Subsetting General Social Survey (GSS) data
Nathan Alexander
Center for Applied Data Science and Analytics (CADSA)
Howard University
2024-11-17
Overview
This file will help you subset data from the gssr
package. The gssr R package provides a useful set of tools
to efficiently access, manipulate, and analyze the General Social Survey
(GSS) dataset. The package simplifies the process of working with
complex survey data by offering a set of functions for subsetting,
filtering, and transforming your desired GSS data.
The package gssr is located here.
Set up your work enviornment
Open up a new .Rmd file.
Use {r setup, include=F} in your first code chunk.
## Warning: package 'gssr' was built under R version 4.4.2## Package loaded. To attach the GSS data, type data(gss_all) at the console.
## For the codebook, type data(gss_dict).
## For the panel data and documentation, type e.g. data(gss_panel08_long) and data(gss_panel_doc).
## For help on a specific GSS variable, type ?varname at the console.##
## Attaching package: 'dplyr'## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, unionWe will run through a sample analysis using the gssr
package.
Explore
We explore variables in the GSSR documentation file
# load the master documentation files
data(gss_all) # note that this is a large file of all GSS data
data(gssrdoc) # this is the documentation for the GSS dataExamine
We examine the years available for any given variable.
# check which years your variables are available
gss_which_years(gss_all, race)## # A tibble: 34 × 2
## year race
## <dbl+lbl> <lgl>
## 1 1972 TRUE
## 2 1973 TRUE
## 3 1974 TRUE
## 4 1975 TRUE
## 5 1976 TRUE
## 6 1977 TRUE
## 7 1978 TRUE
## 8 1980 TRUE
## 9 1982 TRUE
## 10 1983 TRUE
## # ℹ 24 more rowsgss_which_years(gss_all, sex)## # A tibble: 34 × 2
## year sex
## <dbl+lbl> <lgl>
## 1 1972 TRUE
## 2 1973 TRUE
## 3 1974 TRUE
## 4 1975 TRUE
## 5 1976 TRUE
## 6 1977 TRUE
## 7 1978 TRUE
## 8 1980 TRUE
## 9 1982 TRUE
## 10 1983 TRUE
## # ℹ 24 more rowsgss_which_years(gss_all, hrs2)## # A tibble: 34 × 2
## year hrs2
## <dbl+lbl> <lgl>
## 1 1972 FALSE
## 2 1973 TRUE
## 3 1974 TRUE
## 4 1975 TRUE
## 5 1976 TRUE
## 6 1977 TRUE
## 7 1978 TRUE
## 8 1980 TRUE
## 9 1982 TRUE
## 10 1983 TRUE
## # ℹ 24 more rowsgss_which_years(gss_all, educ)## # A tibble: 34 × 2
## year educ
## <dbl+lbl> <lgl>
## 1 1972 TRUE
## 2 1973 TRUE
## 3 1974 TRUE
## 4 1975 TRUE
## 5 1976 TRUE
## 6 1977 TRUE
## 7 1978 TRUE
## 8 1980 TRUE
## 9 1982 TRUE
## 10 1983 TRUE
## # ℹ 24 more rows# when you want to return information for multiple variables
gss_all %>%
gss_which_years(c(race, sex, hrs2, educ)) %>%
print(n = Inf)## # A tibble: 34 × 5
## year race sex hrs2 educ
## <dbl+lbl> <lgl> <lgl> <lgl> <lgl>
## 1 1972 TRUE TRUE FALSE TRUE
## 2 1973 TRUE TRUE TRUE TRUE
## 3 1974 TRUE TRUE TRUE TRUE
## 4 1975 TRUE TRUE TRUE TRUE
## 5 1976 TRUE TRUE TRUE TRUE
## 6 1977 TRUE TRUE TRUE TRUE
## 7 1978 TRUE TRUE TRUE TRUE
## 8 1980 TRUE TRUE TRUE TRUE
## 9 1982 TRUE TRUE TRUE TRUE
## 10 1983 TRUE TRUE TRUE TRUE
## 11 1984 TRUE TRUE TRUE TRUE
## 12 1985 TRUE TRUE TRUE TRUE
## 13 1986 TRUE TRUE TRUE TRUE
## 14 1987 TRUE TRUE TRUE TRUE
## 15 1988 TRUE TRUE TRUE TRUE
## 16 1989 TRUE TRUE TRUE TRUE
## 17 1990 TRUE TRUE TRUE TRUE
## 18 1991 TRUE TRUE TRUE TRUE
## 19 1993 TRUE TRUE TRUE TRUE
## 20 1994 TRUE TRUE TRUE TRUE
## 21 1996 TRUE TRUE TRUE TRUE
## 22 1998 TRUE TRUE TRUE TRUE
## 23 2000 TRUE TRUE TRUE TRUE
## 24 2002 TRUE TRUE TRUE TRUE
## 25 2004 TRUE TRUE TRUE TRUE
## 26 2006 TRUE TRUE TRUE TRUE
## 27 2008 TRUE TRUE TRUE TRUE
## 28 2010 TRUE TRUE TRUE TRUE
## 29 2012 TRUE TRUE TRUE TRUE
## 30 2014 TRUE TRUE TRUE TRUE
## 31 2016 TRUE TRUE TRUE TRUE
## 32 2018 TRUE TRUE TRUE TRUE
## 33 2021 TRUE TRUE TRUE TRUE
## 34 2022 TRUE TRUE TRUE TRUE# loading single year data
gss12 <- gss_get_yr(2012)## Fetching: https://gss.norc.org/documents/stata/2012_stata.zipgss22 <- gss_get_yr(2022)## Fetching: https://gss.norc.org/documents/stata/2022_stata.zip# View(gss22)# select your variables for a single year using cross sectional file
gss12 %>%
select(race, sex, hrs2, educ, wtssall)## # A tibble: 1,974 × 5
## race sex hrs2 educ wtssall
## <dbl+lbl> <dbl+lbl> <dbl+lbl> <dbl+lbl> <dbl+lbl>
## 1 1 [white] 1 [male] NA(i) [iap] 16 [4 years of college] 2.62
## 2 1 [white] 1 [male] NA(i) [iap] 12 [12th grade] 3.50
## 3 3 [other] 1 [male] NA(i) [iap] 12 [12th grade] 1.75
## 4 1 [white] 2 [female] NA(i) [iap] 13 [1 year of college] 1.24
## 5 2 [black] 2 [female] NA(i) [iap] 16 [4 years of college] 0.874
## 6 1 [white] 2 [female] NA(i) [iap] 19 [7 years of college] 0.824
## 7 1 [white] 2 [female] NA(i) [iap] 15 [3 years of college] 0.824
## 8 3 [other] 2 [female] NA(i) [iap] 11 [11th grade] 0.412
## 9 2 [black] 2 [female] NA(i) [iap] 9 [9th grade] 0.412
## 10 1 [white] 2 [female] NA(i) [iap] 17 [5 years of college] 0.412
## # ℹ 1,964 more rowsgss12 %>%
select(race, sex, hrs2, educ, wtssall) %>%
count(race, sex, wtssall)## # A tibble: 66 × 4
## race sex wtssall n
## <dbl+lbl> <dbl+lbl> <dbl+lbl> <int>
## 1 1 [white] 1 [male] 0.412 167
## 2 1 [white] 1 [male] 0.824 280
## 3 1 [white] 1 [male] 0.874 34
## 4 1 [white] 1 [male] 1.00 1
## 5 1 [white] 1 [male] 1.19 1
## 6 1 [white] 1 [male] 1.24 46
## 7 1 [white] 1 [male] 1.65 22
## 8 1 [white] 1 [male] 1.75 95
## 9 1 [white] 1 [male] 2.06 7
## 10 1 [white] 1 [male] 2.62 11
## # ℹ 56 more rowsGet data
We use the gss_all to get all of the data set. We can
use a more direct method but we must be careful about our outputs.
gss_all %>%
select(year, race, sex, hrs2, educ, wtssall, wtss, wtssnr) %>%
filter(year == 2014) %>%
drop_na() -> df
sapply(df, function(x) sum(is.na(x))) # take note that the hrs2 variable will cause issue## year race sex hrs2 educ wtssall wtss wtssnr
## 0 0 0 0 0 0 0 0df## # A tibble: 40 × 8
## year race sex hrs2 educ wtssall wtss wtssnr
## <dbl+lbl> <dbl+lbl> <dbl+lbl> <dbl+lbl> <dbl+lbl> <dbl+lbl> <dbl+lbl> <dbl+>
## 1 2014 1 [white] 2 [female] 40 16 0.448 0.448 0.500
## 2 2014 2 [black] 2 [female] 50 14 1.38 1.38 1.54
## 3 2014 3 [other] 2 [female] 50 13 0.448 0.448 0.500
## 4 2014 2 [black] 2 [female] 45 18 0.690 0.690 0.770
## 5 2014 1 [white] 2 [female] 45 17 0.896 0.896 1.00
## 6 2014 2 [black] 2 [female] 32 12 0.690 0.690 0.735
## 7 2014 3 [other] 2 [female] 40 12 2.07 2.07 2.21
## 8 2014 1 [white] 2 [female] 15 12 0.896 0.896 0.954
## 9 2014 1 [white] 2 [female] 60 20 0.896 0.896 0.888
## 10 2014 1 [white] 1 [male] 40 16 0.896 0.896 0.923
## # ℹ 30 more rowsSubset data
# select your variables for a single year using gss_all
gss_all %>%
filter(year == 2012) %>%
select(race, sex, hrs2, educ, wtssall, wtss, wtssnr)## # A tibble: 1,974 × 7
## race sex hrs2 educ wtssall wtss wtssnr
## <dbl+lbl> <dbl+lbl> <dbl+lbl> <dbl+lbl> <dbl+lbl> <dbl+lbl> <dbl+lbl>
## 1 1 [white] 1 [male] NA(i) [iap] 16 2.62 2.62 2.87
## 2 1 [white] 1 [male] NA(i) [iap] 12 3.50 3.50 3.83
## 3 3 [other] 1 [male] NA(i) [iap] 12 1.75 1.75 1.91
## 4 1 [white] 2 [female] NA(i) [iap] 13 1.24 1.24 1.35
## 5 2 [black] 2 [female] NA(i) [iap] 16 0.874 0.874 0.957
## 6 1 [white] 2 [female] NA(i) [iap] 19 0.824 0.824 0.902
## 7 1 [white] 2 [female] NA(i) [iap] 15 0.824 0.824 0.902
## 8 3 [other] 2 [female] NA(i) [iap] 11 0.412 0.412 0.451
## 9 2 [black] 2 [female] NA(i) [iap] 9 0.412 0.412 0.451
## 10 1 [white] 2 [female] NA(i) [iap] 17 0.412 0.412 0.451
## # ℹ 1,964 more rowsgss_all %>%
filter(year == 2018) %>%
select(race, sex, hrs2, educ, wtssall, wtss, wtssnr)## # A tibble: 2,348 × 7
## race sex hrs2 educ wtssall wtss wtssnr
## <dbl+lbl> <dbl+lbl> <dbl+lbl> <dbl+lbl> <dbl+lbl> <dbl+lbl> <dbl+lbl>
## 1 1 [white] 1 [male] 41 14 2.36 2.36 2.75
## 2 1 [white] 2 [female] NA(i) [iap] 10 0.943 0.943 1.10
## 3 1 [white] 1 [male] NA(i) [iap] 16 0.943 0.943 1.10
## 4 1 [white] 2 [female] NA(i) [iap] 16 0.943 0.943 1.10
## 5 2 [black] 1 [male] NA(i) [iap] 18 0.943 0.943 1.10
## 6 1 [white] 2 [female] NA(i) [iap] 16 0.943 0.943 1.10
## 7 2 [black] 2 [female] NA(i) [iap] 13 0.943 0.943 1.10
## 8 1 [white] 1 [male] NA(i) [iap] 12 0.471 0.471 0.551
## 9 1 [white] 2 [female] NA(i) [iap] 8 0.471 0.471 0.551
## 10 1 [white] 1 [male] NA(i) [iap] 12 0.943 0.943 1.10
## # ℹ 2,338 more rowsClean data
# clean data for analysis
gss12 %>%
select(race, sex, hrs2, educ, wtssall, wtss, wtssnr) %>%
drop_na() -> gss12_analysis# check for missing values
sapply(gss12_analysis, function(x) sum(is.na(x)))## race sex hrs2 educ wtssall wtss wtssnr
## 0 0 0 0 0 0 0str(gss12_analysis)## tibble [35 × 7] (S3: tbl_df/tbl/data.frame)
## $ race : dbl+lbl [1:35] 2, 1, 3, 1, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 2, 1, 1...
## ..@ label : chr "race of respondent"
## ..@ format.stata: chr "%29.0g"
## ..@ labels : Named num [1:16] 1 2 3 NA NA NA NA NA NA NA ...
## .. ..- attr(*, "names")= chr [1:16] "white" "black" "other" "don't know" ...
## $ sex : dbl+lbl [1:35] 2, 1, 1, 2, 1, 2, 2, 2, 2, 2, 2, 1, 1, 2, 1, 1, 2, 2, 1...
## ..@ label : chr "respondents sex"
## ..@ format.stata: chr "%29.0g"
## ..@ labels : Named num [1:15] 1 2 NA NA NA NA NA NA NA NA ...
## .. ..- attr(*, "names")= chr [1:15] "male" "female" "don't know" "iap" ...
## $ hrs2 : dbl+lbl [1:35] 25, 40, 40, 50, 45, 45, 38, 32, 30, 40, 35, 35, 16, 35,...
## ..@ label : chr "number of hours usually work a week"
## ..@ format.stata: chr "%29.0g"
## ..@ labels : Named num [1:14] 89 NA NA NA NA NA NA NA NA NA ...
## .. ..- attr(*, "names")= chr [1:14] "89+ hours" "don't know" "iap" "I don't have a job" ...
## $ educ : dbl+lbl [1:35] 16, 16, 12, 16, 12, 14, 14, 12, 14, 14, 9, 10, 14, 18,...
## ..@ label : chr "highest year of school completed"
## ..@ format.stata: chr "%29.0g"
## ..@ labels : Named num [1:34] 0 1 2 3 4 5 6 7 8 9 ...
## .. ..- attr(*, "names")= chr [1:34] "no formal schooling" "1st grade" "2nd grade" "3rd grade" ...
## $ wtssall: dbl+lbl [1:35] 1.748, 0.824, 3.496, 0.874, 0.824, 0.874, 0.874, 0.412,...
## ..@ label : chr "weight variable"
## ..@ format.stata: chr "%29.0g"
## ..@ labels : Named num [1:13] NA NA NA NA NA NA NA NA NA NA ...
## .. ..- attr(*, "names")= chr [1:13] "don't know" "iap" "I don't have a job" "dk, na, iap" ...
## $ wtss : dbl+lbl [1:35] 1.748, 0.824, 3.496, 0.874, 0.824, 0.874, 0.874, 0.412,...
## ..@ label : chr "weight variable"
## ..@ format.stata: chr "%29.0g"
## ..@ labels : Named num [1:13] NA NA NA NA NA NA NA NA NA NA ...
## .. ..- attr(*, "names")= chr [1:13] "don't know" "iap" "I don't have a job" "dk, na, iap" ...
## $ wtssnr : dbl+lbl [1:35] 1.761, 0.845, 3.740, 0.807, 0.887, 0.895, 1.110, 0.408,...
## ..@ label : chr "weight variable"
## ..@ format.stata: chr "%29.0g"
## ..@ labels : Named num [1:13] NA NA NA NA NA NA NA NA NA NA ...
## .. ..- attr(*, "names")= chr [1:13] "don't know" "iap" "I don't have a job" "dk, na, iap" ...# exploratory analysis
gss12 %>%
select(race, sex, hrs2, educ, wtssall, wtss, wtssnr) %>%
count(race, sex)## # A tibble: 6 × 3
## race sex n
## <dbl+lbl> <dbl+lbl> <int>
## 1 1 [white] 1 [male] 673
## 2 1 [white] 2 [female] 804
## 3 2 [black] 1 [male] 114
## 4 2 [black] 2 [female] 187
## 5 3 [other] 1 [male] 98
## 6 3 [other] 2 [female] 98gss12_analysis %>%
count(race, sex)## # A tibble: 5 × 3
## race sex n
## <dbl+lbl> <dbl+lbl> <int>
## 1 1 [white] 1 [male] 9
## 2 1 [white] 2 [female] 19
## 3 2 [black] 2 [female] 5
## 4 3 [other] 1 [male] 1
## 5 3 [other] 2 [female] 1Model 1
model <- lm(gss12_analysis$hrs2 ~ gss12_analysis$educ, weight = gss12_analysis$wtssall)
summary(model)##
## Call:
## lm(formula = gss12_analysis$hrs2 ~ gss12_analysis$educ, weights = gss12_analysis$wtssall)
##
## Weighted Residuals:
## Min 1Q Median 3Q Max
## -21.953 -5.699 1.695 5.578 17.203
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 41.8008 8.0522 5.191 1.05e-05 ***
## gss12_analysis$educ -0.4010 0.6014 -0.667 0.509
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 10.01 on 33 degrees of freedom
## Multiple R-squared: 0.0133, Adjusted R-squared: -0.0166
## F-statistic: 0.4447 on 1 and 33 DF, p-value: 0.5095Plot the variable relationships
Here, I use some dated code, namely the base R abline() function discussed in class.
# Plot the regression line
plot(gss12_analysis$educ, gss12_analysis$hrs2)
# Some alternate code
plot(gss12_analysis$educ, gss12_analysis$hrs2,
main="Education vs Hours Worked",
xlab="Years of Education",
ylab="Hours Worked")
# Add the regression line
abline(lm(hrs2 ~ educ, data = gss12_analysis), col = "red")
Here, I use some more frequent code with additional parameters given that we’ve made a ggplot object. The code also provides some labels to our plot.
Plot the variable relationships
# Create a basic scatter plot
plot(gss12_analysis$educ, gss12_analysis$hrs2,
main="Education vs Hours Worked",
xlab="Years of Education",
ylab="Hours Worked",
pch=16, # Use filled circles for points
col="blue", # Make the points blue
cex=0.8) # Slightly smaller point size
# Add the regression line
abline(model, col = "red", lwd = 2)
# Add a legend
legend("topleft", legend=c("Data points", "Regression line"),
col=c("blue", "red"), pch=c(16, NA), lty=c(NA, 1))
# Add text for R-squared value
r_squared <- summary(model)$r.squared
text(x = max(gss12_analysis$educ), y = min(gss12_analysis$hrs2),
labels = paste("R-squared =", round(r_squared, 3)),
pos = 2)
Examine residuals
We now examine the residuals.
# Check residuals
resids <- residuals(model)
plot(gss12_analysis$hrs2, resids)
And then view the diagnostic plots.
# Diagnostic plots
par(mfrow=c(2,2))
plot(model)
# Check significance of predictor
anova(model)## Analysis of Variance Table
##
## Response: gss12_analysis$hrs2
## Df Sum Sq Mean Sq F value Pr(>F)
## gss12_analysis$educ 1 44.6 44.575 0.4447 0.5095
## Residuals 33 3307.8 100.236Model 2
# run your model and diagnostic plots
model2 <- lm(df$educ ~ df$hrs2)
summary(model2)##
## Call:
## lm(formula = df$educ ~ df$hrs2)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.2179 -2.8213 -0.1369 1.7238 5.1510
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 14.11117 1.68095 8.395 3.48e-10 ***
## df$hrs2 0.01845 0.04141 0.445 0.659
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 3.211 on 38 degrees of freedom
## Multiple R-squared: 0.005195, Adjusted R-squared: -0.02098
## F-statistic: 0.1984 on 1 and 38 DF, p-value: 0.6585pred <- predict(model2)
# Plot the scatterplot
plot(df$educ, df$hrs2)
# Check residuals
resids2 <- residuals(model2)
plot(df$educ, resids2)
# Diagnostic plots
par(mfrow=c(2,2))
plot(model2)
# Check significance of predictor
anova(model2)## Analysis of Variance Table
##
## Response: df$educ
## Df Sum Sq Mean Sq F value Pr(>F)
## df$hrs2 1 2.05 2.0456 0.1984 0.6585
## Residuals 38 391.73 10.3087Multivariate analysis
Try to examine the assumptions of the multivariate analysis and model below.
## conducting a multivariate linear analysis
mlm1 <- lm(cbind(hrs2, educ) ~ race + sex, data = df, weight = wtssall)
summary(mlm1)## Response hrs2 :
##
## Call:
## lm(formula = hrs2 ~ race + sex, data = df, weights = wtssall)
##
## Weighted Residuals:
## Min 1Q Median 3Q Max
## -25.903 -5.609 2.901 7.820 26.521
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 48.37616 7.68946 6.291 2.54e-07 ***
## race -0.01122 3.47654 -0.003 0.997
## sex -6.38269 4.44618 -1.436 0.160
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 12.62 on 37 degrees of freedom
## Multiple R-squared: 0.05918, Adjusted R-squared: 0.008322
## F-statistic: 1.164 on 2 and 37 DF, p-value: 0.3235
##
##
## Response educ :
##
## Call:
## lm(formula = educ ~ race + sex, data = df, weights = wtssall)
##
## Weighted Residuals:
## Min 1Q Median 3Q Max
## -5.4129 -2.5731 0.0395 1.4778 5.1925
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 15.6865 1.8605 8.431 3.87e-10 ***
## race -2.2122 0.8412 -2.630 0.0124 *
## sex 1.2441 1.0758 1.156 0.2549
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 3.054 on 37 degrees of freedom
## Multiple R-squared: 0.1591, Adjusted R-squared: 0.1136
## F-statistic: 3.5 on 2 and 37 DF, p-value: 0.04054