designEffect: Design effect and Intraclass Correlation Coeficient for...

In AnguloB/SubscaleExplorer: Psychometric scale explorer and management

Description

Calculate ICC1 and design effect for multilevel analysis

Usage

designEffect(df, cluster=cluster, data=df, round1=3)

Arguments

df	dataframe or matrix
cluster	vector same length as df (e.g Data$team)
round1	round, default to 3
data	dataframe or matrix, default same as df

Details

ICC is the intraclass correlation coefficient 1."ICC1 represents the amount of individual-level variance that can be explained by group membership" (see more details in (ICC1)). Design effect is computed according to Hox's (2010) formula.

If cluster size is not the same across groups, sample size mean is used instead.

Value

n	Cluster size or mean cluster size when groups are different size
table	three columns, first indicates variable name, second ICC1 and third indicates design effect

Author(s)

Ariadna Angulo-Brunet ([e-mail]), Carme Viladrich

References

Hox, J. (2010). Multilevel analysis: techniques and applications. (2nd ed.). New York: NY: Routledge.

See Also

ICC1, mult.icc, designEffectICC

Examples

#generate  some random data
x<-1:5
y<-1:10
v1<-sample(x,800,replace=TRUE)
v2<-sample(x,800,replace=TRUE)
v3<-sample(x,800,replace=TRUE)
v4<-sample(x,800,replace=TRUE)
team<- sample(y, 800, replace=TRUE)
df <- data.frame(v1, v2,v3,v4)

designEffect(df, team )

AnguloB/SubscaleExplorer documentation built on Jan. 3, 2021, 2:57 p.m.