\t33.490\n", + "29.000\t31.940\t38.082\n", + "29.000\t38.130\t38.082\n", + "25.000\t40.000\t31.959\n", + "28.000\t28.890\t36.552\n", + "25.000\t17.500\t31.959\n", + "32.000\t46.360\t42.674\n", + "32.000\t41.780\t42.674\n", + "27.000\t22.920\t35.021\n", + "26.000\t60.530\t33.490\n", + "30.000\t31.030\t39.613\n", + "27.000\t55.560\t35.021\n", + "28.000\t52.800\t36.552\n", + "26.000\t30.000\t33.490\n", + "28.000\t27.470\t36.552\n", + "24.000\t11.590\t30.429\n", + "26.000\t27.340\t33.490\n", + "26.000\t30.770\t33.490\n", + "25.000\t23.380\t31.959\n", + "25.000\t19.260\t31.959\n", + "25.000\t32.260\t31.959\n", + "25.000\t23.260\t31.959\n", + "26.000\t45.900\t33.490\n", + "32.000\t45.120\t42.674\n", + "25.000\t40.830\t31.959\n", + "26.000\t31.820\t33.490\n", + "26.000\t28.890\t33.490\n", + "26.000\t23.300\t33.490\n", + "24.000\t28.240\t30.429\n", + "24.000\t33.330\t30.429\n", + "34.000\t19.300\t45.736\n", + "29.000\t44.860\t38.082\n", + "27.000\t42.860\t35.021\n", + "34.000\t38.100\t45.736\n", + "30.000\t35.970\t39.613\n", + "30.000\t35.480\t39.613\n", + "26.000\t42.570\t33.490\n", + "31.000\t54.290\t41.144\n", + "32.000\t31.250\t42.674\n", + "30.000\t46.250\t39.613\n", + "27.000\t34.090\t35.021\n", + "27.000\t40.000\t35.021\n", + "26.000\t21.740\t33.490\n", + "29.000\t47.480\t38.082\n", + "26.000\t34.380\t33.490\n", + "28.000\t30.120\t36.552\n", + "33.000\t30.970\t44.205\n", + "27.000\t23.810\t35.021\n", + "28.000\t40.000\t36.552\n", + "26.000\t39.290\t33.490\n", + "29.000\t51.430\t38.082\n", + "25.000\t44.230\t31.959\n", + "25.000\t25.640\t31.959\n", + "26.000\t22.790\t33.490\n", + "31.000\t44.700\t41.144\n", + "31.000\t18.920\t41.144\n", + "21.000\t19.790\t25.837\n", + "25.000\t25.000\t31.959\n", + "25.000\t22.730\t31.959\n", + "32.000\t47.830\t42.674\n", + "27.000\t23.130\t35.021\n", + "29.000\t49.300\t38.082\n", + "33.000\t40.660\t44.205\n", + "30.000\t28.160\t39.613\n", + "26.000\t28.380\t33.490\n", + "22.000\t34.020\t27.367\n", + "30.000\t52.590\t39.613\n", + "26.000\t25.490\t33.490\n", + "26.000\t24.070\t33.490\n", + "24.000\t34.860\t30.429\n", + "24.000\t32.760\t30.429\n", + "25.000\t25.640\t31.959\n", + "24.000\t31.580\t30.429\n", + "27.000\t22.080\t35.021\n", + "27.000\t9.870\t35.021\n", + "24.000\t30.160\t30.429\n", + "34.000\t52.080\t45.736\n", + "33.000\t34.380\t44.205\n", + "31.000\t50.850\t41.144\n", + "26.000\t25.630\t33.490\n", + "27.000\t43.480\t35.021\n", + "33.000\t62.500\t44.205\n", + "29.000\t41.380\t38.082\n", + "25.000\t28.480\t31.959\n", + "34.000\t27.140\t45.736\n", + "26.000\t21.650\t33.490\n", + "33.000\t36.670\t44.205\n", + "26.000\t51.580\t33.490\n", + "24.000\t32.610\t30.429\n", + "23.000\t21.670\t28.898\n", + "25.000\t31.430\t31.959\n", + "25.000\t42.150\t31.959\n", + "28.000\t27.030\t36.552\n", + "22.000\t45.450\t27.367\n", + "28.000\t27.500\t36.552\n", + "25.000\t37.040\t31.959\n", + "29.000\t25.640\t38.082\n", + "25.000\t32.350\t31.959\n", + "25.000\t22.730\t31.959\n", + "27.000\t48.840\t35.021\n", + "35.000\t27.910\t47.267\n", + "25.000\t34.620\t31.959\n" + ] + } + ], + "source": [ + "#Print out data for chart. \n", + "model = sm.OLS.from_formula('PctOfTotal ~ IC2013_ACT_Composite_75th_percentile_score', data = niche_adderall_fracs).fit()\n", + "ones_to_plot = niche_adderall_fracs[['IC2013_ACT_Composite_75th_percentile_score', 'PctOfTotal']].dropna()\n", + "ones_to_plot['trendline'] = niche_adderall_fracs['IC2013_ACT_Composite_75th_percentile_score'] * model.params['IC2013_ACT_Composite_75th_percentile_score'] + model.params['Intercept']\n", + "ones_to_plot['trendline'] = ones_to_plot['trendline'] * 100\n", + "print 'ACT\\tAdderall\\tTrendline'\n", + "for i in range(len(ones_to_plot)):\n", + " print '%2.3f\\t%2.3f\\t%2.3f' % (ones_to_plot['IC2013_ACT_Composite_75th_percentile_score'].iloc[i],\n", + " ones_to_plot['PctOfTotal'].iloc[i] * 100, \n", + " ones_to_plot['trendline'].iloc[i])" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "###### \"When I looked at the NSDUH data, I found that college students who had used Adderall non-medically reported significantly higher levels of depression and were more likely to have considered suicide.\"" + ] + }, + { + "cell_type": "code", + "execution_count": 15, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "\n", + "Looking at depression levels in college students\n", + "Category seriously_thought_about_killing_oneself_last_year\n", + "Mean value of ADDERALL for level 0.0 is 0.143; unweighted, 0.136 (3944 values)\n", + "Mean value of ADDERALL for level 1.0 is 0.213; unweighted, 0.191 (366 values)\n", + "Ratio between maximum value and minimum value 1.48723023239\n", + "Optimization terminated successfully.\n", + " Current function value: 0.404718\n", + " Iterations 6\n", + "Optimization terminated successfully.\n", + " Current function value: 0.389079\n", + " Iterations 7\n", + "Optimization terminated successfully.\n", + " Current function value: 0.403017\n", + " Iterations 6\n", + "Optimization terminated successfully.\n", + " Current function value: 0.386725\n", + " Iterations 7\n", + "\n", + "=========================================================================================\n", + " 0 1 2 3 \n", + "-----------------------------------------------------------------------------------------\n", + "seriously_thought_about_killing_oneself_last_year 0.410*** 0.435*** \n", + " (0.141) (0.144) \n", + "worst_K6_score_past_year 0.033*** 0.039*** \n", + " (0.007) (0.007) \n", + "C(IRFSTAMP, Sum)[S.1] -0.059 -0.076 \n", + " (0.089) (0.089) \n", + "C(NEWRACE2, Sum)[S.1] 0.746*** 0.754*** \n", + " (0.186) (0.186) \n", + "C(NEWRACE2, Sum)[S.2] -0.693*** -0.676***\n", + " (0.248) (0.249) \n", + "C(NEWRACE2, Sum)[S.3] 0.152 0.209 \n", + " (0.549) (0.552) \n", + "C(NEWRACE2, Sum)[S.4] -0.499 -0.554 \n", + " (0.892) (0.893) \n", + "C(NEWRACE2, Sum)[S.5] -0.038 -0.056 \n", + " (0.245) (0.245) \n", + "C(NEWRACE2, Sum)[S.6] 0.562** 0.533** \n", + " (0.249) (0.250) \n", + "IRSEX -0.373*** -0.439***\n", + " (0.090) (0.091) \n", + "Intercept -1.852*** -1.781*** -2.058*** -1.948***\n", + " (0.047) (0.236) (0.070) (0.240) \n", + "N 4310 4310 4310 4310 \n", + "=========================================================================================\n", + "Standard errors in parentheses.\n", + "* p<.1, ** p<.05, ***p<.01\n", + "\n", + "\n", + "Looking at depression levels in adults\n", + "Category seriously_thought_about_killing_oneself_last_year\n", + "Mean value of ADDERALL for level 0.0 is 0.026; unweighted, 0.034 (18526 values)\n", + "Mean value of ADDERALL for level 1.0 is 0.067; unweighted, 0.093 (756 values)\n", + "Ratio between maximum value and minimum value 2.54581324356\n", + "Optimization terminated successfully.\n", + " Current function value: 0.153638\n", + " Iterations 7\n", + "Optimization terminated successfully.\n", + " Current function value: 0.148860\n", + " Iterations 8\n", + "Optimization terminated successfully.\n", + " Current function value: 0.147275\n", + " Iterations 8\n", + "Optimization terminated successfully.\n", + " Current function value: 0.142961\n", + " Iterations 8\n", + "\n", + "=========================================================================================\n", + " 0 1 2 3 \n", + "-----------------------------------------------------------------------------------------\n", + "seriously_thought_about_killing_oneself_last_year 1.074*** 0.908*** \n", + " (0.132) (0.135) \n", + "worst_K6_score_past_year 0.098*** 0.097*** \n", + " (0.005) (0.006) \n", + "C(IRFSTAMP, Sum)[S.1] 0.292*** 0.157*** \n", + " (0.046) (0.048) \n", + "C(NEWRACE2, Sum)[S.1] 0.565*** 0.529*** \n", + " (0.111) (0.112) \n", + "C(NEWRACE2, Sum)[S.2] -0.848*** -0.825***\n", + " (0.190) (0.191) \n", + "C(NEWRACE2, Sum)[S.3] 0.490* 0.433 \n", + " (0.263) (0.267) \n", + "C(NEWRACE2, Sum)[S.4] 0.225 0.309 \n", + " (0.446) (0.448) \n", + "C(NEWRACE2, Sum)[S.5] -0.461* -0.432 \n", + " (0.268) (0.270) \n", + "C(NEWRACE2, Sum)[S.6] 0.750*** 0.668*** \n", + " (0.199) (0.201) \n", + "IRSEX -0.262*** -0.415***\n", + " (0.078) (0.080) \n", + "Intercept -3.357*** -3.090*** -3.948*** -3.505***\n", + " (0.041) (0.158) (0.061) (0.163) \n", + "N 19282 19282 19282 19282 \n", + "=========================================================================================\n", + "Standard errors in parentheses.\n", + "* p<.1, ** p<.05, ***p<.01\n" + ] + } + ], + "source": [ + "def lookAtDepressionLevels(nsduh_data, idxs):\n", + " nsduh_data['seriously_thought_about_killing_oneself_last_year'] = 1.* (nsduh_data['MHSUITHK'] == 1)\n", + " nsduh_data['worst_K6_score_past_year'] = 1.*nsduh_data['K6SCMAX']\n", + " data_to_use = nsduh_data.loc[idxs]\n", + " weights = data_to_use['ANALWT_C']\n", + " stratifyByCat(data_to_use, 'seriously_thought_about_killing_oneself_last_year') \n", + "\n", + " model1 = sm.Logit.from_formula('ADDERALL ~ seriously_thought_about_killing_oneself_last_year', \n", + " weights = weights, \n", + " data = data_to_use).fit()\n", + " model2 = sm.Logit.from_formula('ADDERALL ~ seriously_thought_about_killing_oneself_last_year + IRSEX + C(IRFSTAMP, Sum) + C(NEWRACE2, Sum)', \n", + " weights = weights, \n", + " data = data_to_use).fit()\n", + " model3 = sm.Logit.from_formula('ADDERALL ~ worst_K6_score_past_year', \n", + " weights = weights, \n", + " data = data_to_use).fit()\n", + " model4 = sm.Logit.from_formula('ADDERALL ~ worst_K6_score_past_year + IRSEX + C(IRFSTAMP, Sum) + C(NEWRACE2, Sum)', \n", + " weights = weights, \n", + " data = data_to_use).fit()\n", + "\n", + "\n", + " models = [model1, model2, model3, model4]\n", + " print summary_col(models, model_names = range(len(models)), stars = True,\n", + " regressor_order = ['seriously_thought_about_killing_oneself_last_year', 'worst_K6_score_past_year'],\n", + " float_format='%0.3f',\n", + " info_dict={'N':lambda x: \"{0:d}\".format(int(x.nobs))})\n", + "college_idxs = nsduh_data['COLLENR'] == 1\n", + "adult_idxs = nsduh_data['CATAG6'] >= 3\n", + "\n", + "print '\\n\\nLooking at depression levels in college students'\n", + "lookAtDepressionLevels(nsduh_data, college_idxs)\n", + "print '\\n\\nLooking at depression levels in adults'\n", + "lookAtDepressionLevels(nsduh_data, adult_idxs)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "###### \"I found that Google searches for Adderall in college towns spiked during exam months and drop during summer months.\"" + ] + }, + { + "cell_type": "code", + "execution_count": 16, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "0 towns processed out of 2084\n", + "Hamilton, New York not in index\n", + "Farmville, Virginia not in index\n", + "10 towns processed out of 2084\n", + "Canton, New York not in index\n", + "Orono, Maine not in index\n", + "20 towns processed out of 2084\n", + "Menomonie, Wisconsin not in index\n", + "Lexington, Virginia not in index\n", + "30 towns processed out of 2084\n", + "40 towns processed out of 2084\n", + "Geneseo, New York not in index\n", + "Kutztown, Pennsylvania not in index\n", + "50 towns processed out of 2084\n", + "60 towns processed out of 2084\n", + "Frostburg, Maryland not in index\n", + "70 towns processed out of 2084\n", + "town 4355\n", + "ADHD 4355\n", + "ADHD_zero_meaned_by_town 4355\n", + "month 4355\n", + "ADHD_normalized_by_town 4355\n", + "year 4355\n", + "Adderall 4355\n", + "Adderall_normalized_by_town 4355\n", + "Adderall_zero_meaned_by_town 4355\n" + ] + }, + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVMAAAFSCAYAAABPFzzRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkqkxxrjAkqkxxrjAkqkxxrjg/wF8PwJFsLeceAAAAABJRU5ErkJggg==\n", + "text/plain": [ + "" + ] + }, + "metadata": {}, + "output_type": "display_data" + }, + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVMAAAFSCAYAAABPFzzRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xd4VGX2wPHvCah0EQlVULDRpIiggEiwgKy7WNEVkGZZ\nRVAXfzbUFbAB6i5WVqWIoIAuRRQVEAkgAlJCEUEFBelVOgSSOb8/3pswhJSZ5E4m5XyeZx7mlrnv\nmWFy5r33LVdUFWOMMTkTE+0AjDGmILBkaowxPrBkaowxPrBkaowxPrBkaowxPrBkaowxPrBkmseI\nyAci8rz3vKWIrIl2TOZUIhIvInd7zzuJyDSfj3+eiAREJNO/0eA4THRZMs1F3hd/j4icnslu6j1Q\n1bmqWsuHcteLyNU5PU42y+4mInOjUXaEBf8/faSqbaMdh1+8JF7Tz2MWBpZMc4mInAc0BXYA7bPa\n3efiNQLHzDfEE+048hn7vMJkyTT3dAG+AUYDXVNWikgjEVkqIvtFZBxQLGhbnIhsDFo+qcaQ5pJA\neRH5QkT+FJHdIjLHyyGjgerA5yJyQET+L+gUspuI/OHtf7+INBGRFd4x3gwOXkR6iMhPXs36axGp\nniauf4jIL95r3/LW1waGAs28sveE84F57+9t733tF5EFad5/cxFZJCJ7ReQHEWkWtC1eRF4QkXnA\nQaCmF+cDIvKrd7wBInK+iMz3jjFORE7zXl/WK3eH954/F5GqGcSZWvsWkce995ryOC4iI71tZ4rI\ncBHZIiKbROT5lNN4EYkRkVdFZKeIrANuCOOjukBEForIPhGZLCJnececKiK90sS6QkRuzOQzn+M9\nXe7F38Fbf6/3ue0Wkc9EpLK3vr+IvOE9P01EDonIYG+5uIgc9T7LlO9cFxHZ4L3PvmG8x7xPVe2R\nCw9gLdAJuBA4BsQCpwMbgIeBIsCt3rYB3mvigI1BxwgANYOWRwbt+zIucRXxHi2C9vsduDpo+Tzv\nWO94MVwHJAKTgPJAFWA7cJW3/43Ar8DFuB/gp4F5aeKaApQBquFq3229bV2BuWk+i47A8hA+sw+A\nXcBl3nsaA4z1tpUD/vQ+0xjg78Ae4CxvezywHqjtbT/Ni3MSUAqo473nb73PowywCugSdPybcT9u\npYBPgElBsc0CenjPu6V9j976c4DNQZ/FJO//qLj3/78QuM/bdj+wGqgKnOUdPxmIyeIzigc2ee+n\nBPA/YLS3rQOwIGjfBt7nWTSLY6b9nl0N7AQaet+XN4DZ3rbWwArveXPc93xB0OsS0nzn3gXOAOoD\nR4Fa0f7b9O1vPNoBFIYHcCVwBCjtLS8DHgGuAjan2Xce2Uum/YHJwPnplJ9RMq0ctG4X0CFo+X/A\nQ97zr1ISh7ccAxwCqgXF1Txo+3jgCe95uokmxM9tJPBe0HI7YLX3/K7gROGt+x7o6j2fBfRLsz0A\nNAtaXgw8FrT8KvCfDGJpCOwJWs40meIS5pKU4wMVveRRLGifO4Fvveff4iVWb/k6L96skuks4KWg\n5dq4HwnB/RDsSflOeO/vrRA+97Tfs+HAwKDlkrgf/ere+zyC+/F5AngK2Ojt0x8YkuY7VyXoOAuB\nO6L5t+nnw07zc0dXYLqqHvCWP/XWVcbVXIJtCPPYKde2XsHVCqaLyDoReSKE124Pen4kneVS3vNz\ngde9U/g/gd3e+uDT3m1Bzw/j/pj8kFFMVYA/0uy7wVufYiOnCuk9i0gJEXlXXOPdPmA2cKZIyNde\nh+MS/yve8rm42vHWoM/xv7gaKrjvQnC8ad9bZtK+7jSgvKoexdWo7/Li/jvuMlO4KhP0vVTVQ7jv\nQFVVPYL7UWqFqxzMxv2otQhaDhap70nUFY12AAWdiBQHbgdiRGSrt/oM4ExgKycnJHB/dGszONxh\n3KlcitQ/QFU9CPwf8H8iUhf4VkR+UNVZ5Ly19w/geVUdm43XRmpass3ALWnWnYurRftR9qPARUBT\nVd0hIg2Bpbgfr0yPKyJPAhcALYNWb8TVGM9W1UA6L9uKq+mlqJ7OPhlJ+7rjuDMNgFHAh7gznsOq\nujCM46bYgqtZAiAiJYGzOVERmA1cAzQCFnnL1+MaXOdQSFjNNPJuApJwp18NvEdt4DvcNbkkEXnI\nu3h/C9Akk2MtAzqJSBERuR73yw+AiPxVRC7waiD7cdfbUv5otwPnZyP2lFrYf4G+IlLHK+vMlIaJ\nTF6X8trtwDkpDTvZLD89XwEXicidIlJURO4AagFfhPj69PYJfl4KV1PdJyLlgOdCClikHdAbuEVV\nE1PWq+pWYDrwbxEp7TU4nS8iKf+HnwAPiUhVrwHpyVDK82LuLCK1RaQEMAD4VL3zaFWdj0v+r+KS\naijSfl/GAt1FpIGInAG8hLvEklJ7no1rYF2lqsdx13HvAX5T1d1krsD0GrBkGnldgBGquklVd3iP\n7cBbwB24hNoNd9p0OzAhk2M9DPwN1/DSEdegkeICYAZwAHea9baqppxivQw8451e9vHWhVJrS/mD\nnAwMAsZ5p7wrgbZp90uznLJuJq5hZ5uI7IDUTu4/hlh+esfG+yP9K64GuQtXK/+rqu5Ju28my2nX\nBZc3BHc9cBfu8/wqg9enfd3tuEa81UEt+u9427rgGnB+wl3L/BSo5G17H5gGLMedNk/IpLy0ZX+I\na6zb6h3/oTT7fAhcgmvAC0U/YJT3fblNVWcCz3oxbQFq4C4ZpJiPuz6bUgtdjfshSlsrzerzz9fE\n+wGLTuEiI3BdQHao6iUZ7PMGruHhMNBNVRNyMcSoEtfR/n1VzU6t0hgAROQu4F5VvSrLnU22Rbtm\nOhJ3bSVdIvIX4AJVvRC4D9etpDCpB/wW7SBM/uWd+j8IvBftWAq6qCZTVZ2LO2XNSHvcBXS8C+dl\nRaRibsQWbSLyOu60vn+0Y4kkEVmVppN7yuPOaMeWV4jIwQw+oxZZvK4trs/vVuDjoPUtMzje/gi/\nlQItr7fmV+Xkbh+bcB2ht6e/e8Ghqg/jkmmBpqp1ox1DXqeqpbLeK93XTeNEV7Lg9XOB0jmNy5ws\nrydTOLW175SLvCJSYC5iG2PyFlUNqcdBtK+ZZmUzbnhiipTheaeI9OiG5557rkCUYe+l8JZRkN5L\nbn1e4cjryXQKrjsJInIFsFddtyJjjMlTonqaLyJjccPQyoubHek53FA4VPVdVf1SRP4iImtxY8G7\nRy9aY4zJWFSTqapm2WKrqr2y2ic3xMXFFYgycqscey95r4zcKqeglBGuqHba94uIaEF4H8aYvEVE\n0ALSAGWMMfmCJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPGB\nJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPGBJVNj\njPGBJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPGB\nJVNjjPGBJVNjjPGBJVNjjPGBJVNjjPFBWMlURMqJSP1IBWOMMflVlslURGaLSBkRKQcsAYaJyH8i\nH5oxxuQfodRMz1TV/cAtwIeq2hS4NrJhGWNM/hJKMi0iIpWB24Gp3jqNXEjGGJP/hJJMBwDTgHWq\n+oOInA/8GtmwjDEmfxHV/F/JFBEtCO/DGJO3iAiqKqHsWzSEg1UA7gXOC9pfVbVHtiM0xpgCJstk\nCnwGzAFmAAFvnVUDjTEmSJan+SKyTFUb5lI82WKn+caYSAjnND+UBqgvROSGHMZkjDEFWig104NA\nCeAYcNxbrapaJsKxhcxqpsaYSPC1AUpVS+U8JGOMKdhCGU46RkTuFZFauRGQMcbkR6Gc5l8NtASu\nBC4AlgJzVXVI5MMLjZ3mG2MiIZzT/JA67YtIUeAy4GrgfuCIql6coyh9ZMnUGBMJfnfanwmUBOYD\n3wGXqeqOnIVojDEFSyhdo1bgWvHrAfWBeiJSPKJRGWNMPhPy2HwRKQ10A/4PqKSqZ0QwrrDYab4x\nJhL8Ps3vjWuAagz8DowA5uYoQmOMKWBCGZtfDHgNWKKqSRGOxxhj8qVQW/Mb4mqniusWtTzSgYXD\nTvONMZHg69h8EXkYGAPEAhWBMSLyUM5CNMaYgiWUTvsrgStU9ZC3XBJYoKqX5EJ8IbGaqTEmEvye\nNQpOzGOa9rkxxhgyaYASkQ9UtRswElgoIhMBAW7CtegbY4zxZHiaLyIJqtrIe94YaOFtmquqCbkU\nX0jsNN8YEwm+jM0XkTVAx+BV3r8KoKpLcxKknyyZGmMiwa9kegBYnNELVbV19sLznyVTY0wk+DUC\nam2kE6aIXA8MAYoAw1R1UJrtcbgb+v3mrZqgqi9EMiZjjMmOUEZARYSIFAHeAq4FNgOLRGSKqq5O\ns+tsVW2f6wEaY0wYMusa9WSEy26Kq/2uV9XjwDjgxnT2C6mKbYwx0ZRhMlXVaREuuyqwMWh5k7fu\npDCA5iKyXES+FJE6EY7JGGOyJWqn+Xi9ArKwFKimqodFpB0wGbgovR379euX+jwuLo64uDgfQjTG\nFCbx8fHEx8dn67XhzGdaQlUPZ6uU9I93BdBPVa/3lp8CAmkbodK85negsaruSbPeWvONMb7ze6KT\n5iLyE/Czt9xQRN7JYYzgul1dKCLnicjpwB3AlDRlVxQR8Z43xSX/PaceyhhjoiuU0/whwPW4Lkqo\n6jIRaZXTglU1SUR6AdNwXaOGq+pqEfmHt/1d4DbgARFJAg4Df89pucYYEwmhzBr1g6o2TTO8dLmq\nNsiVCENgp/nGmEjw9bYlwB8i0sI78OnAQ0DavqDGGFOohTIF3wPAg7huS5uBRt6yMcYYT8it+XmZ\nneYbYyLB77uTvonrEyqc6Bu6H1ikqp9lO0pjjClAQjnNLwY0BH4B1gINgHOAu0VkSARjM8aYfCOU\n1vyFQIuU2zyLSFHgO+BKYKWq1o54lFmw03xjTCT4fQ+oskCpoOVSQDkvuR7NRnzGGFPghNI1ajCQ\nICKzveVWwEveXUq/iVhkxhiTj4TUmi8iVYAm3uIiVd0S0ajCZKf5xphI8HtsfgxwDdDAa70v6o2T\nN8YY4wnlmuk7QDPgTm/5oLfOGGOMJ5RrpperaiMRSQBQ1T0iclqE4zLGmHwllJrpMe9+TQCISCwQ\niFxIxhiT/4SSTN8EJgEVROQlYB7wckSjMsaYfCbU1vzauEYogJnp3EE0qqw13xgTCeG05meYTEWk\nXNpV3r8K7tpptiP0mSVTY0wk+DXRyVIyv+ldjbCiMsaYAsym4DPGmAz4UjMVkUsze6GqLg03MGOM\nKagyu2YaTyan+araOkIxhc1qpsaYSPClASo/sWRqjIkEv2faPx13H6irvFXxwH9V9Xi2IzTGmAIm\nlMmhh+OS7ihc96i7gCRVvSfy4YXGaqbGmEjw9TRfRFaoav2s1kWTJVNjTCT4PdN+kohcEHTw84Gk\n7AZnjDEFUSizRj0GfCsiv3vL5wHdIxaRMcbkQ6GOzS8GXIzrKvWzqiZGOrBw2Gm+MSYS/Bqbf5e3\n/cN01ier6sc5jtQnlkyNMZHgVzL9AbhGVQ+kWV8KmKOqmY6Qyk2WTI0xkeBXA9RpaRMpgKoeBGym\nfWOMCZJZMi3m1UJPIiKlsWRqjElHQAMcPn442mFERWat+cOBT0XkAVVdDyAiNYC3vW3GGHOSzhM7\n8+lPn3JBuQtoVKkRDSs1TP03tmRstMOLqAyTqaq+KiIHgdlebRTcnUlfVtWhuRKdMSbf+Oa3b5i/\naT67HtvF+r3rSdiWwLJty5j661SWbVtG6dNL06hyIxpWbOj+rdSQGmVrIBLSJck8L9SuUWUAVHV/\nxCPKBmuAMia6jiUfo8F/GzDwmoHcWOvGU7arKr/v/Z1l25aRsDUhNdEePHaQhpUanlSDrRNbh9OK\n5I0riTZrlDEmV736/avM/H0mX3b8Mqya5s5DO12C9ZJrwrYENuzdQK3ytWhUqVFqDbZBxQaUPqN0\n1gf0mSVTY0yu2XJgC/WH1mf+3fM5t/SFbNgAxYtDsWInHkVDGWvpOXTsECt3rDypFrtq5yqqlK6S\nWnttfV5rmlVrFrk35bFkajL0ww8QEwONG0MBuVRloqzjhI7UKFuDF695kTvugLlz3Xfs6FH3OHLE\nfdfSJthixUJfd3qxJPaf/jM7Y5axhQQWHf2Id/42hDvq3RHR9+brfKbeAVvgxuSn7K9pR0aZvG/0\naHjsMShTBhIT4ZZb3KN5cyhSJNrRmfwofn088zbO4/2/vc/ixfDdd7B2LZQocfJ+SUkuqaYk2OBE\nm3ZdeusP7i9K4pG6nH60LpWOdqLkqq50P3IdRQIluK3+36Lz5tMIZQq+MUBNYBmQnLJeVXtHNrTQ\nWc00a++8Ay+/DNOnQ61asGoVTJzoHtu2wU03ucTaujWcljeu/Zs87njycRq924j+cf25tc6tXHcd\n3Hor3H9/5Ms+eBA6PPIDM2Jv4L024+jR+pqIlOP3fKargTp5OVtZMs3coEHw7rvwzTdQs+ap29eu\nhUmTXGL95Rf4619dYm3Txp1yGZOeIQuGMPXXqUzvPJ2ZM4UHHoCffsq9H2NVeHLobF754zb61pjC\nC//w/xpqOMkUVc30AXwKVMlqv2g+3NswaQUCqn37qtaurbppU2iv2bhR9c03VVu3Vi1TRrVDB9Wx\nY1X37YtsrCZ/2Xpgq5YfXF5X71ytgYBq48aq48ZFJ5ahM77UIk9W0L/ek6AHD/p7bC+3hJSHQqmZ\nxgMNgR+AlKn3VFXbZyvVR4DVTE8VCMA//+kaA6ZNg9hsDD7ZuROmTHE11rlzoVUrV2Nt3x7OPtv/\nmE3+0WVSFyqXqsyg6wbxv//BSy/B4sWu4Skaxiz9H/dN6k2labOYMqIW9er5c1y/T/Pj0luvqvFh\nRxYhlkxPlpwM997rTtm/+ALKlj2xbeO+jVQqVSnsTtH79sHUqS6xzpgBTZq4xHrTTVClis9vwORp\n3/3xHXdOuJPVD66mWEwp6taFN990l4WiadSyUfT54hl0xBxeeboGPXrkvMeKdY0qxI4dg86d4c8/\nYfJkKFnS9dv79KdPGZ4wnMVbFtPm/Db8r8P/sj3K5PBhV9udONEl2Fq1XMPDzTenf03WFBxJgSQa\nv9eYvlf25Y56d/D++zB2LMycmTe62r39w9sMmvNvio+bS5NaVRg6FErnoK+/L1Pwicg879+DInIg\nzSNPDist7I4ccTXFY8dgyhRl9b7F3P/F/VT7TzUmrJ7Ao80eZffjuzmefJweU3oQ0EC2yilRwiXO\n0aNdT4B//Qt+/hmuuAIaNYLnn4eNG31+cyZP+O/i/3J28bO5ve7tHDkC/fvDwIF5I5ECPNj0QXpe\ncS8x3a5FS+zksstg+fLcKdtqpgXEgQPwt79B+Wp7uPKBj/hgxXD2J+7n7kZ3061hN6qWqZq67+Hj\nh2k7pi2NKjXi9etf922iieRk18/www9hxQpYtMiXw5o8YsehHdR9py7xXeOpW6EugwfDwoUwYUK0\nIztV35l9+Xrt1zxQfBZ9Hz2TF16A++4LP+nbaX4hs2t3gJadZ5PUYBg7y06l3YXtuKfRPbSu0ZoY\nSf/kY+/RvbQe1Zr2F7Wnf+v+vsaTnAzVq7trq3Xq+HpoE0U9PuvBWcXO4rW2r/Hnn3DRRTBnDtSu\nHe3ITqWqPPz1wyzdupQ3L59G1ztLUqcOvPeeG7QSKr9v9WzyqC0HttD3q5epOvAidjV5iF43Xc66\nh9Yx9taxXFPzmgwTKUDZYmWZ1nka41aNY8iCIb7GVaQIdOrkaqimYJi/cT7T1k3jubjnABg8GG68\nMW8mUnBJcMj1Q7jw7At5fOlNzJ53lDPPdMOoExIiVGiofajy8oNC1M/0ePJx/WzNZ/q3j/+mZV4q\nq2U63av39V+oycmBbB1vw94NWv0/1XVkwkhf41y5UrVqVdWkJF8Pa6IgKTlJL333Uh2zfIyqqm7e\nrFqunOoff0Q5sBAcTz6ut31ym9449kY9lnRMx45VLV9e9e23XT/srBBGP9OQaqYicp6IXOs9L5Ey\nv6nJPWv3rKXvzL5U/091Bs0bRLOzbqLMsI30v+w93v1XU2Jisnfds/qZ1ZnWeRpPzXyKSasn+RZv\nvXpQsSJ8+61vhzRR8v7S9yl5Wkk6XtIRcA2MPXpAtWpRDiwERWOK8tEtH3E8cJxun3Wjw+3JfP89\nvP8+3H676/Lnm6yyLXAfsAhY5y1fBMwMNVvnxoMCWjM9fOywjlk+RuM+iNPYwbHa5+s+umrHKl2x\nQrVKFdVhw/wra8mWJRo7OFZnrJvh2zFff121UyffDmeiYOehnRo7OFaXb1uuqqq//KJ69tmqu3ZF\nObAwHT52WFuNbKX3TrlXA4GAHjmi2rOnas2aqosWZfw6wqiZhpKolgNnAAlB61aGWkBuPApaMl22\ndZn2mtpLyw0qp21Gt9FPfvxEjx4/qqqqCxeqVqwYmaF7s9fP1vKDy+uCjQt8Od6OHapnnqm6f78v\nhzNRcO+Ue/WhLx9KXb7jDtUXXohiQDmw/+h+bfp+U+3zdR8NeOf4n3ziTvtffz390/5wkmkoI6B+\nUNWmIpKgqo1EpCiwVFXr+1hBzpG83JqvqiQmJ3I06Wi6jyPHj6Q+37R/Ex+u+JDtB7fTo1EPujfs\nzrllz0091uzZ0KEDjBjhJiOJhKm/TOXuKXfzTZdvqFch52PybrzR9X3t3t2H4EyuWrR5Ee3HtWf1\ng6spW6wsS5e6792vv7rBIPnRniN7aD2qNbfUuiW1MW3dOnfKf+65MHw4nHXWif39Hk76CrAX6AL0\nAnoCP6nq09l6NxEQ6WS6fNtyRi0fdXISTDqSYYIMTpSJyYmcUeQMihUtluGj+GnFKVa0GGcVO4vb\n697OdTWvo0jMyROMfvkldOsG48e7afIiaezKsTw24zHmdJ9DzbNyNqRp4kR44w2Ij/cnNpM7Ahrg\nimFX8GCTB+nasCsAbdu6H8eePaMcXA5tP7idliNbcv9l99OnWR/Aze/72GPw+efub6xpU7ev38k0\nBrgHSBl5Ow0YlpeqgpFMptsObqPxf5sQd1ZXLr2oCmeVSpMIixbPNFGeUfSMTLsoheLTT6FXL/js\nMzfKKDcMXTSUV+e/ytzuc6lSOvuD7xMToWpVNwnGeef5F5+JrGFLhzFy2Ujmdp9LjMTw7beu0/vq\n1QVjvts/9v3BVSOvom/LvtzX+L7U9RMnuvlYn3oKHnkEYmL8nYLv4VDWRfNBhK6ZHks6pi1HXKVV\nO/9LL7lEtWRJ1YYNVXv3Vh0/3nURibQRI1QrV1ZdtizyZaX14pwXtd479XT34d05Ok7PnqoDBvgU\nlIm43Yd3a4VXKujSLUtV1V1LbNpU9eOPoxyYz37d/atWea2KfrTio5PW//abapMmqu3b+98AlZDO\numWhFpAbj0gl00e+ekQv7PcXvebaZE1OVk1MVP3+e9XBg90HXa6cao0aqnfdpfree6o//RRa37VQ\nvfGGarVqqmvW+HfMcAQCAf2/af+nl79/uR5IPJDt4yxcqHrBBf5+NiZyHvjiAe35Rc/U5QkTXCUi\nOTmKQUXIyu0rteIrFXXy6sknrU9MVH3qKZ+SKXAn8DnueunnQY94CkHXqI9XfKxVB9fUs8/Zoxs3\npr9PcrLqqlWq777rEmqNGq7bSPv2qq+8ojp/vvtPCVcgoPrii6rnn6/6++85ehs5FggE9J7P7tFr\nRl2T2qMg/GOoXnyx6rx5PgdnfLdkyxKt+EpF3XN4j6qqHj+uWquW6ldfRTmwCFq0eZHGDo7V6Wun\nn7LNr2R6LhAHLABaec/jgMZA0VALyI2H38l0xbYVWn5QeT338mVhd0HatMl1W+rVy/2alyyp2qqV\n6jPPqH79ddYz1gcCqk88oVq3ruqWLdl+C75KSk7SDp900JvH3azHk49n6xgvvaR6330+B2Z8lRxI\n1iuGXaHDlpzowDxsmPv+FvSzijnr52j5weX1uw3fnbQ+nGRqE52ksffoXi577zJqrO9PxR2dGDMm\nZ8fbtw/mz3ezKX33nWuIufBCaNkSrrzSPVImVw4EoHdvdzvmr7/OW7PZJyYl0n5ce6qUrsLw9sPD\nblTbuBEaNIAtW9yte03e88GyDxi6eCjz755PjMRw5IibzOTTT3Ov4TOapq2dxl2T7uLrzl9zaeVL\nAf9b85sBbwC1cZ33iwAHVTXPDCn1K5kGNMCN424kZl9Nlr38OsuXnzxLvR+OHYMlS04k1+++gzPP\ndEn1wAHYvdvNjh/OzDa55dCxQ7QZ04amVZry77b/DnvqvmuvdS3Ct98eoQBNtu09upfab9fm8zs/\n57IqlwHw2mvu+znJv1HGed7E1RPpObUn33b9ljqxdXy/od4S4EIgAZdIuwMDQ636ZnHs64E1wK/A\nExns84a3fTnQKIN9clrLV1XV/vH99fJ3r9TKVY/prFm+HDJLwdddn31W9dCh3Ck3u/Yc3qP1h9bX\nAfHhN8+PGqV6ww0RCMrkWO8ve+t9U05ch9m7VzU21n03C5sPl32oVV+rquv2rPO9NX+J9++KoHU5\nbs33EvNa4DzgNGAZUDvNPn8BvvSeXw4syOBYOf4Ap/4yVau+VlXbddiijz6a48MVaFsPbNUL3rhA\n31jwRlivO3BAtWxZ1W3bIhSYyZZlW5dphVcq6K5DJwbcP/20avfuUQwqyt754R2tMaRGWMm0aAiV\n10MicgawXEQGA9sAP6ZmbwqsVdX1ACIyDrgRWB20T3tglJctF4pIWRGpqKrbfSg/1bo96+j+WXfu\nKTWJz9dUZtJoP49e8FQqVYkZd83gqpFXUbZYWe5qcFdIrytVyo2g+egj6NMnwkGakKgqvb7qxYC4\nAZxdwl2k37oVhg6N4Lyf+cADTR6g5Okl6UrXkF8TSitCF2+/XsBh4Bzg1mxFeLKqQPCdgjZ567La\n5xwfyk51+PhhbvnkFnrW/RfvPducjz6CM87ws4SC6byy5/F15695bMZjTPl5Ssiv69LFJo3OSz5a\n+RFHjh/6aQ+rAAAgAElEQVThnkvvSV33wgvQtau7W0Jh1qVBl7D2z7JmmlJzBI4A/USkFPAgMCjc\n4NIeOsT90taC031dv379Up/HxcURFxeXdQCq3Pv5vdSv0IDpL/bkySfhkktCjMpQJ7YOn9/5OTd8\nfAPjbxtP6xpZTxoQFwd79ribnDVoEPkYTcb2J+7niW+eYMLtE1Lngli3zo1NX7MmysFFSXx8PPHZ\nnEgiw9Z8EakCPAWcD/wIDADuBR4FJqrqQ9kq8cTxrwD6qer13vJTQEBVBwXt818gXlXHectrgFZp\nT/Oz25r/xsI3GLlsJDftnsfsb0rwzTcQYzdyCVv8+nhu//R2pnacSpOqTbLc/+mn4ehR11psoqfP\ntD7sO7qP4TcOT13XsaO7Fcmzz0YxsDzEl9Z84BugH67FfQiwHhgHVAr1gmxmD1yteB2uAep0sm6A\nugIfG6DmrJ+jFV6poFPm/qaxsaobNoR9CBPkszWfacVXKuqqHVk3/65Z4+ZkPZ69/v/GByu3r9TY\nwbG64+CO1HUJCaqVKrmGQuPg0wioZWmWNwFFQj1wSIVDO+BnXKv+U966fwD/CNrnLW/7cuDSDI4T\n1ge0ef9mrfJaFZ3841dau7bqmDFhvdxkYPTy0XrOv8/R3//8Pct9r7hCderUyMdkThUIBLTVyFb6\n1sK3Tlrfrp3qm29GKag8Kpxkmtlp/grc8FFw1y1nBS2jqntCqvrmgnBO848lH6P1qNa0u6AdOyc8\nw/btMHZs+PfTNul764e3eH3h68ztPpdKpSpluN/QoW6O0/Hjcy8244z7cRyD5g1i8b2LU6+Vzp7t\nJvBeswZOPz3KAeYhvoyAEpH1ZNxIpKqas1mDfRROMu31ZS827t9Iz3KTuOfuGFasOHlmbZNzz89+\nnvGrxvNNl28yTKh79kCNGrB+vX3+uelA4gFqv12b8beNp0X1FgCoQvPmbs7cTp2iHGAeE04yzbA1\nX1XP8y2iPOLD5R8yfd10pt+2iCubxDBqlP0hR8KzrVzrRasPWvFtl2+pWiZtjzcoVw6uu86N+77v\nvlM2mwh5fs7zXFPzmtRECm7S8cOH4c47oxhYQRDq9YC8/CCEa6ZLtyzV8oPL64ptK7VDB9VHHsny\nJSaHBn83WGu+XlPX/7k+3e1Tpqg2b57LQRViP+34ScsPLq/bDpwYgpaUpFq7tl2/zghhXDMtFB2B\n9hzZw62f3Mrbf3mb5TPqsWoVvPRStKMq+B5r8RiPXP4IrT5oxdo9a0/Zfv31sHate5jI2nZwG/d8\nfg/PtHyGiqUqpq4fPRrKl4d27aIYXAFR4JNpciCZjhM6ckvtW7i81O306eOGMxYvHu3ICofel/em\nb8u+tB7VmjW7Tu4Jftpp7tTSRkRFzvHk4/x7/r+p9049rqx2JQ82fTB129Gj8NxzMHCgNcD6IZSx\n+YhIEaBi8P6q+kekgvJTv/h+JCYn8mLrgbS9Dh59FBo2jHZUhct9je/jjCJncPWoq5nWeRqXVDwx\nzKxrV7j5ZujXzwZM+O2b377hoa8eovqZ1ZnXYx4Xl7/4pO1Dh7q/hebNoxRgQZPVdQCgN7AL+AlY\nmfII9TpCbjzI4Jrp5NWTtdq/q+n2g9t18GDVli3dNSITHWNXjtWKr1TUJVuWpK4LBFTr1VONj49i\nYAXM+j/X663jb9UaQ2ro5NWTNZDONPn79qlWqKC6cmUUAsxH8HkKvnXA2aEeMBqP9JLpz7t+1tjB\nsbpg4wJdtky1fPno30/JqE78aaJWeKWCLti4IHXdK68U7une/HLk+BEdED9Ayw0qp/3j++vhY4cz\n3PfZZ1W7ds292PIrv5PpLOC0UA8YjUfaZHog8YDWebuOvrv4XT1yxN1P6YMPcvSZGh998fMXGjs4\nVudumKuq7l5XZcvm/Ymx86pAIKCfrflMawypobeOvzXD3hMptm1zd9a1ykXWwkmmmXXaf9R7Wgeo\nBXwBHDtxdUD/7eflhpwI7rSvqvx9wt8pdVophrUfxv/9n7Bhg+vPaBfZ844Z62bQcWJHxt82nqtr\nXE27dtC5s3UaD9cvu3/h4a8fZv3e9bxx/Rtcd/51Wb6md28oUgSGDMmFAPM5v0ZA9ePECCghzWgo\nVe2fgxh9FZxMX/v+NcatGsfc7nP5fk4xunRx073lpZvTGWf2+tnc9ultjL55NHsXX8+IETB9erSj\nyh8OHjvIC3NeYNjSYTx15VP0vrw3pxfJehzob79BkyZu2GhsbC4Ems/5eg+o/PDAO83/9rdvteIr\nFXX9n+t1zx7VatXc7ZVN3jXvj3kaOzhWxy+frGed5W6VbTIWCAT0oxUfadXXqupdE+/SLfvDux94\n586q/fpFKLgCCD9v9SwiM4AOqrrXWy4HjFXVtjlK+T4SEf1j7x80HdaU0TeP5tqa19Kxo6uNvvlm\ntKMzWVm8ZTE3fHwD9f54izZVO/DEE9GOKG9asX0Fvb/qzf7E/bzV7q2ThoSG9PoV0KYN/PorlC4d\noSALGL/vTnrKzfPSWxfNB6BN32+qA+cOVFXVjz9WrVXLGjTyk2Vbl2m5lyppletHazo9eQq1PYf3\naK+pvbTCKxV06KKhmpQcfv++QMBNsff66xEIsADD5+GkySJyblCmPg8IhJffI++cMufweIvH2bgR\nHn4YxoyBEiWiHZUJVYNKDZhz90y2X/Ik/5o0POsXFAIBDTBs6TBqv12bpEASP/X8ifsvuz912rxw\njBoFGzfCP/4RgUANENoIqKeBuSIyG9cQdRWQ5+b5GXnjSFSFbt1cMm3cONoRmXDVrVCHniVnMWTZ\nNVSulkjPJj2jHVLULNy0kF5f9eL0IqfzVaevaFS5UbaP9dtv8NhjMHOm3SwykjJNpiISA5wJNMbd\nNkSBf6rqzlyILSxlzijDf/7jxhvbNbf865G7LmTMtbMZXO5qEpMS+Wezf0Y7pFy1/eB2npr5FNPW\nTWPgNQPpXL8zkoM+fUlJ7o6wTz0F9ev7GKg5RabJVFUDIvK4qo4HPs+lmLLlxx/dTFALF0LRkGYc\nMHlRzZpQ75wadKkyh0GLr+ZI0hH6tuwb7bAiLimQxNs/vM0Lc1+ga4OurH5wNWXOKJPj4w4a5Gqj\njzziQ5AmU6GknRki8n/AeOBQykrNQ7ctAdfZe9Ag98do8rcuXeDLcdWYM2oO13x4DUeTjtI/rn+O\namh5VUADfLbmM/4V/y8qlarEnG5zqB1b25djL14Mr78OS5faJDK5IZSuUetJ5/YlqlojQjGFTUT0\nppuUiRNtlFNBsG8fnHuuu4d7crEdXPvhtbS7oB0Drx1YYBJqYlIiH638iMHzBlPmjDI83fJp2l/c\n3rf3d+gQXHopDBgAd9zhyyELJV9GQOUnIqI7dqiN6ChAOnaEFi3gwQdh9+HdtB3TlhbVWjDk+iH5\nOqHuT9zPe0veY8iCIdSrUI8nWjxB3Hlxvr+nnj3hwAE3+bPJPt+TqYjUw43RL5ayTlXzzJS+4dxQ\nz+QPX38N//oX/PCDW957dC/tPmpH/Qr1GfrXocRI/jpv3X5wO28sfIN3l7zLdedfx+PNH89RC31m\npk51P0LLl8OZZ0akiELD12TqjdFvBdQFpuLudf+dqt6Wwzh9Y8m04ElKgurVXXee2t4lxAOJB/jr\n2L9So2wNhrcfnq3+lrnttz9/49XvX2Xcj+P4e72/82izRzm/3PkRK2/HDjfh89ix0KpVxIopNMJJ\npqH8vN8GXAtsVdXuQAOgbA7iMyZLRYu6RsXgW5qUPqM0X3X6is0HNtN5UmeOJx+PXoBZSNiawN//\n93eavt+Us4qdxeoHV/PODe9ENJGqwr33ugY8S6S5L5Sa6SJVbSIiS4Crgf3AGlW9ONMX5iKrmRZM\nP/7obrq3YYObMi7F0aSj3PrJrcRIDA9f/jCXV72c0mdEf7C5qjJr/SwGzRvEjzt+pM8Vfbiv8X25\nFtv778M777jugadnPYGUCYHfp/nv4EZB3QE8iuseleDVUvMES6YFV+PGrsvbtdeevD4xKZGX5r7E\nt+u/JWFrAueXO59m5zSj2TnNaF6tOReUuyDXGqqSA8lMXjOZQfMGsT9xP4+3eJxOl3TijKK5N9zo\n11+hWTOYMwfq1Mm1Ygu8iLXmi0gNoLSqrshucJFgybTgev11118ys1bpY8nHWLZtGfM3zmf+pvl8\nv/F7jiQd4YpzrqD5Oc1pVq0ZTao0oeTpJX2NLTEpkdErRvPK969QtlhZnmzxJDfWujHXG8eOH4cr\nr3STa/funatFF3h+10xjgE5ADVUdICLVgUqq+kPOQ/WHJdOCa8cOuOgiN0lHONPGbd6/mfmb5jN/\n43y+3/Q9K7av4OKzL6Z5teauBlutGTXK1shW7XV/4n7eXfwuQxYOoX7F+jzR4glandsqal22+vWD\nBQvgyy+tc77f/E6m/8XNEnW1qtby5jOdrqqX5TxUf1gyLdjat4dbboFu3bJ/jMSkRJZuXZpac52/\naT5JgaQTyfWcZlxW5TKKn1Y8w2NsO7iN1xe8zvtL36fN+W14vMXjNKwU3fuGz58PN90ECQlQpUpU\nQymQ/E6mCaraKOVfb91yVW3gQ6y+sGRasE2YAG+9BbNm+XdMVWXj/o0usXqXB1btXEXd2LqpNdfm\n1ZpTrUw11v25jle/f5Xxq8bTsV5HHm3+KDXPiv645YMHXTeowYPdj43xn9/JdCHQHFjsJdVYXM00\nMj2Os8GSacGWmAhVq8KSJW6YaaQcOX6EJVuXpNZc52+cj4hwPPk49192Pw9d/hAVSlaIXABhuvde\nSE6GESOiHUnB5Xcy7QzcjpuGbxSu3+kzqvpJTgP1iyXTgq9nT3ca+8wzuVemqrJh3wbOLn52nuh6\nFWzyZHj0UVi2zG5BEkmRGE5aG7jGW5ypqqtzEJ/vLJkWfAsXwl13wc8/22Q227a50/uJE6F582hH\nU7D5MgJKREqKyOkAXvL8Bjgd8Gd+MGPC0LSpS6ILFkQ7kuhShR493Cm+JdK8JbOOFF8D5wKIyAXA\nfKAG8KCIDMyF2IxJJQJdu7p7GRVmQ4fCzp1uEhiTt2R4mi8iK1X1Eu/580A5VX3Qq60uVdV6uRhn\npuw0v3D44w9o1Ag2b4ZixbLev6BZvRpatoR58+DiPDOYu2Dza6KT4Ox0De40H1U9Rh68O6kp+KpX\nd9cKP8/TN9CJjGPH3AinF16wRJpXZZZMV4rIqyLSBzgfmA4gImeRzsz7xuSGrl1PnkmqsOjfHypX\ntls152WZneaXAB4GKgEjVHW5t745cL6q5pk5vO00v/A4eBDOOce16lesGO1ocsfcuXD77a4bVGF5\nz3lFobxtSUF4HyY0Xbu60/1/FoK7QO/fDw0awBtvwN/+Fu1oCh+/J4c2Jk/p0qXwnOo/9BC0aWOJ\nND+wO8ybfKd1a9i9G1asgPr1ox1N5Hz6KXz/vZvExOR9VjM1+U5MjBsNVZBrp5s3Q69eMGYMlPR3\nGlYTIZk1QAV3QFEg+LqBqmr7SAYWDrtmWvj8/LO7z9HMmVC3brSj8VcgAG3buj6l1jk/uvy6Zvqa\n9/gNOAK8B7wPHPTWGRM1F18MTz/tTvnvugvWrYt2RP55803Xa6Fv32hHYsIRyqxRS1S1cVbroslq\npoXX/v3wn/+41u7bboNnn3Vdp/KrH390PxALFsD5kbuRqQmR3635JUQk9b9VRGoCJbIbnDF+KlMG\nnnsOfvkFypZ1DVJ9+rjbneQ3iYnu9tYDB1oizY9CSab/BGaJyGwRmQ3MAh6JbFjGhOfss91dTFet\nckMva9d2c5/u3RvtyEL3zDNQs6abFcrkP6HOZ1oMSBkRvEZVEyMaVZjsNN+ktX49DBjgxvH36eP6\na+blVvFZs9zY++XLoXz5aEdjUvh6mi8iJYHHgF7ekNLqIvLXHMZoTESdd567ncfcuW4Y5gUXuNtG\nHz0a7chOlpTkZoHq1g2GD7dEmp+Fcpo/EjiGuw8UwBbgxYhFZIyPatWC8ePhq6/gm2/cbaOHDXNJ\nLFo2bID33oNbb4XYWNef9LHH4PrroxeTybmQW/Pt7qSmIJg/33Wp2rjRXQa4447I32v+0CGIj4fp\n02HaNNizxw0RbdsWrrsOKlWKbPkm+/y+od73uPlMv/fuTno+MFZVm+Y8VH9YMjXhmjnTJdXDh+H5\n56F9e//uLaXqhrpOm+YeP/wAjRu75Nm2rZukJdIJ3PjD72TaBngaqAPMAFoA3VTVx7uY54wlU5Md\nqvDFFy6pFi/uJl6+9trsJdWdO2HGDJc8p0+HUqVc4mzTxvUbtTuI5k+RuDtpeeAKb3GBqu7KQXy+\ns2RqciIQgE8+cUM3q1aFF1/M+mZ1x4+7SwYptc+1ayEu7kTts2bNXAndRJjfNdNvgddUdWrQuvdU\n9b6chekfS6bGD0lJ7oZ9AwbAJZe4mmrDhie2r1t3InnGx8OFF55Ins2awWmnRS10EyF+J9PfgY3A\nTFXt761LbYzKCyyZGj8lJrrW9pdecpONxMa6BHro0MkNR7Gx0Y7URJrfyTQBaAK8AVQD7gJmWTI1\nBd2hQ/Duu67G2ratG6rqVyOVyR98T6ZBXaK6AY8CZ6lqnplOwpKpMSYSwkmmocy0/27KE1X9QERW\nAg9mNzhjjCmIMpscuoyq7heRszn11s6iqruzXahIOWA8cC6wHrhdVU+ZkkJE1gP7gWTgeEZ9W61m\naoyJBF9O80Vkqqre4CW0tDupqma784eIDAZ2qepgEXkCd9ngyXT2+x1orKp7sjieJVNjjO/y/K2e\nRWQN0EpVt4tIJSBeVWuls9/vwGVZ1YItmRpjIsGvmumlmb1QVZdmI7aUY/+pqmd5zwXYk7KcZr/f\ngH240/x3VfX9DI5nydQY4zu/GqD+zamn98FaZxHEDCC9KRyeDl5QVRWRjMppoapbRSQWmCEia1R1\nbmblGmNMNGSYTFU1LicHVtXrMtomIttFpJKqbhORykC6N5lQ1a3evztFZBLQFEg3mfbr1y/1eVxc\nHHFxcdkP3hhTKMXHxxMfH5+t14Y6Nv8SoDZQLGWdqmb7ruVeA9RuVR0kIk8CZdM2QIlICaCIqh7w\nJqieDvRX1enpHM9O840xvvO7034/oBVQF5gKtAO+U9XbchBgOeAToDpBXaNEpArwvteLoCYw0XtJ\nUeAjVX05g+NZMjXG+M7vZPoj0ABYqqoNRKQiLrFdm/NQ/WHJ1BgTCX7f6vmIqiYDSSJyJu76ZrWc\nBGiMMQVNKMNJF4nIWcD7wGLgEPB9RKMyxph8JqxO+yJSAyitqisiF1L47DTfGBMJkZhpvwFwHlAE\nEFz30ImZvigXWTI1xkSCr7NGichI4BJgFRAI2pRnkqkxxkRbKNdMLwfqWtXPGGMyFkpr/iLcnUmN\nMcZkIJSa6UhgvohsAxK9daqq9SMXljHG5C+hJNPhQGfgR06+ZmqMMcYTSjLdoapTIh6JMcbkY6EM\nJx0KnAl8DhzzVlvXKGNMgef3DfWK4a6VtkmzPs8kU2OMibZMk6mIFMHNgv9oLsVjjDH5UqZdo7wJ\nTlp4txYxxhiTgVBO85cBn4nIp8Bhb12eumZqjDHRFuo10z3A1WnWWzI1xhhPVG717DdrzTfGRIKv\nk0OLSDURmSQiO73HBBE5J+dhGmNMwRHK2PyRwBSgivf43FtnjDHGE0qn/eWq2iCrddFkp/nGmEjw\n+x5Qu0XkLhEpIiJFRaQzsCtnIRpjTMESSjLtAdwObAO2Ah2A7pEMyhhj8htrzTfGmAz4MjZfRJ7L\nYJMCqOqAbMRmjDEFUmad9g/hJc4gJYG7gfKAJVNjjPGEenfSMsBDuET6CfCaqu6IcGwhs9N8Y0wk\n+DYFn4icDfwT6AR8CFyqqn/mPERjjClYMrtm+ipwM/AeUF9VD+RaVMYYk89keJovIgHczPrH09ms\nqlomkoGFw07zjTGR4MtpvqqG0gfVGGMMoXXaN8YYkwVLpsYY4wNLpsYY44NQZtrPdwrrLausEc6Y\n6CmQyRQKX2IprD8gxuQVdppvjDE+sGRqjDE+sGSaQxdeeCHjx48/ZX3r1q1DWheq+Ph4nn32WQCu\nvPLKbB/HGBMZlkxzYPny5bRu3ZrPP//ct2MGAoF01wdfE7Xro8bkPZZMc2DSpEn84x//4OjRoxw7\ndoz33nuPZs2a0adPn9R9vvjiCy677DJ69OjB8eNuZO7atWtp27YtcXFxvPjiiwB069aN3r17065d\nO7Zu3crVV19Ny5YtefDBB4HC16BmTH5TaJKpSPiPrCQkJNC4cWPatm3LV199xYgRI5g3bx4dOnRI\n3WfgwIHMmTOHAQMGsH37dgCefvppRowYQXx8PKtWrWLz5s2ICFdeeSXTpk2jfPnyzJgxg7lz57J/\n/37Wrl1rtVFj8rgC2zUqLb8rdmvXrmXlypW0a9eOxMRELrroIs4991xiYmK49NJLU/eLiYmhRIkS\nlChRgtjYWAB+/vlnOnfuDMC+ffvYvHkzAI0bNwZg165dPPDAA+zbt4/169ezZcsWf4M3xviu0CRT\nv02cOJHhw4enNir95S9/Yffu3QQCARISElL3CwQCHD58mD179rBz504AatWqxZAhQ6hUqRKBQAAR\nYejQoam1z7Fjx3LzzTfTtWtXOnfubKf4xuQDlkyz6csvv+Thhx9OXW7QoAHFixenefPmtGrVKjUx\nPvHEE1x11VVceumlVK5cGYAXX3yRHj16kJiYyGmnncaECROAEw1LV199NV26dGHy5MnW8GRMPlEg\n707qzUEYxYhyX2F8z8ZEWjjzmRaaBihjjIkkS6bGGOMDS6bGGOMDS6bGGOMDS6bGGOMDS6Y5FDzR\nSVxcHK1bt+a6666jS5cu7NixA4Du3buzbt261Ne0bNky0/2NMfmPJdMcSDvRiYgwc+ZMZsyYQffu\n3XnggQdS902vj2hm+xtj8hdLpjmQMtFJYmIix44dA05MSNK6dWv27duXOgtURn1AM9rfGJO/FJoR\nUNI//NFD+lzmneATEhLo168fbdq0YcaMGadsr1ChArt27UJV6dSpE8WLF3exZDCqKWX/ChUqhB2r\nMSa6Ck0yzSoxhiu9iU7g5OS4Y8cOypcvj4jw8ccfU7NmTeDENdO0duzYkToZijEmfyk0ydRvaSc6\nad++PYFAIPW0ffbs2ZQrV46YGHclJavT/JT9bfy9MfmTJdNsSjvRSd26dRk0aBDXXnstRYsWpXLl\nyrz99tup2zM6tc9of2NM/mITnRQQhfE9GxNpNtGJMcbkMkumxhjjA0umxhjjg6gkUxHpICKrRCRZ\nRC7NZL/rRWSNiPwqIk+EWUaheoQjPj4+rP2zIzfKyK1yCkoZuVVOQSkjXNGqma4EbgbmZLSDiBQB\n3gKuB+oAd4pI7VAOrqq+P5577rmIHNfPMkJVkL7sBeW92OeV98oIV1S6RqnqGsjynkZNgbWqut7b\ndxxwI7A60vEZY0y48vI106rAxqDlTd46Y4zJcyLWz1REZgCV0tnUV1U/9/aZBTyqqkvTef2twPWq\neq+33Bm4XFV7p7OvdbA0xkREqP1MI3aar6rX5fAQm4FqQcvVcLXT9MqyMZjGmKjKC6f5GSXCxcCF\nInKeiJwO3AFMyb2wjDEmdNHqGnWziGwErgCmishX3voqIjIVQFWTgF7ANOAnYLyqWuOTMSZPKhBj\n840xJtrywmm+McbkewUmmaZcKvDhOGeKyEARGSMiHdNse8enMqqJyDCvnLIiMlJEfhSR0SIS0Wn2\n/T6+iFwf9LysiAwXkZUi8rGIVPSzrHTKPtvn4yWIyDMicr6fxzWFQ75KpiJyaQaPxkAjn4oZ6f07\nATfqaoKIFPPWNfOpjA+A5cA+YAHwM/AX4AdgqE9lICLl0jzOBn5IWfapmJeDnr8GbAX+BiwC3vWp\nDERkkIjEes8vE5HfgIUi8oeIxPlUTFnvMUtEFonIP0Wkik/HBkBEmojILO/HupqIzBCRfV55fn2H\nEZHSIjJA3LDt/SKyS0QWikg3H8so61UI1ojInyKyx3s+UETK+lVOJuX7UoHyjpXjSlS+umYqIslk\nPAT1ClUt7kMZy1W1QdDy07hEdyMwQ1Vz/IUXkWWq2tB7/oeqVk9vmw/lBIANaVafg+tipqpa04cy\nElI+ExFZDjRMmVw27WeZw3J+VNV63vN44DFVXSQiFwFjVbWxD2UkqGojcUPzWgJ34oY9r/bKeM+H\nMhYB/8Il7VeAfwL/A64GXlBVX36wRWQKMAn4BugAlALGAc8Am1S1rw9lTAdmAqOA7aqqIlIZ6Apc\nraptfCgjo7k7BJiqqun1Zc9OOROBX4CFQA/gGNBJVY8Gf8czFenx5n4+gFXARRls2+hTGauBmDTr\nunllb/CpjOVBz19Ms22lj5/Xo8DXQP2gdb/7/H+yCejjlbUe7wfa27bCx3JWA6d5zxdE4jMDEtJZ\nVxQ3P8RIv8sA/kizbZmPn9eKNMuLvX9jgJ99KuOX7GwLs4xkYFYGjyM+fl7L0yw/DcwDyqf3vUjv\nkd9uW9KPjC9NPORTGV8A1wCptxtV1Q9EZBvwpk9lTBGR0qp6QFWfTlkpIhfiTvl9oaqvicgnwL9F\nZBPwnF/HDjIMKO09H4n78u30aijLfCznHeBLEXkZ+FpEXgcm4mp0fpXzS9oV6rrofe09/HBcRNoC\nZwIxInKzqk4SkVZAok9lABwSkZaqOldEbgR2A6hqQPy7z9gGEXkcGKWq2wFEpBKuZvqHT2WsAf6h\nqqf833jdK/1yuojEqGoAQFVfFJHNwGxcrT5L+eo0H0DczFE3cmKc/iZgivrYB7WglJGmvBuBp4Aa\nquprw1BuvRcRaQ3cD1yEqzFuAiYDI1T1uE9lRPS9iEhTYDCwBXgSGA5cDqwF7lPVxT6V0wD3Q3ch\n8CNwt6r+7F137qiqr/tQRjnce2gPpHyntuMG1wxU1T0+lNEBd+axJp1tN6nq5JyW4R3rFWC6qs5I\ns+/H59AAAAMFSURBVP564E1VvTDLY+SnZCpuTtM7cdd+UoaWVsONjhqvqi9n9NrCVkZQWcHJoQTw\nOzDBx+QQrfcCbsjxZ/ntvWTwPqao6k9+HD9NOTcBKY1oEf3BTlN2d1UdmfWeOSqjh6qOiGQZ4ZST\n35Lpr0CdtLUQccNNf1LVC6yMk46XGz8M9l7CKyO3Enau/chlUP5GVa2W9Z55u4xwyslv10yTcb/m\n69Osr+JtszJOdg/pJ4fXcEN0/fiDsvcSntx4H7lSjoiszGSzL5eScqMMv8rJb8n0EeAbEVnLiblO\nq+GuC/WyMk6RG8nB3kt4cuvHJzfKqYDr6fBnOtu+z0dl+FJOvkqmqvq1iFyMm4W/KqC4602LvVZX\nK+NkEU8O9l7Clls/PrlRzlSglKompN0gIrPzURm+lJOvrpma8Im7l1akE12uKCjvJbfeR0H5vPIL\nS6bGGOODfDU23xhj8ipLpsYY4wNLpsYY4wNLpiZfEpGAiIwOWi4qIjtF5PNsHu9MEXkgaDkuu8cy\nhZMlU5NfHQLqyom5Zq/Dm1owm8c7C+jpR2CmcLJkavKzL4EbvOd3AmPx7nYrbgLsySKyXETmi8gl\n3vp+IjJC3ATN60Skt/f6gcD54mbbH4xLyqVE5FMRWS0iY3L3rZn8xpKpyc/GA38XkTOAS3AT+6bo\nDyxRNzl1X+DDoG0XAW1wfTCf8/pjPgGsU9VGqvo4Lik3Ah4G6gA1RaRFpN+Qyb8smZp8S1VXAufh\naqVT02xuAYz29psFnC0ipXE1zqmqelxVdwM7cGOv05vk8wdV3aKuM/Yyryxj0pWvhpMak44pwKtA\nKyA2zbaMZkE+FvQ8mYz/DhJD3M8Yq5mafG8E0E9VV6VZPxfoBK5lHtipqgfIOMEe4MQdA4wJm/3S\nmvxKAVR1M/BW0LqU1vx+wAhxN/k7hLuVRtp9ThxMdbeIzPOmYvvSe6Tdz8ZemwzZ2HxjjPGBneYb\nY4wPLJkaY4wPLJkaY4wPLJkaY4wPLJkaY4wPLJkaY4wPLJkaY4wP/h8bsm0zbfs4YgAAAABJRU5E\nrkJggg==\n", + "text/plain": [ + "" + ] + }, + "metadata": {}, + "output_type": "display_data" + }, + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVMAAAFSCAYAAABPFzzRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n", + "text/plain": [ + "" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "college_towns_to_percentages = {}\n", + "#read in college towns. \n", + "for line in open('college_town_percentages'):\n", + " line = ' '.join(line.split()[1:])\n", + " city_name = line.split('(')[0].strip()\n", + " percentage = float(line.split('(')[1].replace('%)', ''))\n", + " if city_name in d.index:\n", + " college_towns_to_percentages[city_name] = percentage\n", + "#reprocess google data into easier-to-analyze format. \n", + "def createGoogleSearchDataframe(all_searches, college_towns_to_percentages):\n", + " df = {'month':[], 'year':[], 'town':[]}\n", + " for k in all_searches:\n", + " df[k] = []\n", + " df[k + '_normalized_by_town'] = []\n", + " df[k + '_zero_meaned_by_town'] = []\n", + " \n", + " for i, town_name in enumerate(college_towns_to_percentages.keys()):\n", + " if i % 10 == 0:\n", + " print i,'towns processed out of', len(d.index)\n", + " \n", + " mean_vals_by_town = {}\n", + " bad_town = False\n", + " for k in all_searches:\n", + " mean_vals_by_town[k] = {}\n", + " if town_name not in all_searches[k].index:\n", + " print town_name, 'not in index'\n", + " bad_town = True\n", + " if town_name in all_searches[k].index:\n", + " mean_vals_by_town[k]['mu'] = all_searches[k].loc[town_name].mean()\n", + " mean_vals_by_town[k]['sigma'] = all_searches[k].loc[town_name].std()\n", + " if bad_town:\n", + " continue\n", + " for date in all_searches['Adderall'].columns:\n", + " month = date.split('-')[1]\n", + " year = date.split('-')[0]\n", + " df['month'].append(month)\n", + " df['year'].append(year)\n", + " df['town'].append(town_name)\n", + " for k in all_searches.keys():\n", + " search_volume = float(all_searches[k].loc[town_name][date])\n", + " df[k].append(search_volume)\n", + " df[k + '_normalized_by_town'].append((search_volume - mean_vals_by_town[k]['mu']) / mean_vals_by_town[k]['sigma'])\n", + " df[k + '_zero_meaned_by_town'].append((search_volume - mean_vals_by_town[k]['mu']))\n", + "\n", + " for a in df.keys():\n", + " print a, len(df[a])\n", + " df = pd.DataFrame(df)\n", + " return df\n", + "google_search_dataframe = createGoogleSearchDataframe(google_data, college_towns_to_percentages)\n", + "\n", + "#Do regressions for various adjustments to the Google data to confirm we see the same pattern. \n", + "#Eg, if we subtract off the mean for each town, do we still see it? What if we divide by each town's standard deviation...? Pattern is robust to all these things. \n", + "for adjustment_to_use in ['', '_normalized_by_town', '_zero_meaned_by_town']:\n", + " figure(figsize = [5, 5])\n", + " for k in ['Adderall', 'ADHD']:\n", + " model = sm.OLS.from_formula('%s ~ C(month, Sum) + C(year, Sum)' \n", + " % (k + adjustment_to_use), data = google_search_dataframe).fit()\n", + " month_coefs = []\n", + " month_ticks = []\n", + " for param in model.params.index:\n", + " if 'month' in param:\n", + " month_coefs.append(model.params[param])\n", + " month_ticks.append(param.split('S.')[1].replace(']', ''))\n", + " month_coefs.append(- np.mean(month_coefs))\n", + " month_ticks.append('12')\n", + " plot(range(len(month_coefs)), month_coefs, label = k)\n", + " xticks(range(len(month_coefs)), month_ticks, rotation = 90)\n", + " ylim([-1, 1])\n", + " legend(loc = 3, fontsize = 8)\n", + "\n", + " xlabel('Month')\n", + " ylabel('Normalized Search Rate in College Towns')\n", + " title('Adjustment:' +adjustment_to_use)\n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "###### \"In fact, NSDUH data showed that adults whose family incomes were below 10,000 had the highest rates of nonmedical Adderall use, and those whose family incomes were greater than 75,000 had the lowest.\"" + ] + }, + { + "cell_type": "code", + "execution_count": 17, + "metadata": { + "collapsed": false + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Examining Adderall use frequency in adults as a function of family income\n", + "Ritalin results are qualitatively similar although the effects seem to be weaker\n", + "\n", + "Categories for total family income:\n", + "1: <10k, 2: 10-20, 3: 20-30; 4: 30-40; 5: 40-50; 6: 50-75; 7: 75\n", + "Category IRFAMIN3\n", + "Mean value of ADDERALL for level 1 is 0.037; unweighted, 0.053 (1132 values)\n", + "Mean value of ADDERALL for level 2 is 0.025; unweighted, 0.036 (2308 values)\n", + "Mean value of ADDERALL for level 3 is 0.032; unweighted, 0.042 (2102 values)\n", + "Mean value of ADDERALL for level 4 is 0.032; unweighted, 0.040 (2100 values)\n", + "Mean value of ADDERALL for level 5 is 0.029; unweighted, 0.036 (2085 values)\n", + "Mean value of ADDERALL for level 6 is 0.028; unweighted, 0.033 (3418 values)\n", + "Mean value of ADDERALL for level 7 is 0.024; unweighted, 0.031 (6137 values)\n", + "Ratio between maximum value and minimum value 1.52577800068\n", + "Category IRFAMIN3\n", + "Mean value of RITALIN for level 1 is 0.030; unweighted, 0.038 (1132 values)\n", + "Mean value of RITALIN for level 2 is 0.018; unweighted, 0.025 (2308 values)\n", + "Mean value of RITALIN for level 3 is 0.024; unweighted, 0.031 (2102 values)\n", + "Mean value of RITALIN for level 4 is 0.026; unweighted, 0.031 (2100 values)\n", + "Mean value of RITALIN for level 5 is 0.018; unweighted, 0.022 (2085 values)\n", + "Mean value of RITALIN for level 6 is 0.022; unweighted, 0.026 (3418 values)\n", + "Mean value of RITALIN for level 7 is 0.020; unweighted, 0.025 (6137 values)\n", + "Ratio between maximum value and minimum value 1.67773828665\n", + "Optimization terminated successfully.\n", + " Current function value: 0.154529\n", + " Iterations 7\n", + "Optimization terminated successfully.\n", + " Current function value: 0.154656\n", + " Iterations 7\n", + "Optimization terminated successfully.\n", + " Current function value: 0.132304\n", + " Iterations 11\n", + "Optimization terminated successfully.\n", + " Current function value: 0.124139\n", + " Iterations 8\n", + "Optimization terminated successfully.\n", + " Current function value: 0.124316\n", + " Iterations 8\n", + "Optimization terminated successfully.\n", + " Current function value: 0.111418\n", + " Iterations 10\n", + "\n", + "==========================================================================================\n", + " ADDERALL I ADDERALL II ADDERALL III RITALIN I RITALIN II RITALIN III\n", + "------------------------------------------------------------------------------------------\n", + "C(CATAG6, Sum)[S.3] 2.208*** 1.373*** \n", + " (0.186) (0.104) \n", + "C(CATAG6, Sum)[S.4] 0.742*** 0.467*** \n", + " (0.191) (0.110) \n", + "C(CATAG6, Sum)[S.5] -0.150 -0.244* \n", + " (0.215) (0.139) \n", + "C(IRFAMIN3)[T.2] -0.406** -0.387** -0.426** -0.426** \n", + " (0.173) (0.181) (0.205) (0.210) \n", + "C(IRFAMIN3)[T.3] -0.236 -0.365** -0.213 -0.360* \n", + " (0.171) (0.179) (0.200) (0.206) \n", + "C(IRFAMIN3)[T.4] -0.283 -0.458** -0.196 -0.401* \n", + " (0.173) (0.181) (0.200) (0.205) \n", + "C(IRFAMIN3)[T.5] -0.405** -0.650*** -0.560*** -0.837*** \n", + " (0.177) (0.185) (0.215) (0.221) \n", + "C(IRFAMIN3)[T.6] -0.484*** -0.825*** -0.390** -0.785*** \n", + " (0.163) (0.172) (0.189) (0.195) \n", + "C(IRFAMIN3)[T.7] -0.572*** -0.834*** -0.421** -0.811*** \n", + " (0.152) (0.161) (0.175) (0.182) \n", + "C(NEWRACE2, Sum)[S.1] 0.795*** 1.001*** \n", + " (0.114) (0.139) \n", + "C(NEWRACE2, Sum)[S.2] -0.798*** -1.727*** \n", + " (0.191) (0.346) \n", + "C(NEWRACE2, Sum)[S.3] 0.424 0.441 \n", + " (0.269) (0.334) \n", + "C(NEWRACE2, Sum)[S.4] 0.156 0.411 \n", + " (0.451) (0.516) \n", + "C(NEWRACE2, Sum)[S.5] -0.606** -0.480 \n", + " (0.270) (0.329) \n", + "C(NEWRACE2, Sum)[S.6] 0.858*** 1.182*** \n", + " (0.203) (0.226) \n", + "IRFAMIN3 -0.068*** -0.042* \n", + " (0.019) (0.022) \n", + "IRSEX -0.264*** -0.326*** \n", + " (0.080) (0.090) \n", + "Intercept -2.883*** -2.963*** -3.964*** -3.232*** -3.379*** -3.748*** \n", + " (0.133) (0.096) (0.274) (0.155) (0.112) (0.256) \n", + "N 19282 19282 19282 19282 19282 19282 \n", + "==========================================================================================\n", + "Standard errors in parentheses.\n", + "* p<.1, ** p<.05, ***p<.01\n" + ] + } + ], + "source": [ + "#run regressions + stratify by income levels to confirm that we observe income effects.\n", + "adult_idxs = nsduh_data['CATAG6'] >= 3\n", + "\n", + "data_to_use = nsduh_data.loc[adult_idxs]\n", + "weights = data_to_use['ANALWT_C']\n", + "print 'Examining Adderall use frequency in adults as a function of family income'\n", + "print 'Ritalin results are qualitatively similar although the effects seem to be weaker'\n", + "print '\\nCategories for total family income:\\n1: <10k, 2: 10-20, 3: 20-30; 4: 30-40; 5: 40-50; 6: 50-75; 7: 75'\n", + "stratifyByCat(data_to_use, 'IRFAMIN3') \n", + "stratifyByCat(data_to_use, 'IRFAMIN3', drug_to_measure = 'RITALIN')\n", + "\n", + "\n", + "model1 = sm.Logit.from_formula('ADDERALL ~ C(IRFAMIN3)', data = data_to_use, weights = weights).fit()#treat income as categorical variable\n", + "model2 = sm.Logit.from_formula('ADDERALL ~ IRFAMIN3', data = data_to_use, weights = weights).fit()#treat as continuous variable (bad, bad, bad)\n", + "model3 = sm.Logit.from_formula('ADDERALL ~ IRSEX + C(CATAG6, Sum) + C(NEWRACE2, Sum) + C(IRFAMIN3)', data = data_to_use, weights = weights).fit()#control for other covariates: age, race, sex. \n", + "model4 = sm.Logit.from_formula('RITALIN ~ C(IRFAMIN3)', data = data_to_use, weights = weights).fit()\n", + "model5 = sm.Logit.from_formula('RITALIN ~ IRFAMIN3', data = data_to_use, weights = weights).fit()\n", + "model6 = sm.Logit.from_formula('RITALIN ~ IRSEX + C(CATAG6, Sum) + C(NEWRACE2, Sum) + C(IRFAMIN3)', data = data_to_use, weights = weights).fit()\n", + "\n", + "models = [model1, model2, model3, model4, model5, model6]\n", + "print summary_col(models, model_names = ['ADDERALL', 'ADDERALL', 'ADDERALL', 'RITALIN', 'RITALIN', 'RITALIN'], stars = True,\n", + " float_format='%0.3f',\n", + " info_dict={'N':lambda x: \"{0:d}\".format(int(x.nobs))})" + ] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 2", + "language": "python", + "name": "python2" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 2 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython2", + "version": "2.7.10" + } + }, + "nbformat": 4, + "nbformat_minor": 0 +}