Skip to content

Instantly share code, notes, and snippets.

@AllenDowney

AllenDowney/political_party_soln.ipynb

Created December 28, 2018 17:35
Show Gist options
  • Star 1 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save AllenDowney/476b9c77a6581d689528c99d688685e9 to your computer and use it in GitHub Desktop.
Save AllenDowney/476b9c77a6581d689528c99d688685e9 to your computer and use it in GitHub Desktop.
Download ZIP
Display the source blob
Display the rendered blob
Raw
political_party_soln.ipynb
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Think Bayes\n",
"\n",
"This notebook presents example code and exercise solutions for Think Bayes.\n",
"\n",
"Copyright 2018 Allen B. Downey\n",
"\n",
"MIT License: https://opensource.org/licenses/MIT"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Race, religion and politics\n",
"\n",
"In their November 3, 2018 issue, *The Economist* published the [following figure](https://www.economist.com/graphic-detail/2018/11/03/how-to-forecast-an-americans-vote) showing results from their analysis of data from [YouGov](https://today.yougov.com/).\n",
"\n",
"![title](./figs/economist.png)\n",
"\n",
"These results are probably based on logistic regression, or something like it. As an exercise in conditional probability, I will try to replicate their results using data from the General Social Survey (GSS). Rather than use a regression model, I will just count the number of respondents in each group.\n",
"\n",
"This is just an exercise; my results should not be taken too seriously. \n",
"\n",
"- First, I am using GSS data from the entire history of the survey, going back to 1972.\n",
"\n",
"- Second, many of the conditions I use are only rough matches for the condition *The Economist* uses.\n",
"\n",
"- Also, the way I am using the GSS does not make it a representative survey.\n",
"\n",
"- Finally, some of the conditional probabilities I compute are based on small sample sizes.\n",
"\n",
"The point of the exercise is to practice thinking about and computing conditional probabilities.\n",
"\n",
"Here are the libraries I'll use."
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"# Configure Jupyter so figures appear in the notebook\n",
"%matplotlib inline\n",
"\n",
"# Configure Jupyter to display the assigned value after an assignment\n",
"%config InteractiveShell.ast_node_interactivity='last_expr_or_assign'\n",
"\n",
"import pandas as pd\n",
"import numpy as np\n",
"import matplotlib.pyplot as plt"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Here are functions to compute probabilities, counts, and conditional probabilities."
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
"def prob(A):\n",
" \"\"\"Probability of A\"\"\"\n",
" return A.mean()"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [],
"source": [
"def count(A):\n",
" \"\"\"Number of instances of A\"\"\"\n",
" return A.sum()"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [],
"source": [
"def conditional(A, B):\n",
" \"\"\"Conditional probability of A given B\"\"\"\n",
" return prob(A & B) / prob(B)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### GSS data\n",
"\n",
"The GSS data I'm using is from [this extract](https://gssdataexplorer.norc.org/projects/54786)."
]
},
{
"cell_type": "code",
"execution_count": 5,
"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>year</th>\n",
" <th>relig</th>\n",
" <th>srcbelt</th>\n",
" <th>region</th>\n",
" <th>adults</th>\n",
" <th>wtssall</th>\n",
" <th>ballot</th>\n",
" <th>cohort</th>\n",
" <th>feminist</th>\n",
" <th>polviews</th>\n",
" <th>partyid</th>\n",
" <th>race</th>\n",
" <th>sex</th>\n",
" <th>educ</th>\n",
" <th>age</th>\n",
" <th>indus10</th>\n",
" <th>occ10</th>\n",
" <th>id_</th>\n",
" <th>realinc</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>1972</td>\n",
" <td>3</td>\n",
" <td>3</td>\n",
" <td>3</td>\n",
" <td>1</td>\n",
" <td>0.4446</td>\n",
" <td>0</td>\n",
" <td>1949</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>2</td>\n",
" <td>1</td>\n",
" <td>2</td>\n",
" <td>16</td>\n",
" <td>23</td>\n",
" <td>5170</td>\n",
" <td>520</td>\n",
" <td>1</td>\n",
" <td>18951.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>1972</td>\n",
" <td>2</td>\n",
" <td>3</td>\n",
" <td>3</td>\n",
" <td>2</td>\n",
" <td>0.8893</td>\n",
" <td>0</td>\n",
" <td>1902</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>1</td>\n",
" <td>1</td>\n",
" <td>10</td>\n",
" <td>70</td>\n",
" <td>6470</td>\n",
" <td>7700</td>\n",
" <td>2</td>\n",
" <td>24366.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>1972</td>\n",
" <td>1</td>\n",
" <td>3</td>\n",
" <td>3</td>\n",
" <td>2</td>\n",
" <td>0.8893</td>\n",
" <td>0</td>\n",
" <td>1924</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>3</td>\n",
" <td>1</td>\n",
" <td>2</td>\n",
" <td>12</td>\n",
" <td>48</td>\n",
" <td>7070</td>\n",
" <td>4920</td>\n",
" <td>3</td>\n",
" <td>24366.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>1972</td>\n",
" <td>5</td>\n",
" <td>3</td>\n",
" <td>3</td>\n",
" <td>2</td>\n",
" <td>0.8893</td>\n",
" <td>0</td>\n",
" <td>1945</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>1</td>\n",
" <td>2</td>\n",
" <td>17</td>\n",
" <td>27</td>\n",
" <td>5170</td>\n",
" <td>800</td>\n",
" <td>4</td>\n",
" <td>30458.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>1972</td>\n",
" <td>1</td>\n",
" <td>3</td>\n",
" <td>3</td>\n",
" <td>2</td>\n",
" <td>0.8893</td>\n",
" <td>0</td>\n",
" <td>1911</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>0</td>\n",
" <td>1</td>\n",
" <td>2</td>\n",
" <td>12</td>\n",
" <td>61</td>\n",
" <td>6680</td>\n",
" <td>5020</td>\n",
" <td>5</td>\n",
" <td>50763.0</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" year relig srcbelt region adults wtssall ballot cohort feminist \\\n",
"0 1972 3 3 3 1 0.4446 0 1949 0 \n",
"1 1972 2 3 3 2 0.8893 0 1902 0 \n",
"2 1972 1 3 3 2 0.8893 0 1924 0 \n",
"3 1972 5 3 3 2 0.8893 0 1945 0 \n",
"4 1972 1 3 3 2 0.8893 0 1911 0 \n",
"\n",
" polviews partyid race sex educ age indus10 occ10 id_ realinc \n",
"0 0 2 1 2 16 23 5170 520 1 18951.0 \n",
"1 0 1 1 1 10 70 6470 7700 2 24366.0 \n",
"2 0 3 1 2 12 48 7070 4920 3 24366.0 \n",
"3 0 1 1 2 17 27 5170 800 4 30458.0 \n",
"4 0 0 1 2 12 61 6680 5020 5 50763.0 "
]
},
"execution_count": 5,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"from utils import read_gss\n",
"\n",
"gss = read_gss('data/gss_bayes')\n",
"gss.head()"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [],
"source": [
"def replace_invalid(series, bad_vals, replacement=np.nan):\n",
" series.replace(bad_vals, replacement, inplace=True)\n",
" \n",
"replace_invalid(gss.partyid, [3, 7, 8, 9])\n",
"replace_invalid(gss.relig, [98, 99])\n",
"replace_invalid(gss.educ, [98, 99])"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [],
"source": [
"def values(series):\n",
" return series.value_counts().sort_index()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### relig\n",
"\n",
"https://gssdataexplorer.norc.org/projects/54786/variables/287/vshow"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"1.0 35968\n",
"2.0 15181\n",
"3.0 1246\n",
"4.0 7254\n",
"5.0 1069\n",
"6.0 177\n",
"7.0 89\n",
"8.0 38\n",
"9.0 136\n",
"10.0 112\n",
"11.0 762\n",
"12.0 30\n",
"13.0 135\n",
"Name: relig, dtype: int64"
]
},
"execution_count": 8,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"values(gss.relig)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### srcbelt\n",
"\n",
"https://gssdataexplorer.norc.org/projects/54786/variables/121/vshow"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"1 5572\n",
"2 8670\n",
"3 7113\n",
"4 9348\n",
"5 23583\n",
"6 8180\n",
"Name: srcbelt, dtype: int64"
]
},
"execution_count": 9,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"values(gss.srcbelt)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### region\n",
"\n",
"https://gssdataexplorer.norc.org/projects/54786/variables/119/vshow"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"1 2976\n",
"2 9057\n",
"3 11502\n",
"4 4559\n",
"5 12039\n",
"6 4121\n",
"7 5923\n",
"8 3882\n",
"9 8407\n",
"Name: region, dtype: int64"
]
},
"execution_count": 10,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"values(gss.region)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### partyid\n",
"\n",
"https://gssdataexplorer.norc.org/projects/52787/variables/141/vshow"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.0 9999\n",
"1.0 12942\n",
"2.0 7485\n",
"4.0 5462\n",
"5.0 9661\n",
"6.0 6063\n",
"Name: partyid, dtype: int64"
]
},
"execution_count": 11,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"values(gss.partyid)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### race\n",
"\n",
"https://gssdataexplorer.norc.org/projects/52787/variables/82/vshow"
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"1 50340\n",
"2 8802\n",
"3 3324\n",
"Name: race, dtype: int64"
]
},
"execution_count": 12,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"values(gss.race)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### sex\n",
"\n",
"https://gssdataexplorer.norc.org/projects/52787/variables/81/vshow"
]
},
{
"cell_type": "code",
"execution_count": 13,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"1 27562\n",
"2 34904\n",
"Name: sex, dtype: int64"
]
},
"execution_count": 13,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"values(gss.sex)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Missing data\n",
"\n",
"To keep things simple, I'm dropping rows that are missing any of the data I need."
]
},
{
"cell_type": "code",
"execution_count": 14,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(51397, 19)"
]
},
"execution_count": 14,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"subset = gss.dropna(subset=['sex', 'race', 'partyid', 'region', 'relig', 'educ', 'realinc', 'age'])\n",
"subset.shape"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Boolean variables\n",
"\n",
"The following line makes the columns from `subset` available as global variables. Kids, don't try this at home!"
]
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {},
"outputs": [],
"source": [
"globals().update(subset)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Now I need boolean Series for each of the conditions *The Economist* uses."
]
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.8095414129229332"
]
},
"execution_count": 16,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"white = race==1\n",
"prob(white)"
]
},
{
"cell_type": "code",
"execution_count": 17,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.14559215518415472"
]
},
"execution_count": 17,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"black = race==2\n",
"prob(black)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Rather than \"born again, church-going Protestant\", I am just using \"Protestant\"."
]
},
{
"cell_type": "code",
"execution_count": 18,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.5979921007062668"
]
},
"execution_count": 18,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"prot = relig==1\n",
"prob(prot)"
]
},
{
"cell_type": "code",
"execution_count": 19,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"False 28788\n",
"True 22609\n",
"Name: sex, dtype: int64"
]
},
"execution_count": 19,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"male = sex==1\n",
"values(male)"
]
},
{
"cell_type": "code",
"execution_count": 20,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"False 22609\n",
"True 28788\n",
"Name: sex, dtype: int64"
]
},
"execution_count": 20,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"female = sex==2\n",
"values(female)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Rather than \"age 25\", and I am using \"young\", defined as less than 30."
]
},
{
"cell_type": "code",
"execution_count": 21,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.1974045177734109"
]
},
"execution_count": 21,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"young = age<30\n",
"prob(young)"
]
},
{
"cell_type": "code",
"execution_count": 22,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.13115551491332178"
]
},
"execution_count": 22,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"rural = srcbelt==6\n",
"prob(rural)"
]
},
{
"cell_type": "code",
"execution_count": 23,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.6042376014164251"
]
},
"execution_count": 23,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"urban = srcbelt.isin([1,2,5])\n",
"prob(urban)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"For \"never attended college\" I am using 12 or fewer years of education. "
]
},
{
"cell_type": "code",
"execution_count": 24,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.5199330700235423"
]
},
"execution_count": 24,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"no_college = educ<=12\n",
"prob(no_college)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"For \"post-graduate degree\" I am using more than 17 years of education."
]
},
{
"cell_type": "code",
"execution_count": 25,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.07578263322762029"
]
},
"execution_count": 25,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"postgrad = educ>17\n",
"prob(postgrad)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"*The Economist* considers two specific incomes, $100,000 and $15,000. Instead, I define conditions for high and low income:"
]
},
{
"cell_type": "code",
"execution_count": 26,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.07444014242076386"
]
},
"execution_count": 26,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"high_income = realinc>=80000\n",
"prob(high_income)"
]
},
{
"cell_type": "code",
"execution_count": 27,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.3670836819269607"
]
},
"execution_count": 27,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"low_income = realinc<=15000\n",
"prob(low_income)"
]
},
{
"cell_type": "code",
"execution_count": 28,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.19641224195964746"
]
},
"execution_count": 28,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"west = region.isin([8,9])\n",
"prob(west)"
]
},
{
"cell_type": "code",
"execution_count": 29,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.2568048718796817"
]
},
"execution_count": 29,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"midwest = region.isin([3,4])\n",
"prob(midwest)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The GSS has 7 levels of party identification, with \"Independent\" in the middle. I exclude \"Independent\" and \"Other party\" and classify everyone else as Democrat or Republican. "
]
},
{
"cell_type": "code",
"execution_count": 30,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.5891005311594062"
]
},
"execution_count": 30,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"democrat = partyid.isin([0,1,2])\n",
"prob(democrat)"
]
},
{
"cell_type": "code",
"execution_count": 31,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.4108994688405938"
]
},
"execution_count": 31,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"republican = partyid.isin([4,5,6])\n",
"prob(republican)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### The blue path\n",
"\n",
"Here are the conditional probabilities that make up the blue path in the figure."
]
},
{
"cell_type": "code",
"execution_count": 32,
"metadata": {},
"outputs": [],
"source": [
"blue_path = pd.Series([]);"
]
},
{
"cell_type": "code",
"execution_count": 33,
"metadata": {},
"outputs": [],
"source": [
"blue_path['Average'] = prob(democrat)"
]
},
{
"cell_type": "code",
"execution_count": 34,
"metadata": {},
"outputs": [],
"source": [
"blue_path['Black'] = conditional(democrat, black)"
]
},
{
"cell_type": "code",
"execution_count": 35,
"metadata": {},
"outputs": [],
"source": [
"blue_path['Protestant'] = conditional(democrat, black&prot)"
]
},
{
"cell_type": "code",
"execution_count": 36,
"metadata": {},
"outputs": [],
"source": [
"blue_path['Male'] = conditional(democrat, black&prot&male)"
]
},
{
"cell_type": "code",
"execution_count": 37,
"metadata": {},
"outputs": [],
"source": [
"blue_path['Young'] = conditional(democrat, black&prot&male&young)"
]
},
{
"cell_type": "code",
"execution_count": 38,
"metadata": {},
"outputs": [],
"source": [
"blue_path['Rural'] = conditional(democrat, black&prot&male&young&rural)"
]
},
{
"cell_type": "code",
"execution_count": 39,
"metadata": {},
"outputs": [],
"source": [
"blue_path['No college'] = conditional(democrat, black&prot&male&young&rural&no_college)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"I'm not able to compute the last two probabilities because there is no one in the dataset that matches the conditions."
]
},
{
"cell_type": "code",
"execution_count": 40,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0"
]
},
"execution_count": 40,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"count(black&prot&male&young&rural&no_college&high_income)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Nevertheless, here are the results for the blue path."
]
},
{
"cell_type": "code",
"execution_count": 41,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Average 0.589101\n",
"Black 0.901376\n",
"Protestant 0.907050\n",
"Male 0.894713\n",
"Young 0.839367\n",
"Rural 0.821429\n",
"No college 0.894737\n",
"dtype: float64"
]
},
"execution_count": 41,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"blue_path"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Here's some code to plot the results."
]
},
{
"cell_type": "code",
"execution_count": 42,
"metadata": {},
"outputs": [],
"source": [
"def arrow(series, index, **options):\n",
" \"\"\"Draw an arrow showing the effect of a condition.\n",
" \n",
" series: Series that maps from label to probability\n",
" index: which step in the Series to plot\n",
" options: passed to plt.plot\n",
" \"\"\"\n",
" label = series.index[index]\n",
" y1 = index\n",
" y2 = index+1\n",
" x1 = series.iloc[index]\n",
" try:\n",
" x2 = series.iloc[index+1]\n",
" plt.plot([x1, x1, x2], [y1, y2, y2], **options)\n",
" plt.text(x1, y1, label)\n",
" except IndexError:\n",
" plt.text(x1, y1, label)"
]
},
{
"cell_type": "code",
"execution_count": 43,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"for i in range(len(blue_path)):\n",
" arrow(blue_path, i, color='blue')\n",
" \n",
"plt.gca().invert_xaxis()\n",
"plt.gca().invert_yaxis()\n",
"plt.xlabel('Probability of Democrat');"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Most of these results are qualitatively consistent with *The Economist*, with a few exceptions:\n",
"\n",
"* The effect of \"Protestant\" is smaller and in the wrong direction, probably because my condition is not limited to \"born-again, church-going\" Protestants.\n",
"\n",
"* The effect of \"No college\" is in the wrong direction, but at this point in the analysis, we are down to a small number of respondents, so this result should not be taken seriously."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Red path\n",
"\n",
"Here's the same analysis for the red path."
]
},
{
"cell_type": "code",
"execution_count": 44,
"metadata": {},
"outputs": [],
"source": [
"red_path = pd.Series([]);"
]
},
{
"cell_type": "code",
"execution_count": 45,
"metadata": {},
"outputs": [],
"source": [
"red_path['Average'] = prob(republican)"
]
},
{
"cell_type": "code",
"execution_count": 46,
"metadata": {},
"outputs": [],
"source": [
"red_path['White'] = conditional(republican, white)"
]
},
{
"cell_type": "code",
"execution_count": 47,
"metadata": {},
"outputs": [],
"source": [
"red_path['Protestant'] = conditional(republican, white&prot)"
]
},
{
"cell_type": "code",
"execution_count": 48,
"metadata": {},
"outputs": [],
"source": [
"red_path['Female'] = conditional(republican, white&prot&female)"
]
},
{
"cell_type": "code",
"execution_count": 49,
"metadata": {},
"outputs": [],
"source": [
"red_path['Young'] = conditional(republican, white&prot&female&young)"
]
},
{
"cell_type": "code",
"execution_count": 50,
"metadata": {},
"outputs": [],
"source": [
"red_path['Urban'] = conditional(republican, white&prot&female&young&urban)"
]
},
{
"cell_type": "code",
"execution_count": 51,
"metadata": {},
"outputs": [],
"source": [
"red_path['Postgrad'] = conditional(republican, white&prot&female&young&urban&postgrad)"
]
},
{
"cell_type": "code",
"execution_count": 52,
"metadata": {},
"outputs": [],
"source": [
"red_path['Low income'] = conditional(republican, white&prot&female&young&urban&postgrad&low_income)"
]
},
{
"cell_type": "code",
"execution_count": 53,
"metadata": {},
"outputs": [],
"source": [
"red_path['Midwest'] = conditional(republican, white&prot&female&young&urban&postgrad&low_income&midwest)"
]
},
{
"cell_type": "code",
"execution_count": 54,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"4"
]
},
"execution_count": 54,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"count(white&prot&female&young&urban&postgrad&low_income&midwest)"
]
},
{
"cell_type": "code",
"execution_count": 55,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Average 0.410899\n",
"White 0.473899\n",
"Protestant 0.552054\n",
"Female 0.531102\n",
"Young 0.551869\n",
"Urban 0.537722\n",
"Postgrad 0.545455\n",
"Low income 0.600000\n",
"Midwest 0.500000\n",
"dtype: float64"
]
},
"execution_count": 55,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"red_path"
]
},
{
"cell_type": "code",
"execution_count": 56,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"for i in range(len(red_path)):\n",
" arrow(red_path, i, color='red')\n",
" \n",
"plt.gca().invert_yaxis()\n",
"plt.xlabel('Probability of Republican');"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Again, some of the results are consistent with *The Economist*, but there are several exceptions:\n",
"\n",
"* The effect of \"Young\" is in the wrong direction, but my definition of young is different from theirs.\n",
"\n",
"* The effect of \"Postgrad\" is in the wrong direction, but it is based on a small sample size.\n",
"\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"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": "3.6.7"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment