Longitudinal person-period data from a study of adolescent alcohol use, tracking 82 adolescents from age 14 to age 16. Each row represents one person at one time point, resulting in 246 observations (82 persons x 3 occasions). The data include measures of alcohol consumption, peer alcohol use, and whether the adolescent is a child of an alcoholic (COA). This dataset is commonly used in multilevel/growth curve modeling textbooks to illustrate individual growth trajectories and the effects of time-varying and time-invariant predictors.
Format
A tibble with 246 rows and 9 columns:
- id
Person identifier. Type: numeric. Range: (1, 82). Uniquely identifies each of the 82 adolescents.
- age
Age in years at the time of measurement. Type: numeric. Values: 14, 15, 16. Three measurement occasions.
- coa
Child of an alcoholic. Type: numeric. Binary indicator (0/1) where 1 = child of an alcoholic parent, 0 = not a child of an alcoholic. Time-invariant predictor.
- male
Sex of adolescent. Type: numeric. Binary indicator (0/1) where 1 = male, 0 = female. Time-invariant predictor.
- age_14
Age centered at 14 (i.e., age minus 14). Type: numeric. Values: 0, 1, 2. Used as the time variable in growth models so that the intercept represents alcohol use at age 14.
- alcuse
Alcohol use composite score. Type: numeric. Range: (0.00, 3.61). Higher values indicate greater alcohol consumption. This is the primary outcome variable.
- peer
Peer alcohol use score. Type: numeric. Range: (0.00, 2.53). Mean = 1.02. Higher values indicate greater peer alcohol use. Time-varying covariate.
- cpeer
Peer alcohol use, grand-mean centered. Type: numeric. Range: (-1.02, 1.51). Mean approximately 0. This is the peer variable centered by subtracting the grand mean (1.02) to aid interpretation of the intercept in regression models.
- ccoa
Child of alcoholic, grand-mean centered. Type: numeric. Values: -0.45, 0.55. Mean approximately 0. This is the COA variable centered by subtracting the grand mean (0.45) so that the intercept represents the expected outcome for an average person on this predictor rather than for a non-COA.
Source
Singer, J. D., & Willett, J. B. (2003). Applied Longitudinal
Data Analysis: Modeling Change and Event Occurrence. Oxford University
Press. Original data file: alcohol1_pp.dta
Details
This dataset is used in Chapters on interactions and multilevel modeling to illustrate growth curve models and person-period data structures. Key analyses include: fitting individual growth curves of alcohol use over time, examining the effects of COA status and peer influence on alcohol use trajectories, and demonstrating grand-mean centering of predictors.
The centered variables (cpeer, ccoa) illustrate the common practice of centering predictors to improve interpretability of the intercept in multilevel and longitudinal models. Grand-mean centering shifts the meaning of the intercept from the value when all predictors are zero to the value at the sample average of the predictor.
Examples
data(alcohol1_pp)
head(alcohol1_pp)
#> # A tibble: 6 × 9
#> id age coa male age_14 alcuse peer cpeer ccoa
#> <int> <int> <int> <int> <int> <dbl> <dbl> <dbl> <dbl>
#> 1 1 14 1 0 0 1.73 1.26 0.247 0.549
#> 2 1 15 1 0 1 2 1.26 0.247 0.549
#> 3 1 16 1 0 2 2 1.26 0.247 0.549
#> 4 2 14 1 1 0 0 0.894 -0.124 0.549
#> 5 2 15 1 1 1 0 0.894 -0.124 0.549
#> 6 2 16 1 1 2 1 0.894 -0.124 0.549
# Growth model: alcohol use as a function of time and COA status
lm(alcuse ~ age_14 + ccoa, data = alcohol1_pp)
#>
#> Call:
#> lm(formula = alcuse ~ age_14 + ccoa, data = alcohol1_pp)
#>
#> Coefficients:
#> (Intercept) age_14 ccoa
#> 0.6512 0.2707 0.6938
#>
# Interaction between time and centered peer alcohol use
lm(alcuse ~ age_14 * cpeer, data = alcohol1_pp)
#>
#> Call:
#> lm(formula = alcuse ~ age_14 * cpeer, data = alcohol1_pp)
#>
#> Coefficients:
#> (Intercept) age_14 cpeer age_14:cpeer
#> 0.6516 0.2706 0.7588 -0.1514
#>