Getting data frame from loadings of factor analysis (fa function in psych)

Question

I have a data frame such as follows:

x <- c(1, 2, 1, 2)
y <- c(1, 2, 3, 4)
z <- c(4, 3, 2, 1)
df <- data.frame(x, y, z)

I am running a factor analysis with the fa funciton from the psych package:

fit <- fa(df, nfactors = 2)
fit$loadings

This results in the following output:

Loadings:
  MR1    MR2   
x  0.448       
y  0.999       
z -0.999       

                 MR1   MR2
SS loadings    2.195 0.000
Proportion Var 0.732 0.000
Cumulative Var 0.732 0.732

I would like to save the table with MR1 and MR2 as a data frame. Does anyone know how could this be done? Thank you.

Manifestative

Answer 1

I believe the short answer should simply be:

as.data.frame(unclass(fit$loadings))

Giving:

         MR1         MR2
x  0.4502976  0.10314193
y  0.9984784 -0.02325767
z -0.9984784  0.02325767

Huskey

Answer 2

x <- c(1, 2, 1, 2)
y <- c(1, 2, 3, 4)
z <- c(4, 3, 2, 1)
df <- data.frame(x, y, z)

fit <- psych::fa(df, nfactors = 2)

x <- fit$loadings

via stats:::print.loadings:

Lambda <- unclass(x)
p <- nrow(Lambda)
factors <- ncol(Lambda)

vx <- colSums(x^2)
varex <- rbind(`SS loadings` = vx)

if (is.null(attr(x, "covariance"))) {
  varex <- rbind(varex, `Proportion Var` = vx/p)
  if (factors > 1) 
    varex <- rbind(varex, `Cumulative Var` = cumsum(vx/p))
}

tibble::rownames_to_column(as.data.frame(varex), "x")
##                x       MR1          MR2
## 1    SS loadings 2.1954555 3.000000e-30
## 2 Proportion Var 0.7318185 1.000000e-30
## 3 Cumulative Var 0.7318185 7.318185e-01

And, for the first table:

cutoff <- 0.1 # (the default for the `print.loadings()` function)
Lambda <- unclass(x)
p <- nrow(Lambda)
fx <- setNames(Lambda, NULL)
fx[abs(Lambda) < cutoff] <- NA_real_
fx <- as.data.frame(fx)
rownames(fx) <- NULL
fx
##          MR1 MR2
## 1  0.4476761  NA
## 2  0.9987596  NA
## 3 -0.9987596  NA

Sledge

This works well. I'm sorry I wasn't specific enough: I would like to turn the first table into a data frame (the one with MR1, MR2, x, y, z). Could this be done in a way that doesn't require to list out all the variables from the the original df data frame (x, y, z)? – Manifestative

updated to account for needing the first table and using NA vs the blanks for the other column. – Sledge