I have a dataset with 137 observations and 4 binary variables (0, 1). I want to check their factor structure.

I have tried out two possibilities:

Option 1:

fit_efa <- efa(data = df_children_td_first, 
               nfactors = 1,
               rotation = "promax", 
               estimator="WLSMV", 
               ordered = TRUE, 
               output = "efa") 

In this case, however, I get the following warning: “Warning: lavaan WARNING: some estimated ov variances are negative”.
These are the Standardized loadings:

                       f1      unique.var   communalities

tom_first_test1_binary 1.007 -0.015 1.015
tom_first_test2_binary 0.789 0.378 0.622
tom_first_test3_binary 0.873 0.238 0.762
tom_first_test4_binary 0.862 0.256 0.744

This may not work.

So I tried:
Option 2:

efa.model <- '
efa("efa")*f1 =~ tom_first_test1_binary + tom_first_test2_binary + tom_first_test3_binary + tom_first_test4_binary
'

fit <- cfa(efa.model, 
           data = df_children_td_first,
               rotation = "promax", 
               estimator = "WLSMV", 
               ordered = TRUE)

summary(fit, standardized = TRUE)

I don’t get an error but this output:
Latent Variables:
Estimate Std.lv Std.all
f1 =~ efa
tm_frst_tst1_b 1.007 1.007 1.007
tm_frst_tst2_b 0.789 0.789 0.789
tm_frst_tst3_b 0.873 0.873 0.873
tm_frst_tst4_b 0.862 0.862 0.862

How do I need to proceed in this case? I have never seen estimates above 1.

See description above.