htestClust (version 0.2.2)

screen8: Example data for informative cluster size

Description

Simulated hypothetical clustered data created for illustration of functions in the htestClust package.

Usage

data(screen8)

Format

A data frame with 2224 rows and 12 columns:

sch.id

identification variable for school (clusters).

stud.id

identification variable for students within schools (observations within clusters).

age

student age in years.

gender

binary student gender.

height

student height in inches.

weight

student weight in lbs.

math

score from standardized math test.

read

score from standardized reading test.

phq2

ordinal (0-6) score from a mental health screening; higher scores correspond to higher levels of depression.

qfit

age-adjusted fitness quartile from physical health assessment at end of school year.

qfit.s

age-adjusted fitness quartile from physical health assessment at beginning of school year.

activity

student's primary after-school activity.

Details

Hypothetical data simulated for the following scenario. An urban school district has collected demographic, biometric, and academic performance data from graduating 8th grade students. screen8 contains a sample of this data from 2224 students across 73 schools. Student-level observations are clustered within schools. The school district has implemented an incentive program in which schools with higher participation rates are prioritized for classroom and technology upgrades. Cluster size could be informative in this data, as resource-poor schools might have higher participation rates (larger cluster size), but also tend to have worse health metrics and lower standardized test scores.

Examples

# NOT RUN {
data(screen8)
head(screen8)

## plot average math scores by cluster size
cl.size <- as.numeric(table(screen8$sch.id))
ave.math <- tapply(screen8$math, list(screen8$sch.id), mean)
plot(cl.size, ave.math)

# }