NlsyLinks (version 2.2.2)

Survey79: Dataset containing survey details for each subject, for each year

Description

Each row represents a survey that a subject completed (or didn't complete). It can be very helpful when restructuring the NLS investigator extracts into a longitudinal dataset that's aligned by age (instead of by survey wave). The Age variables can help to align other response variables across subjects. While the 'SurveySource' indicates where to look for their responses.

These variables are useful to many types of analyses (not just behavior genetics), and are provided to save users time.

Format

A data frame with 580,752 observations on the following 6 variables.

• SubjectTag see the variable of the same name in Links79Pair

• SurveySource The location of that subject's survey responses that year. Values are NoInterview, Gen1, Gen2C or Gen2YA.

• SurveyYear The year/wave of the survey.

• Survey79 The exact date of the administered survey.

• Age The subject's age at the time of the survey, rounded to 1 decimal. See Details below.

Author

Will Beasley

Download CSV If you're using the NlsyLinks package in R, the dataset is automatically available. To use it in a different environment, download the csv, which is readable by all statistical software. links-metadata-2017-79.yml documents the dataset version information.

Details

Age uses a variable called AgeCalculateYears (according to a their own response, or their mother's response) if it's available, and uses AgeSelfReportYears (according to a their own response, or their mother's response) if not. The values usually agree, but not always. These two separated variables are available for download, but withheld from the R package (otherwise we would exceed CRAN's maximum size).

The exact date of birth isn't public (only the subject's month of birth). To balance the downward bias of two weeks, their birthday is set to the 15th day of the month to produce AgeCalculateYears.

In the Gen2 Child dataset, self-reported age is stated by month (eg, the child is 38 months old); a constant of 0.5 months has been added to balance the downward bias. In the Gen2 YA and Gen1 datasets, self-reported age is stated by year (eg, the subject is 52 years old); a constant of 0.5 years has been added.

Examples

library(NlsyLinks) # Load the package into the current R session.

summary(Survey79)
table(Survey79$SurveyYear, Survey79$SurveySource)
table(is.na(Survey79$AgeSelfReportYears), is.na(Survey79$AgeCalculateYears))

if (require(ggplot2) & require(dplyr)) {
  dsSourceYear <- Survey79 %>%
    dplyr::count(SurveyYear, SurveySource) %>%
    dplyr::filter(SurveySource != "NoInterview")

  Survey79 %>%
    dplyr::filter(SurveySource != "NoInterview") %>%
    dplyr::group_by(SurveySource, SurveyYear) %>%
    dplyr::summarize(
      age_min = min(Age, na.rm = TRUE),
      age_max = max(Age, na.rm = TRUE)
    ) %>%
    dplyr::ungroup() %>%
    ggplot(aes(x = SurveyYear, ymin = age_min, ymax = age_max, color = SurveySource)) +
    geom_errorbar() +
    scale_color_brewer(palette = "Dark2") +
    theme_minimal() +
    theme(legend.position = c(0, 1), legend.justification = c(0, 1))
}