best.rsa {RSAtools}

Compare a list of polynomial models against the data

Description

Compares any number of predefined or user-specific polynomial models and extracts their fit indices, thereby establishing best-fitting solutions.

Usage

best.rsa(RSA_object, order = c("wAIC", "R2adj"), robust = TRUE)

Arguments

RSA_object

x an object of class "RSA_object" generated by RSAmodel()

order

Single or vector of fit indices used to determine best-fitting polynomial families. The output matrix is ordered based on this fit index

robust

Should robust fit indices should be extracted? (default= TRUE)

Details

This function compares models based on information-theoretic criteria and statistical tests. The cubic saturated polynomial provides a benchmark reference for fit, against which predefined polynomial families (37 to date) or user-specific variants of these families are compared for absolute fit (likelihood ratio test), parsimony (wAIC), explained variance (adjusted R2), and ordinary SEM criteria (e.g., CFI, TLI, RMSEA, SRMR).

Value

A table containing fit indices for each model

Examples

#####ESTIMATE RSA OBJECT
RSA_step1 <-  RSAmodel(engagement ~ needs*supplies,
data= sim_NSfit, model= c("CUBIC","FM8_INCONG","FM9_INCONG","FM20_ASYMCONG",
"FM21_ASYMCONG","FM26_PARALLELASYMWEAK"))
##### COMPARE POLYNOMIAL FAMILIES FROM THE RSA OBJECT
RSA_step1_fit <- best.rsa(RSA_step1,order=c("wAIC"))
names(RSA_step1$models)
#Inspect best-fitting family model
summary(RSA_step1$models$FM26_PARALLELASYMWEAK)