Is the value of AVE less than but close to 0.5 acceptable?

In one of my measurement CFA models (using AMOS) the factor loading of two items are smaller than 0.3. I found some scholars that mentioned only the ones which are smaller than 0.2 should be considered for deletion. However, there are various ideas in this regard. Some said that the items which their factor loading are below 0.3 or even below 0.4 are not valuable and should be deleted.

The measurement I used is a standard one and I do not want to remove any item. What should I do? 

Thank you for your valuable time.

Regards,

Hi All,

I am conducting CFA using AMOS with the data set of 593 and five latent variables. The CFA model fit is good (Cmin/DF= 2.28; p=0.00; CFI=0.935; GFI=0.968; AGFI=0.947; RMSEA=0.047; PCLOSE=0.69) but when checking for the Composite Reliability (CR) and Average Variance Extracted (AVE), I am getting the values less than 0.7 and 0.5 respectively (see attached summary of result). In the other hand, Discriminant Validity is looking good. Would you please suggest whether I can go ahead with these results (any reference)? If not, is there any way to improve this?

I have already performed the EFA and removed some indicators to make the pattern matrix better. I tried hard but unable to make total cumulative variance explained more than 45% and KMO value 0.636.

In addition to this, I have a dependent variable which is in the form of bivariate nominal variable i.e. not on a Linkert scale. Can I add this variable while solving structural model after completing CFA in AMOS?

Thank you in advance.

Hello,

Is it possible to continue with a low average variance extracted (AVE) value if the values of composite reliability (CR) and Cronbach's alpha fall within the excellent range? And if not, then what is the best way forward?

Cronbach alpha acceptability values

It is desirable that for the normal distribution of data the values of skewness should be near to 0. What if the values are +/- 3 or above?

his very interesting professional discussion. i have tried to construct SEM for my study. i have 5 latent variables in my model, depression (9 questions,), General anxiety (7 question), social anxiety (10 question) and PTSD (17 questions) and also somatic symptom (15 questions). depression and anxiety are my dependent variable and used second order SEM because anxiety measured using general anxiety, social anxiety and PTSD). The sample size of this study is 217. i had conduct data cleaning activity like missing record, outlier, unengaded response and common bias and other also check sample size adequate using KMO (Kmo=0.89). even tried to determain the SEM but the model not fit the required mode fit criteria, could you please help me with any think, What is the minimum acceptable range for factor loading in SEM? Thank you.

I have recently received the following comments on my manuscript by a reviewer but could not comprehend it properly. I am alien to the concept of Common Method Bias. 

"Recent editorial work has stressed the potential problem of common method bias, which describes the measurement error that is compounded by the sociability of respondents who want to provide positive answers (Chang, v. Witteloostuijn and Eden, 2010). The paper study collected data on both the independent and dependent variables from the same respondents at one point in time, thus raising potential common method variance as false internal consistency might be present in the data. The authors however, failed to tell the reader how they countered common method bias."

Does anyone have a better ideas how to explain the used of Heterotrait-Monotrait Ratio of Correlations (HTMT) in assessing the discriminant validity in PLS-SEM model?

While estimating convergent validity, one latent factor gets AVE value less than 0.5. however, factor loadings gets values greater than 0.6 and model fitted well with fit indices. In this case, shall I accept the latent factor for model fitting or do I need to remove the latent factor from the model as it has AVE value less than 0.5?