Panel Study of Income Dynamics (PSID)

Build Status Build status

The Panel Study of Income Dynamics is the longest running longitudinal household survey in the world.

  • One cross-year individual with one record per respondent in participating household, many family data tables with one record per family per timepoint.

  • A complex sample survey designed to generalize to residents of the United States.

  • Released either annually or biennially since 1968.

  • Administered by the University of Michigan’s Institute for Social Research and funded by consortium.

Simplified Download and Importation

The R lodown package easily downloads and imports all available PSID microdata by simply specifying "psid" with an output_dir = parameter in the lodown() function. Depending on your internet connection and computer processing speed, you might prefer to run this step overnight.

library(lodown)
lodown( "psid" , output_dir = file.path( path.expand( "~" ) , "PSID" ) , 
    your_email = "email@address.com" , 
    your_password = "password" )

lodown also provides a catalog of available microdata extracts with the get_catalog() function. After requesting the PSID catalog, you could pass a subsetted catalog through the lodown() function in order to download and import specific extracts (rather than all available extracts).

library(lodown)
# examine all available PSID microdata files
psid_cat <-
    get_catalog( "psid" ,
        output_dir = file.path( path.expand( "~" ) , "PSID" ) , 
        your_email = "email@address.com" , 
        your_password = "password" )

# download the microdata to your local computer
lodown( "psid" , psid_cat , 
    your_email = "email@address.com" , 
    your_password = "password" )

Analysis Examples with the survey library

Construct a complex sample survey design:

options( survey.lonely.psu = "adjust" )

library(survey)

# identify the cross-year individual filename
cross_year_individual_rds <- 
    grep( 
        "cross-year individual" ,
        list.files( 
            file.path( path.expand( "~" ) , "PSID" ) , 
            recursive = TRUE , 
            full.names = TRUE 
        ) ,
        value = TRUE
    )

individual_df <- readRDS( cross_year_individual_rds )

ind_variables_to_keep <-
    c( 
        'one' ,         # column with all ones
        'er30001' ,     # 1968 interview number
        'er30002' ,     # 1968 person number
        'er31997' ,     # primary sampling unit variable
        'er31996' ,     # stratification variable
        'er33802' ,     # sequence number, 2005
        'er34302' ,     # sequence number, 2015
        'er32000' ,     # sex
        'er34305' ,     # age in 2015
        'er33813' ,     # employment status in 2005
        'er34317' ,     # employment status in 2015
        'er33848' ,     # 2005 longitudinal weight
        'er34413'       # 2015 longitudinal weight
    )

    

individual_df <- individual_df[ ind_variables_to_keep ] ; gc()

family_2005_df <- 
    readRDS( file.path( path.expand( "~" ) , "PSID" , "family files/2005.rds" ) )

fam_2005_variables_to_keep <- 
    c( 
        'er25002' , # 2005 interview number
        'er28037'   # 2005 total family income
        
    )

family_2005_df <- family_2005_df[ fam_2005_variables_to_keep ] ; gc()

family_2015_df <- 
    readRDS( file.path( path.expand( "~" ) , "PSID" , "family files/2015.rds" ) )

fam_2015_variables_to_keep <-
    c( 
        'er60002' , # 2015 interview number
        'er65349'   # 2015 total family income
    )

family_2015_df <- family_2015_df[ fam_2015_variables_to_keep ] ; gc()

ind_fam_2005 <- 
    merge( 
        individual_df , 
        family_2005_df , 
        by.x = 'er33802' , 
        by.y = 'er25002' 
    )

ind_fam_2015 <- 
    merge( 
        individual_df , 
        family_2015_df , 
        by.x = 'er34302' , 
        by.y = 'er60002' 
    )

psid_df <- merge( ind_fam_2005 , ind_fam_2015 , all = TRUE )

psid_design <- 
    svydesign( 
        ~ er31997 , 
        strata = ~ er31996 , 
        data = psid_df , 
        weights = ~ er33848 , 
        nest = TRUE 
    )

Variable Recoding

Add new columns to the data set:

psid_design <- 
    update( 
        psid_design , 
        
        employment_2005 =
            factor( er33813 , levels = 1:8 ,
                labels = c( 'working now' , 'only temporarily laid off' ,
                'looking for work, unemployed' , 'retired' , 'permanently disabled' ,
                'housewife; keeping house' , 'student' , 'other' )
            ) ,
            
        employed_in_2015 = 
            factor( er34317 , levels = 1:8 ,
                labels = c( 'working now' , 'only temporarily laid off' ,
                'looking for work, unemployed' , 'retired' , 'permanently disabled' ,
                'housewife; keeping house' , 'student' , 'other' )
            ) ,
            
        female = as.numeric( er32000 == 2 )

    )

Unweighted Counts

Count the unweighted number of records in the survey sample, overall and by groups:

sum( weights( psid_design , "sampling" ) != 0 )

svyby( ~ one , ~ employment_2005 , psid_design , unwtd.count )

Weighted Counts

Count the weighted size of the generalizable population, overall and by groups:

svytotal( ~ one , psid_design )

svyby( ~ one , ~ employment_2005 , psid_design , svytotal )

Descriptive Statistics

Calculate the mean (average) of a linear variable, overall and by groups:

svymean( ~ er28037 , psid_design , na.rm = TRUE )

svyby( ~ er28037 , ~ employment_2005 , psid_design , svymean , na.rm = TRUE )

Calculate the distribution of a categorical variable, overall and by groups:

svymean( ~ employed_in_2015 , psid_design , na.rm = TRUE )

svyby( ~ employed_in_2015 , ~ employment_2005 , psid_design , svymean , na.rm = TRUE )

Calculate the sum of a linear variable, overall and by groups:

svytotal( ~ er28037 , psid_design , na.rm = TRUE )

svyby( ~ er28037 , ~ employment_2005 , psid_design , svytotal , na.rm = TRUE )

Calculate the weighted sum of a categorical variable, overall and by groups:

svytotal( ~ employed_in_2015 , psid_design , na.rm = TRUE )

svyby( ~ employed_in_2015 , ~ employment_2005 , psid_design , svytotal , na.rm = TRUE )

Calculate the median (50th percentile) of a linear variable, overall and by groups:

svyquantile( ~ er28037 , psid_design , 0.5 , na.rm = TRUE )

svyby( 
    ~ er28037 , 
    ~ employment_2005 , 
    psid_design , 
    svyquantile , 
    0.5 ,
    ci = TRUE ,
    keep.var = TRUE ,
    na.rm = TRUE
)

Estimate a ratio:

svyratio( 
    numerator = ~ er28037 , 
    denominator = ~ er65349 , 
    psid_design ,
    na.rm = TRUE
)

Subsetting

Restrict the survey design to senior in 2015:

sub_psid_design <- subset( psid_design , er34305 >= 65 )

Calculate the mean (average) of this subset:

svymean( ~ er28037 , sub_psid_design , na.rm = TRUE )

Measures of Uncertainty

Extract the coefficient, standard error, confidence interval, and coefficient of variation from any descriptive statistics function result, overall and by groups:

this_result <- svymean( ~ er28037 , psid_design , na.rm = TRUE )

coef( this_result )
SE( this_result )
confint( this_result )
cv( this_result )

grouped_result <-
    svyby( 
        ~ er28037 , 
        ~ employment_2005 , 
        psid_design , 
        svymean ,
        na.rm = TRUE 
    )
    
coef( grouped_result )
SE( grouped_result )
confint( grouped_result )
cv( grouped_result )

Calculate the degrees of freedom of any survey design object:

degf( psid_design )

Calculate the complex sample survey-adjusted variance of any statistic:

svyvar( ~ er28037 , psid_design , na.rm = TRUE )

Include the complex sample design effect in the result for a specific statistic:

# SRS without replacement
svymean( ~ er28037 , psid_design , na.rm = TRUE , deff = TRUE )

# SRS with replacement
svymean( ~ er28037 , psid_design , na.rm = TRUE , deff = "replace" )

Compute confidence intervals for proportions using methods that may be more accurate near 0 and 1. See ?svyciprop for alternatives:

svyciprop( ~ female , psid_design ,
    method = "likelihood" )

Regression Models and Tests of Association

Perform a design-based t-test:

svyttest( er28037 ~ female , psid_design )

Perform a chi-squared test of association for survey data:

svychisq( 
    ~ female + employed_in_2015 , 
    psid_design 
)

Perform a survey-weighted generalized linear model:

glm_result <- 
    svyglm( 
        er28037 ~ female + employed_in_2015 , 
        psid_design 
    )

summary( glm_result )

Analysis Examples with srvyr

The R srvyr library calculates summary statistics from survey data, such as the mean, total or quantile using dplyr-like syntax. srvyr allows for the use of many verbs, such as summarize, group_by, and mutate, the convenience of pipe-able functions, the tidyverse style of non-standard evaluation and more consistent return types than the survey package. This vignette details the available features. As a starting point for PSID users, this code replicates previously-presented examples:

library(srvyr)
psid_srvyr_design <- as_survey( psid_design )

Calculate the mean (average) of a linear variable, overall and by groups:

psid_srvyr_design %>%
    summarize( mean = survey_mean( er28037 , na.rm = TRUE ) )

psid_srvyr_design %>%
    group_by( employment_2005 ) %>%
    summarize( mean = survey_mean( er28037 , na.rm = TRUE ) )

Replication Example