fa.sort.RdAlthough the print.psych function will sort factor analysis loadings, sometimes it is useful to do this outside of the print function. fa.sort takes the output from the fa or principal functions and sorts the loadings for each factor. Items are located in terms of their greatest loading. The new order is returned as an element in the fa list. fa.organize allows for the columns or rows to be reorganized.
fa.sort(fa.results,polar=FALSE)
fa.organize(fa.results,o=NULL,i=NULL,cn=NULL,echelon=TRUE,flip=TRUE)The output from a factor analysis or principal components analysis using fa or principal. Can also just be a matrix of loadings. Can also organize the output of cluster loadings from iclust.
Sort by polar coordinates of first two factors (FALSE)
The order in which to order the factors
The order in which to order the items
new factor names
Organize the factors so that they are in echelon form (variable 1 .. n on factor 1, n+1 ...n=k on factor 2, etc.)
Flip factor loadings such that the colMean is positive.
The fa.results$loadings are replaced with sorted loadings.
fa.organize takes a factor analysis or components output and reorganizes the factors in the o order. Items are organized in the i order. This is useful when comparing alternative factor solutions.
The flip option works only for the case of matrix input, not for full fa objects. Use the reflect function.
A sorted factor analysis, principal components analysis, or omega loadings matrix.
These sorted values are used internally by the various diagram functions.
The values returned are the same as fa, except in sorted order. In addition, the order is returned as an additional element in the fa list.
See Also as fa, pca, iclust, print.psych, fa.diagram,
test.simple <- fa(sim.item(16),2)
fa.sort(test.simple)
#> Factor Analysis using method = minres
#> Call: fa(r = sim.item(16), nfactors = 2)
#> Standardized loadings (pattern matrix) based upon correlation matrix
#> MR1 MR2 h2 u2 com
#> V12 0.68 0.03 0.46 0.54 1
#> V11 0.65 0.00 0.42 0.58 1
#> V4 -0.64 -0.01 0.41 0.59 1
#> V2 -0.60 -0.04 0.37 0.63 1
#> V1 -0.60 0.03 0.36 0.64 1
#> V10 0.57 -0.04 0.33 0.67 1
#> V3 -0.56 0.04 0.32 0.68 1
#> V9 0.55 0.00 0.30 0.70 1
#> V16 0.01 0.64 0.41 0.59 1
#> V8 -0.05 -0.63 0.41 0.59 1
#> V7 0.00 -0.62 0.39 0.61 1
#> V14 0.01 0.60 0.36 0.64 1
#> V15 0.02 0.59 0.35 0.65 1
#> V13 -0.06 0.59 0.34 0.66 1
#> V6 0.04 -0.56 0.31 0.69 1
#> V5 0.00 -0.51 0.26 0.74 1
#>
#> MR1 MR2
#> SS loadings 2.96 2.82
#> Proportion Var 0.19 0.18
#> Cumulative Var 0.19 0.36
#> Proportion Explained 0.51 0.49
#> Cumulative Proportion 0.51 1.00
#>
#> With factor correlations of
#> MR1 MR2
#> MR1 1.00 0.07
#> MR2 0.07 1.00
#>
#> Mean item complexity = 1
#> Test of the hypothesis that 2 factors are sufficient.
#>
#> df null model = 120 with the objective function = 3.11 with Chi Square = 1531.41
#> df of the model are 89 and the objective function was 0.15
#>
#> The root mean square of the residuals (RMSR) is 0.02
#> The df corrected root mean square of the residuals is 0.03
#>
#> The harmonic n.obs is 500 with the empirical chi square 70.04 with prob < 0.93
#> The total n.obs was 500 with Likelihood Chi Square = 74.82 with prob < 0.86
#>
#> Tucker Lewis Index of factoring reliability = 1.014
#> RMSEA index = 0 and the 90 % confidence intervals are 0 0.014
#> BIC = -478.28
#> Fit based upon off diagonal values = 0.99
#> Measures of factor score adequacy
#> MR1 MR2
#> Correlation of (regression) scores with factors 0.91 0.90
#> Multiple R square of scores with factors 0.83 0.82
#> Minimum correlation of possible factor scores 0.66 0.63
fa.organize(test.simple,c(2,1)) #the factors but not the items have been rearranged
#> Factor Analysis using method = minres
#> Call: fa(r = sim.item(16), nfactors = 2)
#> Standardized loadings (pattern matrix) based upon correlation matrix
#> MR2 MR1 h2 u2 com
#> V1 0.03 -0.60 0.36 0.64 1
#> V2 -0.04 -0.60 0.37 0.63 1
#> V3 0.04 -0.56 0.32 0.68 1
#> V4 -0.01 -0.64 0.41 0.59 1
#> V5 -0.51 0.00 0.26 0.74 1
#> V6 -0.56 0.04 0.31 0.69 1
#> V7 -0.62 0.00 0.39 0.61 1
#> V8 -0.63 -0.05 0.41 0.59 1
#> V9 0.00 0.55 0.30 0.70 1
#> V10 -0.04 0.57 0.33 0.67 1
#> V11 0.00 0.65 0.42 0.58 1
#> V12 0.03 0.68 0.46 0.54 1
#> V13 0.59 -0.06 0.34 0.66 1
#> V14 0.60 0.01 0.36 0.64 1
#> V15 0.59 0.02 0.35 0.65 1
#> V16 0.64 0.01 0.41 0.59 1
#>
#> MR2 MR1
#> SS loadings 2.82 2.96
#> Proportion Var 0.18 0.19
#> Cumulative Var 0.18 0.36
#> Proportion Explained 0.49 0.51
#> Cumulative Proportion 0.49 1.00
#>
#> With factor correlations of
#> MR2 MR1
#> MR2 1.00 0.07
#> MR1 0.07 1.00
#>
#> Mean item complexity = 1
#> Test of the hypothesis that 2 factors are sufficient.
#>
#> df null model = 120 with the objective function = 3.11 with Chi Square = 1531.41
#> df of the model are 89 and the objective function was 0.15
#>
#> The root mean square of the residuals (RMSR) is 0.02
#> The df corrected root mean square of the residuals is 0.03
#>
#> The harmonic n.obs is 500 with the empirical chi square 70.04 with prob < 0.93
#> The total n.obs was 500 with Likelihood Chi Square = 74.82 with prob < 0.86
#>
#> Tucker Lewis Index of factoring reliability = 1.014
#> RMSEA index = 0 and the 90 % confidence intervals are 0 0.014
#> BIC = -478.28
#> Fit based upon off diagonal values = 0.99
#> Measures of factor score adequacy
#> MR2 MR1
#> Correlation of (regression) scores with factors 0.90 0.91
#> Multiple R square of scores with factors 0.82 0.83
#> Minimum correlation of possible factor scores 0.63 0.66