heplots (version 1.8.5)

LearnDis: Reading and Arithmetic Achievement in Children with Learning Disabilities

Description

A small factorial dataset from Tabachnick & Fidell (2013) relating a treatment/control intervention and degree of disability to reading and arithmetic achievement test scores, with IQ also recorded. It is not clear whether this is fictitious data or derived from a real study.

It provides for simple examples of MANOVA, MANCOVA and stepdown analysis

Usage

data("LearnDis")

Format

A data frame with 18 observations on the following 5 variables, a 3 (Disability) x 2 (Treatment) between-subjects factorial with n = 3 per cell.

Disability

Degree of disability, an ordered factor with levels Mild < Moderate < Severe

Treatment

a factor with levels Treatment Control

WRAT_R

Wide Range Achievement Test, Reading subtest score, a numeric vector

WRAT_A

Wide Range Achievement Test, Arithmetic subtest score, a numeric vector

IQ

IQ score, a numeric vector -- used in the source as a MANCOVA covariate, not part of the stepdown analysis itself

References

The Roy-Bargmann stepdown analysis of this data (WRAT-R prioritized over WRAT-A) appears in the same source, §7.5.3.2, Tables 7.7-7.9, pp.273-274.

Roy, S. N. (1958). Step-Down Procedure in Multivariate Analysis. The Annals of Mathematical Statistics, 29(4), 1177-1187. tools:::Rd_expr_doi("10.1214/aoms/1177706449").

Examples

data(LearnDis)
str(LearnDis)

ld.mod <- lm(cbind(WRAT_R, WRAT_A) ~ Disability * Treatment, data = LearnDis)
car::Anova(ld.mod)

heplot(ld.mod, fill = TRUE, fill.alpha = 0.1)

# Roy-Bargmann stepdown: does WRAT-A add anything to WRAT-R for the
# Treatment effect? 
# Needs Type III SS, since the WRAT_R covariate breaks the balanced
# factorial's orthogonality:
options(contrasts = c("contr.sum", "contr.poly"))
step2.mod <- lm(WRAT_A ~ WRAT_R + Disability * Treatment, data = LearnDis)
car::Anova(step2.mod, type = "III")