Input data matrix causes errors

[cwliu007]

Hi Pablo,

I found the following two situations causing different errors for estQ.

(1) first

library("randcorr")
r = randcorr(20)

library("cdmTools")
out1 = estQ(
r = r,
K = 3,
n.obs = NULL,
criterion = "loaddiff",
boot = TRUE,
efa.args = list(cor = "tet", rotation = "oblimin", fm = "uls"),
boot.args = list(N = 0.8, R = 100, verbose = FALSE, seed = 1)
)

Error in colnames<-(*tmp*, value = paste0("Cat", 1:maxK)) :
attempt to set 'colnames' on an object with less than two dimensions

(1) second

values <- c(
1.00000000, 0.2605664, 0.1858627, 0.1399712, 0.09735522,
0.26056643, 1.0000000, 0.1443643, 0.1962672, 0.20538981,
0.18586268, 0.1443643, 1.0000000, 0.1765424, 0.23795236,
0.13997122, 0.1962672, 0.1765424, 1.0000000, 0.17226901,
0.09735522, 0.2053898, 0.2379524, 0.1722690, 1.00000000
)
matrix_var <- matrix(values, nrow = 5, byrow = TRUE)

out1 = estQ(
r = matrix_var,
K = 3,
n.obs = NULL,
criterion = "loaddiff",
boot = TRUE,
efa.args = list(cor = "tet", rotation = "oblimin", fm = "uls"),
boot.args = list(N = 0.8, R = 100, verbose = FALSE, seed = 1)
)

Error in polychoric2(dat = dat, cor.smooth = cor.smooth, maxiter = maxit) :
Minimum value must always zero.

Thanks a lot for this useful package!
Chen-Wei

[Pablo-Najera]

Hi Chen-Wei,

Please note the documentation of the package, which states the following regarding the arguments of the estQ function:
n.obs: Number of individuals if r is a correlation matrix. If n.obs is provided, r will be treated as a correlation matrix. Use NULL if r is raw data. The default is NULL.
Since you are using a correlation matrix as input data for r in the function, you must specify the number of individuals. For instance, if you use n.obs = 100, it works.

In the second example, please note that matrix_var is not a correlation matrix, since it is not perfectly symmetrical. This is causing the code to think that it is not a correlation matrix, but a response matrix.

Kind regards,
Pablo

[cwliu007]

Thank you for pointing out this, Pablo. Now it works on my end.