Dear all,

I am new to STATA and try to test the validity of my measures. I run Stata 16 for Microsoft. My data contains 425 observations of 16 variables, each measured with 2-5 items on a 6-point Likert scale. I tested for reliability with alpha and have values above 0.75. I have tried to conduct an Exploratory Factor Analysis using the factor commandand even specified the number of variables to be created from the items by adding factor(x) to the command, however I did not get the constructs I expected. Almost each item lead to other factors than the ones expected. As I used validated items from the literature, I decided to conduct a Confirmatory Factor Analysis. I used the following command from the Stata conference in 2017 (https://www.stata.com/meeting/france...7_Perrot.pdf):

validscale PE1 PE2 PE3 EE1 EE2 EE3 EE4 SI1 SI2 SI3 FC1 FC2 FC4 PFC1 PFC2 PFC3 IT1 IT2 IT3 DT1 DT2 DT3 DT4 DT5 PFR1 PFR2 PFR3 PVR1 PVR2 PVR3 TR1 TR2 T
> R3 OR1 OR2 OR3 OR4 BI1 BI2 BI3 MF1 MF2 MF3 MF4 IC1 IC2 IC3 IC4 PD1 PD2 PD3 PD4 UA1 UA2, partition(3 4 3 3 3 3 5 3 3 3 4 3 4 4 4 2) scorename(PE EE SI
> FC PFC IT DT PFR PVR TR OR BI MF IC PD UA) graphs compscore(stand) cfa cfamethod(ml) cfasb cfacov(PE1*EE1) alpha(0.7) delta(0.9) h(0.3) hjmin(0.3) tco
> nc(0.4)

And got the results visible in the picture attached.

The Goodness of fit results seem to be acceptable:
Goodness of fit (with Satorra-Bentler correction):

chi2 df chi2/df RMSEA [90% CI] SRMR NFI RNI CFI IFI MCI
2203.11 1256 1.8 0.042 [ ; ] 0.058 0.858 0.933 0.933 0.933 0.327
(p-value = 0.000)

My problem are the factor loadings, as they have values over 1. How should I decide whether they are good or bad? Usually they should be between 0.70 and 1. In the presentation of the conference, I did not find additional information.

The conference paper is available here: https://www.researchgate.net/publica...urement_scales.

I would appreciate any help!

Kind regards,
Ana