Regression Discontinuity Design.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Introduction to Regression Discontinuity Design"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Code and data taken from: https://github.com/natematias/research_in_python/blob/master/regression_discontinuity/Regression%20Discontinuity%20Analysis.ipynb"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Regression Discontinuity** is a method that makes use of a treatment applied at a cutoff point. It takes advantage of the fact that subjects at either side of the cutoff are not likely to substantially differ from one another on confounders. Thus, any difference in outcomes are only attributable to the intervention itself.\n",
    "\n",
    "In this notebook, we will learn how to implement and visualize a RDD design. We are using data from Joshua Angrist and Victor Lavy's \"Using Maimonides Rule to Estimate the Effect of Class Size on Scholastic Achievement.\" A common question in education and social science research is whether class size has an effect on student performance. At both the K-12 and university level, propnents of smaller class sizes argue that large class sizes cause teachers to divide their time and attention too thinly, and students suffer from the lack of individual attention. Moreover, large class sizes might be more distracting and impede student comprehension and achievement in a number of other ways.\n",
    "\n",
    "In this study, Angrist and Lavy make use of the fact that Israel implemented a rule that automatically split classes if they had more than 40 students. They compared student outcomes for students who were in classes with slightly fewer than 40 students to students who were in classes that were just barely split up. \n",
    "\n",
    "In this analysis, we examine a dataset that includes school level data for:\n",
    "\n",
    "- size: fifth-grade cohort size\n",
    "- intended_classsize: average intended class size for each school\n",
    "- observed_classize: observed average class size for each school\n",
    "- read: average reading achievement in cohort"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import modules\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "import statsmodels.formula.api as smf "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load data\n",
    "class_df = pd.read_sas('angrist.sas7bdat')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>read</th>\n",
       "      <th>size</th>\n",
       "      <th>intended_classize</th>\n",
       "      <th>observed_classize</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>58.000000</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>82.629997</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>46.330002</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>66.559998</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>11.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>75.000000</td>\n",
       "      <td>10.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        read  size  intended_classize  observed_classize\n",
       "0  58.000000   8.0                8.0                8.0\n",
       "1  82.629997   9.0                9.0                9.0\n",
       "2  46.330002   9.0                9.0                9.0\n",
       "3  66.559998   9.0                9.0               11.0\n",
       "4  75.000000  10.0               10.0               10.0"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Explore the data\n",
    "class_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>read</th>\n",
       "      <th>size</th>\n",
       "      <th>intended_classize</th>\n",
       "      <th>observed_classize</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>2019.000000</td>\n",
       "      <td>2019.000000</td>\n",
       "      <td>2019.000000</td>\n",
       "      <td>2019.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>74.379171</td>\n",
       "      <td>77.741951</td>\n",
       "      <td>30.955935</td>\n",
       "      <td>29.935116</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>7.678460</td>\n",
       "      <td>38.810731</td>\n",
       "      <td>6.107924</td>\n",
       "      <td>6.545885</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>34.799999</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>8.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>69.855000</td>\n",
       "      <td>50.000000</td>\n",
       "      <td>27.000000</td>\n",
       "      <td>26.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>75.379997</td>\n",
       "      <td>72.000000</td>\n",
       "      <td>31.666666</td>\n",
       "      <td>31.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>79.839996</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>35.666668</td>\n",
       "      <td>35.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>93.860001</td>\n",
       "      <td>226.000000</td>\n",
       "      <td>40.000000</td>\n",
       "      <td>44.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              read         size  intended_classize  observed_classize\n",
       "count  2019.000000  2019.000000        2019.000000        2019.000000\n",
       "mean     74.379171    77.741951          30.955935          29.935116\n",
       "std       7.678460    38.810731           6.107924           6.545885\n",
       "min      34.799999     8.000000           8.000000           8.000000\n",
       "25%      69.855000    50.000000          27.000000          26.000000\n",
       "50%      75.379997    72.000000          31.666666          31.000000\n",
       "75%      79.839996   100.000000          35.666668          35.000000\n",
       "max      93.860001   226.000000          40.000000          44.000000"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class_df.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question 1: What would be the problem with comparing all students in small classes v. all students in large classes? How might students far away from the cutoff differ in important ways?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prepare data for analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The key to a regression discontinuity is that we need to distinguish between units that were given treatment, and units that were under control. The usual approach to doing this is to create dummy indicators for whether an observations falls in the pre- or post- side of the intervention."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question 2: What is a running variable? How does it relate to the treatment cutoff?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x21fa73b7da0>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Write a function that returns a '1' if a class size is large (>40), and a '0' if it is small (<=40)\n",
    "\n",
    "def small(size):\n",
    "    if(size >= 41):\n",
    "        return 1\n",
    "    return 0\n",
    "\n",
    "# Create dummy for whether a class size is small, and csize which measures the difference between observed_classize and the cutoff (41)\n",
    "class_df['small'] = class_df['size'].map(small)\n",
    "class_df['csize'] = class_df['size'].map(lambda x: x-41)\n",
    "\n",
    "# summarize the read variable by each class size group\n",
    "class_df[(class_df['size'] >= 36) & (class_df['size'] <= 46)].boxplot('read', 'csize')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>read</th>\n",
       "      <th>size</th>\n",
       "      <th>intended_classize</th>\n",
       "      <th>observed_classize</th>\n",
       "      <th>small</th>\n",
       "      <th>csize</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>58.000000</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>0</td>\n",
       "      <td>-33.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>82.629997</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0</td>\n",
       "      <td>-32.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>46.330002</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>0</td>\n",
       "      <td>-32.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>66.559998</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>11.0</td>\n",
       "      <td>0</td>\n",
       "      <td>-32.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>75.000000</td>\n",
       "      <td>10.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>0</td>\n",
       "      <td>-31.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        read  size  intended_classize  observed_classize  small  csize\n",
       "0  58.000000   8.0                8.0                8.0      0  -33.0\n",
       "1  82.629997   9.0                9.0                9.0      0  -32.0\n",
       "2  46.330002   9.0                9.0                9.0      0  -32.0\n",
       "3  66.559998   9.0                9.0               11.0      0  -32.0\n",
       "4  75.000000  10.0               10.0               10.0      0  -31.0"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Regression Discontinuity"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we're ready to fit a model! Let's narrow the window down to class sizes between 29 and 53, and then estimate the regression."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                   read   R-squared:                       0.015\n",
      "Model:                            OLS   Adj. R-squared:                  0.010\n",
      "Method:                 Least Squares   F-statistic:                     3.103\n",
      "Date:                Tue, 26 Jun 2018   Prob (F-statistic):             0.0459\n",
      "Time:                        17:37:33   Log-Likelihood:                -1536.6\n",
      "No. Observations:                 423   AIC:                             3079.\n",
      "Df Residuals:                     420   BIC:                             3091.\n",
      "Df Model:                           2                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================\n",
      "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "Intercept     70.1188      1.151     60.908      0.000      67.856      72.382\n",
      "csize         -0.1392      0.119     -1.168      0.243      -0.373       0.095\n",
      "small          3.9531      1.800      2.197      0.029       0.416       7.491\n",
      "==============================================================================\n",
      "Omnibus:                       34.356   Durbin-Watson:                   1.967\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):               40.549\n",
      "Skew:                          -0.711   Prob(JB):                     1.57e-09\n",
      "Kurtosis:                       3.529   Cond. No.                         34.2\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n"
     ]
    }
   ],
   "source": [
    "# Subset the data to size between 29 and 53\n",
    "window = class_df[(class_df['size'] >= 29) & (class_df['size'] <= 53)]\n",
    "\n",
    "# Fit OLS model with smf\n",
    "result = smf.ols(formula = 'read ~ csize + small',\n",
    "                data = window).fit()\n",
    "\n",
    "# Print results\n",
    "print(result.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Regression Discontinuity: Reading Scores by Class Size Before and After the Cutoff')"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 960x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot results with cutoff\n",
    "\n",
    "plt.figure(num = None, figsize = (12, 6), dpi = 80, facecolor = 'w', edgecolor = 'k')\n",
    "plt.scatter(window.csize, window.read, color = 'blue')\n",
    "l = window[window.csize < 0].csize.count()\n",
    "plt.plot(window.csize[0:l], result.predict()[0:l], '-', color = \"r\")\n",
    "plt.plot(window.csize[l:], result.predict()[l:], '-', color = \"r\")\n",
    "plt.axvline(x = -.5, color = 'black', linestyle = '--')\n",
    "plt.title(\"Regression Discontinuity: Reading Scores by Class Size Before and After the Cutoff\", fontsize = \"18\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}