20 Data Sources

This chapter describes the functions and data sets associated with the lecture notes which are contained in the file DataAnalysisPAData.RData.

The following functions are included:

bday(k): This function calculates the probability of two people having the same birthday in a group of $k$ people.
samplesize(me,p=0.5,pop): This function requires the desired margin of error as the input, e.g., $me=0.03$ for $\pm 3\%$. By default, the function calculates the sample size assuming a variance maximizing probability of $p=0.5$ and infinite population. Both parameters can be adjusted.

Some data sets are generic, i.e., randomly generated, to highlight a particular concept while other are either based on public sources or are taken from academic papers. In any case, the sources are clearly indicated. Sometimes modifications were made to the data for ease of use, e.g., removing missing values or unnecessary variables. Some of the data sets are also associated with specific R packages which is indicated. All variable names are lowercase. The following data sets are included:

accidents (N=30): Generic data on the number of $accidents$, $temperature$ in Fahrenheit, and $precipitation$ in inches.
airlines (N=211918): On-time and delay statistics for the 100 largest airports (by number of 2019 arrivals) by airline January 2004 to May 2022. The variable names are mostly self-explanatory. The variables $arrflights$ and $arrdel15$ are the total number of arriving flights and the number of flight delayed at least 15 minutes. The original data is available here.
anscombe (N=11): Anscombe’s Quartet where $yi$ and $xi$ represent dependent and independent variable of set $i$, respectively. This data set is widely available and is also included in R.
apartments (N=147): The data set contains data on furnished apartments for rent. The variables are $rent$, $rooms$, $area$ (in square meters), managing $company$, and $location$ (Southwest Berlin and Potsdam).
aptitude (N=200): Aptitude ($apt$), reading ($read$), and $math$ test scores for students ($id$). The program type is indicated by $prog$. The data set is taken from the UCLA Advanced Research Computing website on Tobit Models.
beer (N=92): Quarterly beer production in Australia. The data is obtained from the lecture notes for STAT 510: Applied Time Series Analysis at Penn State.
biochemistry (N=915): The data set is based on the article The Origins of Sex Differences in Science and assesses the number of articles written by graduate students three years before receiving their Ph.D. in biochemistry.
blm (N=1359): The data is associated with the paper Black Lives Matter: Evidence that Police-Caused Deaths Predict Protest Activity. The variables include city ($geography$), total protests ($totprotests$), population ($pop$), population density ($popdensity$), percent black ($percentblack$), black poverty rate ($blackpovertyrate$), percent college-educated ($percentbachelor$), college students as a percent of the population ($collegeenrollpc$), Democratic vote share ($demshare$), police-caused deaths per 10,000 ($deathspc$), and black police-caused deaths per 10,000 ($deathsblackpc$).
bloodpressure (N=20): The data is taken from STAT 501 Regression Methods at Penn State. It contains blood pressure data for 20 individuals on $bp$ (in mm Hg), $age$, $weight$ (in kilograms), $bsa$ (body surface area in square meters), $dur$ (duration of hypertension in years), $pulse$ (basal pulse in beats per minutes), and $stress$ (index number).
bmw (N=30): Data on the prices and miles of a particular BMW 5-series model in the Indianapolis area. Some of the models have rear-wheel drive ($allwheeldrive=0$) whereas some have all-wheel drive ($allwheeldrive=1$).
boston (N=506): Housing values in suburbs of Boston. The data set is part of the R package MASS. An explanation of the various variables can be found here.
censoring (N=50): Generic data in which values of the response variable below 0 are reported as 0. The researcher observes $y$ and $x$ but not $yreal$.
chicagogrocery (N=77): The data set was represents the information from 2013 grocery stores and [Community Data Snapshots Raw Data from March 2014](https://datahub.cmap.illinois.gov/dataset/community-data-snapshots-raw-data]. It includes the following variables: $cca$ (Name of the Chicago Community Area), $stores$ (number of grocery stores), $sqft$ (sum of store square footage), $income$ (median income), $homevalue$ (median home value), $pop$ (population), $whitepop$ (white population), $laborforce$ (labor force), $unemployed$ (unemployed people), $black$ (black population), and $acres$ (area of the CCA).
childmortality (N=64): Data on child mortality ($cm$, number of deaths of children under age 5 in a year per 1000 live births), female literacy rate ($flfp$, percent), per-capita GNP in 1980 ($pgnp$), and total fertility rate ($tfr$ 1980–1985, the average number of children born to a woman, using age-specific fertility rates for a given year). The data was taken from Table 6.4 in Basic Econometrics (4th edition) by Damodar N. Gujarati.
coffee (N=29): Historical data about the coffee consumption in the United States. The variable $consumption$ is measured in thousand 60-kg bags. The variable $price$ represents the retail price of roasted coffee in US Dollars per pound. Both time series are obtained from the International Coffee Association. The variables $rdi$ and $cpi$ represent real disposable income and the consumer price index, respectively. Those data series are obtained from the St. Louis FRED database. The same is true for the variable $population$. The variables $rprice$ and $pcconsumption$ represent the real coffee prices and the per capita consumption of coffee.
compactcars (N=78): Fuel efficiency of compact cars in 1995 and 2015. The fuel efficiency is expressed in miles per gallon in the columns $automatic$ and $manual$, respectively. The variable $displ$ represents the displacement in liters.
covid: Daily time series of COVID-19 $cases$ and $death$ (cumulative count). The variables $newcases$ and $newdeaths$ correspond to the additional cases each day. The number of observations is variable since it is updated on a daily basis. The data can be obtained from a New York Times GitHub account.
cps7885 (N=1084): Data from the Current Population Survey (CPS) for the year 1978 and 1985. The data accompanies the book Introductory Econometrics: A Modern Approach by Jeffrey M. Wooldridge. It contains the following variables: $educ$ (years of schooling), $south$ (1 if live in south), $nonwhite$ (1 if nonwhite), $female$ (1 if female), $married$ (1 if married), $exper$, $age$, $union$ (1 if belong to union), $wage$, and $year$ (78 or 85).
crime (N=99): Crime rates in North Carolina counties.
discharge (N=7): Generic data set of pollutant discharge into a river measured in gallons.
eggweights (N=61): Weight in grams of Large Grade A Brown Eggs from Whole Foods.
eucrime: Intentional homicide data in Europe from Eurostat.
evdata (N=579): Data about the choice of consumers with respect to alternative fuel vehicles. The variable $choice$ represents the choice by the consumer for gasoline vehicles (1), conventional hybrids (2), plug-in hybrids (3), and electric vehicles (4). For each consumer, you have the following variables: $age$, $level2$ (indicating whether people have a fast charger for electric cars in their community), $female$, and $numcars$ (number of cars). The variable $income$ is coded as follows: Under $15,000 (1), $15,000 to $24,999 (2), $25,000 to $34, 999 (3), $35,000 to $49,999 (4), $50,000 to $74,999 (5), $75,000 to $99,999 (6), $100,000 to $149,999 (7), $150,000 to $199,999 (8), $200,000 to $249,000 (9), above $250,000 (10). The variable $edu$ is coded as follows: Less than High School (1), High School / GED (2), Some College (3), 2-year College Degree (4), 4-year College Degree (5), Masters Degree (6), Doctoral Degree (7). The variable $politics$ is coded as follows: Extremely Liberal (1), Liberal (2), Slightly liberal (3), Moderate (4), Slightly conservative (5), Conservative (6), Extremely conservative (7), Other (8), None (9). See Dumortier et al. (2014) for more details.
exemptorgs (N=42135): Tax exempt organizations in Indiana (Source: IRS Exempt Organizations Business Master File Extract (EO BMF)) updated on 11 December 2023. The variables are $name$, $state$, $deductability$ (1 meaning contributions are deductible, 2 meaning contributions are not deductible, and 4 meaning contributions are deductible by treaty, which may refer to foreign organizations), $organization$ (1-Corporation, 2-Trust, 3-Cooperative, 4-Partnership, 5-Association). The variables $assets$, $income$, and $revenue$ are measured in USD. The variable $ntee$ refers to the National Taxonomy of Exempt Entities and may be missing for some older organizations. A full list is available here.
ez (N=107): This data is taken from Introductory Econometrics: A Modern Approach by Jeffrey M. Wooldridge. The variable $uclms$ represents the unemployment claims in Anderson (Indiana). The variable $ez$ indicates the presence of the enterprise zone in the city. An enterprise zone benefits from tax breaks and other public support to spur economic activity.
factor (N=3199): The data is associated with the analysis in Lane et al. (2018). On a scale from 1–5, respondents had to rank the importance of various vehicle attributes.
fair (N=601): The data comes from the analysis presented in the 1978 paper A Theory of Extramarital Affairs by Ray C. Fair. The variable $affairs$ representing the number of extramarital affairs in the past year. Everything of seven and above is coded as 7. The other variables are $male$, $yearsmarried$ (number of years married), $children$, $religious$ (religiousness on a scale of 1-5 with 1 being basically an atheist), and $marriagehappiness$ (self-rating of marriage with 1=very unhappy to 5=very happy).
faithful (N=272): Eruption and waiting in time minutes of Old Faithful Geyser. This data is also included in R.
fatalcity (N=8131): Each row represents a car accident with at least one fatality according to the Fatality Analysis Reporting System (FARS). The accidents occurred in counties (identified by their $fips$ code) comprising a city. Except $year$, all other variables are dummy variables indicating characteristics associated with the accident.
fatalstate (N=26893): Each row represents a car accident with at least one fatality according to the Fatality Analysis Reporting System (FARS). The accidents occurred in counties (identified by their $fips$ code) in various states. The variable $state$ identifies the FIPS code of the state.
fertil1 (N=1129): Data from the General Social Survey. The data set is accompanying the book Introductory Econometrics: A Modern Approach by Jeffrey Wooldridge. The variables are as follows: $year$ (72 to 84, even), $educ$ (years of schooling), $meduc$ and $feduc$ (mother’s and father’s education), $kids$ (number children ever born), $east$, $northcentral$, and $west$ (equal to 1 if lived in at 16), $farm$ (equal to 1 if on farm at 16), $otherrural$ (equal to 1 if other rural at 16), $town$ (equal to 1 if lived in town at 16), and $smallcity$ (equal to 1 if in small city at 16).
fetransmission (N=9297): Fuel efficiency of cars from 1984 and 2023. This data is similar to compactcars except for all years and all vehicle classes. The fuel efficiency is expressed in miles per gallon in the columns $automatic$ and $manual$, respectively. The variable $displ$ represents the displacement in liters.
fpdata (N=176):
gqtestdata (N=50): Generic (heteroscedastic) home value data of $price$ and living area ($sqft$).
grunfeld (N=200): This is a standard data set to introduce panel data models. It represents investment data from ten firms over the years from 1935 to 1954. The variables are $firm$, $year$, $inv$ (gross investment), $value$ (value of the firm), and $capital$ (stock of plant and equipment).
gss (N=2867): 2016 data from the General Social Survey with the variables that are mostly self-explanatory. The variables $wrkstat$ and $cappun$ refer to the Work status and the stance on the death penalty, respectively. The variables $facebook$ and $instagrm$ indicates if the respondent uses the respective social media platform.
happy (N=631): Data from the 2018 General Social Survey about happiness which includes the following variables: $sexfreq$: Frequency of sex during last year, $gun$: Have gun in home, $sclass$: Subjective class identification, $health1$: Condition of health, $happiness$: General happiness, $party$: Political party affiliation. You may want to combine the “Ind, near democrat” and “Not str democrat” into the same category, e.g., “lean democrat.” Do the same for Republicans., $education$: Highest year of school completed, $age$: Age of respondent
hdi (N=189): Human Development Indicator from 2019.
heating (N=900): R package mlogit. There are five types of heating systems: gas central ($gc$), gas room ($gr$), electric central ($ec$), electric room ($er$), and heat pump ($hp$). There are also the following variables: $idcase$ gives the observation number (1-900), $depvar$ identifies the chosen alternative (gc, gr, ec, er, hp), $ic.alt$ is the installation cost for the 5 alternatives, $oc.alt$ is the annual operating cost for the 5 alternatives, $income$ is the annual income of the household, $agehed$ is the age of the household head, $rooms$ is the number of rooms in the house, $region$ a factor with levels $ncostl$ (northern coastal region), $scostl$ (southern coastal region), $mountn$ (mountain region), $valley$ (central valley region).
henning (N=194): The data set henning is similar to rossi and comes from the paper Cognitive-behavioral treatment of incarcerated offenders: An evaluation of the Vermont Department of Corrections’ Cognitive Self-Change Program by Henning and Frueh (1996). The following variables will be used in this analysis: $months$ (day of re-arrest measured in months), censor (dummy variable ), personal (offense against a person), property (offense against property), and cage (centered age at time of release, i.e., age minus average age).
hhpub (N=129696): The data contains household data from the 2017 National Household Travel Survey (NHTS). Some columns were deleted but the detailed code book can be found here. The variables included are: $homeown$ (home ownership), $hhsize$ (count of household members), $hhvehcnt$ (count of household vehicles), $hhfaminc$ (household income), $bike$ (frequency of bicycle use for travel), $price$ (price of gasoline affects travel), $place$ (travel is financial burden), $ptrans$ (public transporation to reduce financial burden of travel), $hhstate$ (household state), $urbrur$ (household in urban/rural area), and $hbppopdn$ (category of population density (persons per square mile) in the census block group of the household’s home location.
honda (N=81): Prices and miles of used Honda Accords in the Indianapolis area.
housing: Data from FRED
housing1 (N=88): This data is taken from Introductory Econometrics: A Modern Approach by Jeffrey M. Wooldridge. The variables are the following: $price$ (house price in USD 1000), $assess$ (assessed value in USD 1000), $bdrms$ (number of bedrooms), $lotsize$ (in square feet), $sqrft$ (home size in square feet), and $colonial$ (equal to 1 if house is in colonial style).
hprice2 (N=506): This data is taken from Introductory Econometrics: A Modern Approach by Jeffrey M. Wooldridge. The variables are the following: $price$ (median housing price in USD), $crime$ (crimes committed per capita), $nox$ (nitrogen oxide concentration), $rooms$ (number of rooms), $dist$ (weighted distance to five employment centers), $radial$ (accessibility index highways), $proptax$ (property tax per USD 1000), $stratio$ (average student-teacher ratio), and $lowstat$ (percent of people of lower socioeconomic status).
hybrid (N=1000): This is a generic data set with the following variables: $y$ (equal to 1 if person purchased a hybrid car), $gas$ price of gasoline at the time of purchase, $increment$ (price difference compared to gasoline car), $college$ as dummy variable indicating if person has a college degree, and $env$ as a dummy variable indicating whether person is associated with an environmental group.
indyheating (N=24): Contains 24 observations of natural gas usage (in hundreds of cubic feet) and the average outside temperature (in Fahrenheit).
indyhomes (N=102): This data set contains home characteristics from two ZIP codes in Indianapolis.
jcars (N=43): The data is taken from Forecasting: Methods and Applications by Makridakis, Wheelwright, and Hyndman (1998). It contains Japanese motor vehicle production in thousand (1947–1989).
kiel (N=321): This data is taken from Introductory Econometrics: A Modern Approach by Jeffrey M. Wooldridge. It includes the following variables: $year$ (1978 or 1981), $age$ (age of house), $nbh$ (neighborhood, 1-6), $cbd$ (distance to central business district in feet), $intst$ (distance to interstate in feet), $price$ (selling price), $rooms$ (number of rooms in house), $area$ (square footage of house), $land$ (square footage lot), $baths$ (number of bathrooms), $dist$ (distance from house to incinerator in feet), $wind$ (percent time wind from incinerator to house), $y81$ (equal to 1 if year is 1981), $nearinc$ (equal 1 if distance is less than 15840 feet), $rprice$ (real price in 1978 dollars).
lung (N=228): The data is associated with the R package survival and includes the survival of patients with lung cancer. The variables included are the following: $inst$ (institution code), $time$ (survival time in days), $phecog$ (ECOG performance score as rated by the physician: 0=asymptomatic, 1=symptomatic but completely ambulatory, 2=in bed less than 50% of the day, 3=in bed more than 50% of the day but not bed bound, 4 = bed bound), $phkarno$ (Karnofsky performance score from 0=bad to 100=good as rated by physician), $patkarno$ (Karnofsky performance score as rated by the patient), $mealcal$ (calories consumed at meals) and $wtloss$ (weight loss in last six months in pounds).
meatdemand (N=39): Meet demand quantity ($q$) and real price ($p$) data from the U.S. Department of Agriculture. Prices and real disposable income ($rdi$) are in real terms.
mh1 (N=101): Home values in the Meridian Hills area in Indianapolis.
mh2 (N=18): Prices and characteristics of homes in the Meridian Hills area in Indianapolis.
milk (N=50): Randomly generated milk container fillings in ounces.
mpa (N=18): Randomly generated exam scores in a class of MPA students.
nfl (N=1009): Data from Berri et al. (2011) which includes the following variables: $total$ (total salary), $yards$ (passing yards from the prior season), $att$ (pass attempts), $exp$ (total years of experience in the league), $exp2$ (total years of experience in the league squared), $draft1$ (first round draft pick), $draft2$ (second round draft pick), $veteran$ (bargaining status changes after a player has completed three years in the NFL), $changeteam$ (player has changed team), $pbowlever$ (player appeared in the Pro Bowl), and $symm$ (facial symmetry).
ohioincome (N=607): Enrollment and median income in Ohio School districts for 2018/2019. $IRN$ identifies the school district.
ohioscore (N=608): Performance and achievement scores of Ohio schools for 2018/2019. The data is obtained from the Ohio Department of Education. $IRN$ identifies the school district.
organic (N=100): Randomly generated data on the purchase behavior of organic food (binary choice) as a function of income.
quakes (N=99): Time series of the annual number of earthquakes in the world with seismic magnitude over 7.0, for 99 consecutive years. The data is taken from Penn State STAT 510 Applied Time Series Analysis.
quillayute (N=20116): Weather data from Quillayute Airport in Washington State. Weather station USW00094240. The variables are $date$, $temperature$ (average between maximum and minimum temperature in degree Fahrenheit), and $month$ (month of the year).
retail (N=348): Advance Retail Sales: Retail (Excluding Food Services) obtained from the St. Louis FRED.
rossi (N=432): Experimental recidivism study on 432 male prisoners over a period of one year after release from prison (Rossi et al., 1980). The following variables are included: $week$ (week of first arrest after release, or censoring time), $arrest$ (the event indicator, equal to 1 for those arrested during the period of the study and 0 for those who were not arrested), $fin$ (a factor, with levels yes if the individual received financial aid after release from prison, and no if he did not; financial aid was a randomly assigned factor manipulated by the researchers), $age$ (in years at the time of release), $race$ (a factor with levels black and other), $wexp$ (a factor with levels yes if the individual had full-time work experience prior to incarceration and no if he did not), $mar$ (a factor with levels married if the individual was married at the time of release and not married if he was not), $paro$ (a factor coded yes if the individual was released on parole and no if he was not), $prio$ (number of prior convictions), $educ$ (education, a categorical variable coded numerically, with codes 2 (grade 6 or less), 3 (grades 6 through 9), 4 (grades 10 and 11), 5 (grade 12), or 6 (some post-secondary)), and $emp1$ to $emp52$ (factors coded yes if the individual was employed in the corresponding week of the study and no otherwise).
skewness (N=500): Randomly generated data from a beta distribution with various parameters. $beta55$ ($a=5$ and $b=5$), $beta28$ ($a=2$ and $b=8$), and $beta82$ ($a=8$ and $b=2$).
soda (N=25): Randomly generated data of for soda cans (in milliliters).
souvenirs (N=84): The data is taken from Forecasting: Methods and Applications by Makridakis, Wheelwright, and Hyndman (1998). It contains monthly sales for a souvenir shop on the wharf at a beach resort town in Queensland, Australia. The data is also contained in the R package fma as fancy.
statefinhealth (N=51): Data taken from the Urban Institute’s Financial Health and Wealth Dashboard 2022. The data includes the following variables: $state$, $delinquentdebt$ (residents with delinquent debt), $emergencysavings$ (Households with at least USD 2000 in emergency savings), $mediancreditscore$, $mortgageforeclosure$ (mortgage holders in foreclosure), and $mediannetworth$.
states (N=10): Generic income data for three states.
traffic (N=108): This data is taken from Introductory Econometrics: A Modern Approach by Jeffrey M. Wooldridge. It contains the following variables: $year$, $month$, $totacc$ (statewide total accidents), $t$ (time trend), $unem$ (state unemployment rate), $spdlaw$ (equal to 1 after 65 mph in effect), $bltlaw$ (equal to 1 after seatbelt law), and $wkends$ (number of weekends in month).
truncation (n=50): Generic truncated data for which the researcher only observes $yobs$ as the independent variable and $x$ as the dependent variable. $yreal$ are the (partially unobserved) values of the dependent variable.
usdata (N=307): Data from the St. Louis FRED data base of real disposable personal income per capita ($income$) and real personal consumption expenditures per capita ($consumption$).
vehicles (N=45885): The data set contains data about the fuel efficiency of vehicles from model year 1984 to 2021 ($year$). It also includes $make$ and $model$ of the various manufacturers. The variable $displ$ refers to engine displacement in liters, $vclass$ is the vehicle class, $co2tailpipegpm$ indicates the greenhouse gas (GHG) emissions, and $comb08$ is combined MPG. The other variables are self-explanatory. The data is obtained from the U.S. Environmental Protection Agency’ (EPA) Fuel Economy Data webpage. The full data set contains many more variables which are well documented here.
vehpub (N=256115): The data contains vehicle data from the 2017 National Household Travel Survey (NHTS). Some columns were deleted but the detailed code book can be found here. The variables included are: $houseid$ (household identifier), $vehid$ (vehicle identifier), $vehyear$ (vehicle year), $make$ (vehicle make), $fueltype$ (fuel type), $vehtype$ (vehicle type), $odread$ (odometer reading), $hfuel$ (type of hybrid vehicle), $hybrid$ (hybrid vehicle), $homeown$ (home ownership), $hhfaminc$ (household income), $hhstate$ (household state), $urbrur$ (household in urban/rural area).
wage (N=526): This data is taken from Introductory Econometrics: A Modern Approach by Jeffrey M. Wooldridge. It contains the following variables: $income$ (average hourly earnings), $educ$ (years of education), and $exper$ (years potential experience).
wage2 (N=935): This data is taken from Introductory Econometrics: A Modern Approach by Jeffrey M. Wooldridge. It contains the following variables: $wage$ (monthly earnings), $hours$ (average weekly hours), $kww$ knowledge of world work score, $educ$ (years of education), $exper$ (years of work experience), $tenure$ (years with current employer), $age$ (age in years), $married$ (equal to 1 if married), $black$ (equal to 1 if black), $south$ (equal to 1 if live in south), $urban$ (equal to 1 if live in SMSA), $sibs$ (number of siblings), $brthord$ (birth order), $meduc$ (mother’s education), $feduc$ (father’s education).
waterpressure (N=30): Randomly generated data on the water pressure in a city’s water lines.
wdi (N=13671): World Development Indicators (WDI) from the World Bank. The following variables are included in the data set: Country code ($iso2c$), $country$, $year$, gross domestic product per capita in constant 2010 USD ($gdp$), life expectancy at birth in years ($lifeexp$), adult literacy rate (percent of people ages 15 and above) ($litrate$), fertility rate (births per woman) ($fertrate$), mortality rate under age of five (per 1,000 live births) ($mortrate$), geographic region ($region$), and income groupings based on gross national income per capita ($incomeLevel$).
windsolar (N=648): The data is taken from the paper The effect of the feed-in-system policy on renewable energy investments: Evidence from the EU countries by Alolo et al. (2020). The data contains the following variables: $country$, $year$, added wind capacity in Megawatts ($agwind$), presence of either a feed-in-premium (FIP) or a feed-in-tariff (FIT) ($fiswind$), presence of a FIT ($fitwind$), presence of a FIP ($fipwind$), existence of a quota trade system ($quota$), existence of tax benefits ($tax$), existence of a $tender$ mechanism, existence of an electricity price $cap$, gross domestic product ($gdp$), per capita $electricityconsumption$, percent of electricity from various sources ($coalshare$, $renewableshare$, $oilshare$, $nuclearshare$, and $naturalgasshare$), population growth ($popgrowth$), per capita carbon emissions $co2$, prices of various energy sources ($oilprice$, $naturalgasprice$, and $coalprice$), energy dependence defined as energy imports divided by total energy consumption ($energyimport$).