Youth Risk Behavior Surveillance System (YRBSS)

The high school edition of the Behavioral Risk Factor Surveillance System (BRFSS).

• One table with one row per sampled youth respondent.

• A complex sample designed to generalize to US public and private school students in grades 9-12.

• Released biennially since 1993.

• Administered by the Centers for Disease Control and Prevention.

---

Recommended Reading

Four Example Strengths & Limitations:

✔️ 2023 included 68 site-level surveys in 39 states, 3 tribal governments, 5 territories, and 21 districts

✔️ Classrooms sampled in each grade from required subject (e.g., math) or period (e.g., homeroom)

❌ BMI calculated from self-reported height and weight underestimates the prevalence of obesity

❌ Arizona, Georgia, and Kansas did not obtain representative data from their 2023 high school surveys

Three Example Findings:

1. After Needham, Massachusetts outlawed tobacco sales to anyone under 21, the town’s youth smoking rate decline nearly tripled that of its neighbors between implementation of the law and 2010.

2. In 2021, 13% of Mississippi high school students used cannabis, lower than the US average of 28%.

3. 77% of U.S. high school students in 2023 reported using social media at least several times a day.

Two Methodology Documents:

Methodology of the Youth Risk Behavior Surveillance System

Wikipedia Entry

One Haiku:

# maladolescence
# epidemiology
# sex, drugs, rock and roll

---

Download, Import, Preparation

Load the SAScii library to interpret a SAS input program, and also re-arrange the SAS input program:

library(SAScii)

sas_url <-
    "https://www.cdc.gov/yrbs/files/2023/2023XXH_SAS_Input_Program.sas"

sas_text <- tolower( readLines( sas_url ) )

# find the (out of numerical order)
# `site` location variable's position
# within the SAS input program
site_location <- which( sas_text == '@1 site $3.' )

# find the start field's position
# within the SAS input program
input_location <- which( sas_text == "input" )

# create a vector from 1 to the length of the text file
sas_length <- seq( length( sas_text ) )

# remove the site_location
sas_length <- sas_length[ -site_location ]

# re-insert the site variable's location
# immediately after the starting position
sas_reorder <- 
    c( 
        sas_length[ seq( input_location ) ] , 
        site_location , 
        sas_length[ seq( input_location + 1 , length( sas_length ) ) ] 
    )

# re-order the sas text file
sas_text <- sas_text[ sas_reorder ]

sas_tf <- tempfile()

writeLines( sas_text , sas_tf )

Download and import the national file:

dat_tf <- tempfile()

dat_url <-
    "https://www.cdc.gov/yrbs/files/2023/XXH2023_YRBS_Data.dat"
    
download.file( dat_url , dat_tf , mode = 'wb' )

yrbss_df <- read.SAScii( dat_tf , sas_tf )

names( yrbss_df ) <- tolower( names( yrbss_df ) )

yrbss_df[ , 'one' ] <- 1

Save Locally  

Save the object at any point:

# yrbss_fn <- file.path( path.expand( "~" ) , "YRBSS" , "this_file.rds" )
# saveRDS( yrbss_df , file = yrbss_fn , compress = FALSE )

Load the same object:

# yrbss_df <- readRDS( yrbss_fn )

Survey Design Definition

Construct a complex sample survey design:

library(survey)

yrbss_design <- 
    svydesign( 
        ~ psu , 
        strata = ~ stratum , 
        data = yrbss_df , 
        weights = ~ weight , 
        nest = TRUE 
    )

Variable Recoding

Add new columns to the data set:

yrbss_design <- 
    update( 
        yrbss_design , 
        did_not_always_wear_seat_belt = as.numeric( qn8 == 1 ) ,
        ever_used_marijuana = as.numeric( qn46 == 1 ) ,
        tried_to_quit_tobacco_past_year = as.numeric( qn40 == 1 ) ,
        used_tobacco_past_year = as.numeric( q40 > 1 )
    )

---

Analysis Examples with the survey library  

Unweighted Counts

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

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

svyby( ~ one , ~ ever_used_marijuana , yrbss_design , unwtd.count )

Weighted Counts

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

svytotal( ~ one , yrbss_design )

svyby( ~ one , ~ ever_used_marijuana , yrbss_design , svytotal )

Descriptive Statistics

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

svymean( ~ bmipct , yrbss_design , na.rm = TRUE )

svyby( ~ bmipct , ~ ever_used_marijuana , yrbss_design , svymean , na.rm = TRUE )

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

svymean( ~ q2 , yrbss_design , na.rm = TRUE )

svyby( ~ q2 , ~ ever_used_marijuana , yrbss_design , svymean , na.rm = TRUE )

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

svytotal( ~ bmipct , yrbss_design , na.rm = TRUE )

svyby( ~ bmipct , ~ ever_used_marijuana , yrbss_design , svytotal , na.rm = TRUE )

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

svytotal( ~ q2 , yrbss_design , na.rm = TRUE )

svyby( ~ q2 , ~ ever_used_marijuana , yrbss_design , svytotal , na.rm = TRUE )

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

svyquantile( ~ bmipct , yrbss_design , 0.5 , na.rm = TRUE )

svyby( 
    ~ bmipct , 
    ~ ever_used_marijuana , 
    yrbss_design , 
    svyquantile , 
    0.5 ,
    ci = TRUE , na.rm = TRUE
)

Estimate a ratio:

svyratio( 
    numerator = ~ tried_to_quit_tobacco_past_year , 
    denominator = ~ used_tobacco_past_year , 
    yrbss_design ,
    na.rm = TRUE
)

Subsetting

Restrict the survey design to youths who ever drank alcohol:

sub_yrbss_design <- subset( yrbss_design , qn40 > 1 )

Calculate the mean (average) of this subset:

svymean( ~ bmipct , sub_yrbss_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( ~ bmipct , yrbss_design , na.rm = TRUE )

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

grouped_result <-
    svyby( 
        ~ bmipct , 
        ~ ever_used_marijuana , 
        yrbss_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( yrbss_design )

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

svyvar( ~ bmipct , yrbss_design , na.rm = TRUE )

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

# SRS without replacement
svymean( ~ bmipct , yrbss_design , na.rm = TRUE , deff = TRUE )

# SRS with replacement
svymean( ~ bmipct , yrbss_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( ~ did_not_always_wear_seat_belt , yrbss_design ,
    method = "likelihood" , na.rm = TRUE )

Regression Models and Tests of Association

Perform a design-based t-test:

svyttest( bmipct ~ did_not_always_wear_seat_belt , yrbss_design )

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

svychisq( 
    ~ did_not_always_wear_seat_belt + q2 , 
    yrbss_design 
)

Perform a survey-weighted generalized linear model:

glm_result <- 
    svyglm( 
        bmipct ~ did_not_always_wear_seat_belt + q2 , 
        yrbss_design 
    )

summary( glm_result )

---

Replication Example

This example matches statistics, standard errors, and confidence intervals from the “did not always wear a seat belt” row of PDF page 29 of this CDC analysis software document:

unwtd_count_result <-
    unwtd.count( ~ did_not_always_wear_seat_belt , yrbss_design )

stopifnot( coef( unwtd_count_result ) == 15071 )

wtd_n_result <-
    svytotal( 
        ~ one , 
        subset(
            yrbss_design , 
            !is.na( did_not_always_wear_seat_belt ) 
        )
    )

stopifnot( round( coef( wtd_n_result ) , 0 ) == 16917 )

share_result <-
    svymean(
        ~ did_not_always_wear_seat_belt ,
        yrbss_design ,
        na.rm = TRUE 
    )

stopifnot( round( coef( share_result ) , 4 ) == .3958 )

stopifnot( round( SE( share_result ) , 4 ) == .0172 )

ci_result <-
    svyciprop(
        ~ did_not_always_wear_seat_belt ,
        yrbss_design , 
        na.rm = TRUE
    )

stopifnot( round( confint( ci_result )[1] , 4 ) == 0.3621 )

stopifnot( round( confint( ci_result )[2] , 4 ) == 0.4304 )

---

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 YRBSS users, this code replicates previously-presented examples:

library(srvyr)
yrbss_srvyr_design <- as_survey( yrbss_design )

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

yrbss_srvyr_design %>%
    summarize( mean = survey_mean( bmipct , na.rm = TRUE ) )

yrbss_srvyr_design %>%
    group_by( ever_used_marijuana ) %>%
    summarize( mean = survey_mean( bmipct , na.rm = TRUE ) )