13. SEM | SPSS AMOS | Assess Discriminant Validity | Heterotrait Monotrait Ratio

Using SEM based SPSS AMOS, this session discusses in detail the concept and methodology to assess discriminant validity using the Heterotrait Monotrait HTMT ratio.
Previous Lecture: • 12. SPSS AMOS - Assess...
Session Content
00:00 - Channel Introduction
00:15 - Concept of Discriminant Validity
01:09 - Concept of HTMT
02:29 - Calculating HTMT Ratio
08:02 - Formatting HTMT Matrix
#SPSSAMOS #SEM #HTMT #HeterotraitMonotraitRatio #StructuralEquationModelling #DataAnalysis #DiscriminantValidity #SEM #ConstructValidity #ResearchCoach

[@anitalow3628]

Never been known that HTMT can be calculated via Excel. Really open eyes! Thanks!

[@researchwithfawad]

You're welcome!

[@mariaritaafonso]

Thank you so much! I've been using your videos as a way of comprehending the analysis I'm doing for my master thesis. Thanks for the help, greetings from Portugal 😀

[@researchwithfawad]

You're very welcome!

[@farzaanakoodoruth3398]

Very comprehensive sir. thank you

[@researchwithfawad]

Thanks for watching.

[@omaranbari2484]

your explanation is simple yet very comprehensive. thank you

[@researchwithfawad]

Glad it was helpful!

[@FrancisNimoNunoo]

Thank you for your tutorials. You make me love statistics

[@researchwithfawad]

Happy to hear that! Thank you.

[@iskendervarol5672]

Thank you very much, great indeed.

[@researchwithfawad]

Pleasure. I am glad you liked it.

[@shayistamajeed9889]

Indeed very informative n helpful video,,,, May Allah bless u Sir

[@researchwithfawad]

Thanks. I am glad the series is helpful.

[@talhamansoor7108]

very informative and helpful as always

[@researchwithfawad]

Thanks. I am glad you liked it

[@dr.sandeepprabhu9014]

Wow, so well explained!, thanks

[@researchwithfawad]

You're very welcome!

[@yuzeng7158]

it helps me a lot, thks so much!!!

[@researchwithfawad]

Glad it helped!

[@ametorwobright1316]

Thank you for educating us

[@researchwithfawad]

Thanks. I am glad you liked it.

[@ametorwobright1316]

@@researchwithfawad I am a PhD student at that point of data collection and analysis. Your lectures have been beneficial. Thank you

[@sofianelaradi6176]

Thank you ❤ ❤ ❤

[@researchwithfawad]

Pleasure. I am glad you liked it.

[@richardchen8511]

Thank you so much!

[@researchwithfawad]

Glad it helped!

[@anasifti203]

Thank you, this is helpful. Is it possible to assess HTMT through the bootstrap method to see if it is significantly different from 1? Would appreciate it if you could suggest something on this.

[@researchwithfawad]

In my little knowledge, AMOS doesn't offer HTMT yet. It is manually calculated. Bootstrapping is not yet available

[@anasifti203]

@@researchwithfawad really appreciate your response. I am seeing quite a few good papers mentioning it but not sure how they are doing it. Ill just go with the HTMT values.

[@Alhamzah_F_Abbas]

Great explanation prof. so can we not consider AVE and report HTMT only?

[@researchwithfawad]

Thanks. Yes, you may.

[@simonchan2394]

Hi Fawad, can you tell me how I can change the order of the items in the implied correlation table? AMOS is putting Q14_1, then Q13_1 etc. but I need to switch the 2 items around so that I can create a block of only items related to Q14.

[@researchwithfawad]

Hi. You can try to put a number before Q so it can order it.

[@madhumitabanerjee4542]

Thank You so much for the video, Sir. I calculated and found two of my HTMT ratios to have negative values ( - 0.4467 and -0.5937). (Theoretically the constructs should be negatively correlated and that is what I have hypothesized). So are these negative values fine or are they a cause of concern?

[@researchwithfawad]

Pleasure. I am glad you liked it. Nothing to be concerned about.

[@madhumitabanerjee4542]

Thank You so very much, Sir for your wonderful wonderful student friendly videos and prompt replies which are additional blessings.

[@sunnydial1509]

Very helpful video! can you please help me as my doubt is that for having the correct implied correlation matrix do we need to eliminate the latent variable and errors and covarie all the items? because in my measurement model in the implied correlation matrix is showing the latent variable too...

[@researchwithfawad]

No, you do not need to eliminate the latent variable. But when calculating HTMT you will delete the latent variables from the matrix.

[@sunnydial1509]

​@@researchwithfawad Thank you very much i will do it that way then, however another strange thing is actually in AMOS i am getting values of correlations between the items which are different than those calculated directly using correl function in excel... for example if i calculate correlation between item EOU1 and EOU2 for my model in excel i get 0.38, while in the correlation matrix in AMOS it is 0.89... why are they so different and which should be used actually?

[@researchwithfawad]

They may be slightly different but not to that level. Check your formula again in Excel

[@sunnydial1509]

@@researchwithfawad I checked but they are different ... I tried to do it differently as well and I deleted the latent variables and covaried just all the items, and in this case the correlation values ​​I see under the estimates-> scalars-> correlations section are the same of the correlation that I get in excel doing the Pearson 's correlation, however when i go to the implied (for all variables) correlation matrix the values ​​are different again... which should i use?

[@zainabfirdousabbas5674]

Thank you so much sir. I am doing data analysis through your videos.helping alot. Sir, is it normal to have heterotait ratio equal to zero ? Can we report it in thesis? Or is there any issue. Please guide.

[@researchwithfawad]

Thanks for watching. 0 mean no correlation. Did you calculate it manually?

[@michellengo6978]

Hi, thanks for the explanation. May I know whether negative HTMT is acceptable?

[@researchwithfawad]

It is fine. prinicpally HTMT is normed between -1 and +1.

[@michellengo6978]

@@researchwithfawad Thank you Prof

[@mohmoh7078]

Thank you very much for the video; it is really helpful. My factor loadings are all above 0.7, model fitness is achieved, and convergent validity is achieved as well. Still, my HTMT values are very high (one above 0.9 and the others between 0.8 and 0.85). Any recommendation, Sir? * I followed a well-documented theory (TBL for sustainability).

[@researchwithfawad]

Glad it helped. Please watch this for HTMT issue kzbin.info/www/bejne/o6iyZ6aqaKyso6c

[@mohmoh7078]

@@researchwithfawad Thank you for your reply, I really appreciate it. My data is clean from any misconduct or missing values. The minimum SD is 0.43.

[@user-so8yx3ny6y]

All the values of the HTMt was less than 0.9 and 1 is alone 0.98 . Can i show this In my thesis. Wat justification reference can I give for this

[@researchwithfawad]

You may also check Fornell & Larcker Criteria. And if that is OK, you may mention that DV is attained.

[@user-so8yx3ny6y]

Even fornell and larcker is also not satisfying

[@syuhadaredzuan8187]

Hai, already follow your guideline. So, I got one result which is 0.906. Is it okay? but so far other results are below 0.90.

[@researchwithfawad]

This is fine. Just check the cross loadings as well. If cross loadings are good you can say discriminant validity is established.

[@syuhadaredzuan8187]

@@researchwithfawad thank you😊

[@syuhadaredzuan8187]

@@researchwithfawad how to check cross loadings using Amos?