R/DIMTESTS.R

In EFA.dimensions: Exploratory Factor Analysis Functions for Assessing Dimensionality

DIMTESTS <- function(data, tests=c('EMPKC', 'HULL', 'RAWPAR'), corkind='pearson', Ncases=NULL, 
                     HULL_method='PAF', HULL_gof='CAF', HULL_cor_method='pearson',
                     CD_cor_method='pearson', display = NULL) {
       
data <- MISSING_DROP(data)

# set up cormat
cordat <- setupcormat(data, corkind=corkind, Ncases=Ncases)
cormat <- cordat$cormat
ctype  <- cordat$ctype
Ncases <- cordat$Ncases


if (is.null(display)) display = 2

verbose <- ifelse(display == 2, TRUE, FALSE) 

#if (is.null(tests)) tests = 'EMPKC'
		
dimtests <- matrix(-9999,length(tests),1)

for (lupe in 1:length(tests)) {
	
	if (tests[lupe] == 'CD') {		
		 dimtests[lupe,1] <- EFAtools::CD(x=data, n_factors_max = NA, N_pop = 10000, N_samples = 500, alpha = 0.3,
                                          cor_method = c("pearson", "spearman", "kendall"), max_iter = 50)$n_factors
		if (verbose == TRUE)	 {
			message('\n\nCOMPARISON DATA')
	        message('\nThe number of factors according to the Comparison Data test (from EFAtools) = ', dimtests[lupe,1], '\n')
	    }
	}

	if (tests[lupe] == 'EMPKC') dimtests[lupe,1] <- EMPKC(cormat, Ncases=Ncases, verbose=verbose)$NfactorsEMPKC
	
	if (tests[lupe] == 'HULL') {
		 hullOutput <- EFAtools::HULL(x=cormat, N = Ncases, method = HULL_method, gof = HULL_gof, 
		                              eigen_type = c("PCA"), cor_method = HULL_cor_method, n_datasets = 500, 
                                      percent = 50, decision_rule = c("means") )
                                               
		if (HULL_method == 'PAF')  dimtests[lupe,1] <- hullOutput$n_fac_CAF
		if ( (HULL_method == 'ML' | HULL_method == 'ULS') & HULL_gof=='CFI')    dimtests[lupe,1] <- hullOutput$n_fac_CFI
		if ( (HULL_method == 'ML' | HULL_method == 'ULS') & HULL_gof=='CAF')    dimtests[lupe,1] <- hullOutput$n_fac_CAF
		if ( (HULL_method == 'ML' | HULL_method == 'ULS') & HULL_gof=='RMSEA')  dimtests[lupe,1] <- hullOutput$n_fac_RMSEA	
		if (verbose == TRUE)	 {
			message('\n\nHULL METHOD')
	        message('\nThe number of factors according to the Hull method test (from EFAtools) = ', dimtests[lupe,1], '\n')
	    }
	}

	if (tests[lupe] == 'MAP') dimtests[lupe,1] <- MAP(cormat, Ncases=Ncases, verbose=verbose)$NfactorsMAP
	
	if (tests[lupe] == 'NEVALSGT1') dimtests[lupe,1] <- NEVALSGT1(cormat, Ncases=Ncases, verbose=verbose)$NfactorsNEVALSGT1
	
	if (tests[lupe] == 'RAWPAR') dimtests[lupe,1] <- RAWPAR(cormat, Ncases=Ncases, verbose=verbose)$NfactorsPA
	
	if (tests[lupe] == 'SALIENT') dimtests[lupe,1] <- SALIENT(cormat, Ncases=Ncases, verbose=verbose)$NfactorsSALIENT
	
	if (tests[lupe] == 'SESCREE') dimtests[lupe,1] <- SESCREE(cormat, Ncases=Ncases, verbose=verbose)$NfactorsSESCREE
	
	if (tests[lupe] == 'SMT') dimtests[lupe,1] <- SMT(cormat, Ncases=Ncases, verbose=verbose)$NfactorsSMT
}
rownames(dimtests) <- tests
colnames(dimtests) <- '# of Factors:'

if (display > 0) {
	message('\n\nDIMTESTS results:\n')
	print(dimtests)
}	

dimtestOutput <- list(dimtests=dimtests, NfactorsDIMTESTS = dimtests[1,1])

}



# # DIMTESTS(data=data_Harman, tests='CD', corkind='pearson', Ncases=305, display = NULL)

# DIMTESTS(data=data_RSE, tests='CD', corkind='pearson', display = NULL)

# DIMTESTS(data=data_NEOPIR, tests='CD', corkind='pearson', display = NULL)



# DIMTESTS(data=data_Harman, tests='HULL', corkind='pearson', Ncases=305, 
                     # HULL_method='ML', HULL_gof='CFI', HULL_cor_method='pearson',
                     # display = NULL)

# DIMTESTS(data=data_Harman, tests='HULL', corkind='pearson', Ncases=305, 
                     # HULL_method='PAF', HULL_gof='CFI', HULL_cor_method='pearson',
                     # display = NULL)

# DIMTESTS(data=data_Harman, tests='HULL', corkind='pearson', Ncases=305, 
                     # HULL_method='ML', HULL_gof='RMSEA', HULL_cor_method='pearson',
                     # display = NULL)




# DIMTESTS(data = data_RSE, tests = c('EMPKC','MAP','RAWPAR','NEVALSGT1','SALIENT','SESCREE','SMT'), display=1)


# DIMTESTS(data = data_RSE, tests = 'EMPKC')

# DIMTESTS(data = data_RSE, tests = 'EMPKC', display=1)
# DIMTESTS(data = data_RSE, tests = 'MAP', display=1)
# DIMTESTS(data = data_RSE, tests = 'RAWPAR', display=1)
# DIMTESTS(data = data_RSE, tests = 'NEVALSGT1', display=1)
# DIMTESTS(data = data_RSE, tests = 'SALIENT', display=1)
# DIMTESTS(data = data_RSE, tests = 'SESCREE', display=1)
# DIMTESTS(data = data_RSE, tests = 'SMT', display=1)



# if (is.element('EMPKC', tests)) 
	# dimtestOutput <- rbind(dimtestOutput, EMPKC(data, verbose=verbose)$NfactorsEMPKC)

# if (is.element('MAP', tests))  
	# dimtestOutput <- rbind(dimtestOutput, MAP(data, verbose=verbose)$NfactorsMAP)

# if (is.element('RAWPAR', tests))  
	# dimtestOutput <- rbind(dimtestOutput, RAWPAR(data, verbose=verbose)$NfactorsPA)

# if (is.element('NEVALSGT1', tests))  
	# dimtestOutput <- rbind(dimtestOutput, NEVALSGT1(data, verbose=verbose)$NfactorsNEVALSGT1)

# if (is.element('SALIENT', tests))  
	# dimtestOutput <- rbind(dimtestOutput, SALIENT(data, verbose=verbose)$NfactorsSALIENT)

# if (is.element('SESCREE', tests))  
	# dimtestOutput <- rbind(dimtestOutput, SESCREE(data, verbose=verbose)$NfactorsSESCREE)

# if (is.element('SMT', tests))  
	# dimtestOutput <- rbind(dimtestOutput, SMT(data, verbose=verbose)$NfactorsSMT)