SPSS 04: Convergent and Discriminant Validity of Scale

[@wong4728]

Jamaica here. This is a most excellent presentation.

[@kwamegyekye3528]

Best and simplified explanation l have watched so far. Thank you 😊

[@educationalinsightskarachi4969]

I am humbled by your kind word

[@drsureshraghavan9005]

Checking the Convergent and Discriminant validity through Correlation matrix is for the initial stage only right ? How do you differentiate this method through, Average Variants extracted through factor analysis and the Composite Reliability method ? Or through CFA method ?

[@educationalinsightskarachi4969]

Definitely this will be an initial analysis. We should calculate AVE or assess data for FL condition too. You may visit this video on link below. kzbin.info/www/bejne/Y6Sro5iYfryreLcsi=3GhzNUKh3wUzXu7g

[@saharmona785]

very informative👍

[@noonatatao687]

Thank you very much. Please tell me. Can i test these tow types of validity by testing the correlation between each item and its construct and between the constructs? Hope my question is clear.

[@educationalinsightskarachi4969]

@@noonatatao687 Thanks for your question. You will need to calculate correlation among the items for convergent and discriminant validity.

[@komalanand4325]

Your video is really good,,,,You explained that the values should be between 0.3 and 0.7,,,,for discriminate you said it should be below...can you pls tell the reference

[@educationalinsightskarachi4969]

Thanks, Komal. This explanation is based on two references. 1. John, O. P., & Benet-Martínez, V. (2000). Measurement: reliability, construct validation, and scale construction. In H. T. Reis & C. M. Judd (Eds.), Handbook of research methods in social and personality psychology (pp. 339-369). Cambridge: Cambridge University Press. This study stipulates that correlation items within the construct should be greater than outside the construct. 2. Ratner, B. The correlation coefficient: Its values range between +1/−1, or do they?. J Target Meas Anal Mark 17, 139-142 (2009). doi.org/10.1057/jt.2009.5 Values between 0.3 and 0.7 (-0.3 and −0.7) indicate a moderate positive (negative) linear relationship.

[@educationalinsightskarachi4969]

kzbin.info/www/bejne/q3Kck2pvjtpgY8U I have calculated the HTMT ratios based on these correlations in this video based on the following reference. Henseler, J., Ringle, C.M. & Sarstedt, M. A new criterion for assessing discriminant validity in variance-based structural equation modelling. J. of the Acad. Mark. Sci. 43, 115-135 (2015). doi.org/10.1007/s11747-014-0403-8

[@usra9983]

Many thanks for your video , can you please explain how we report this in a written manner ( to publish it in a scientific Journal ). Should we do the PCA for each measure scale alone ? or for all the scales in the same time and how we interpreted the results of the PCA ?

[@educationalinsightskarachi4969]

thank you so much for your kind words. I will share a link to the written interpretation soon.

[@usra9983]

@@educationalinsightskarachi4969 many thanks it's much appreciated

[@kossonouprunelle7576]

PLEASE, Is exploratory factor analysis the same as convergent and discriminant validity? I am a little bit confused. Thanks in advance for your reply. Should we do both convergent and discriminant validity, construct validity using EFA in one study?

[@educationalinsightskarachi4969]

Convergent and discriminant validity is a part of CFA. You may use it after EFA as a part of CFA.

[@rosacapallo7517]

Hi, any specific reason for selecting "one-tailed"; what if we choose to do it with two tailed? Regards

[@WillowSpector]

would love to know as well

[@educationalinsightskarachi4969]

We use one-tailed when testing directional hypotheses

[@wong4728]

Follow up video is needed on reporting findings. The actual report.

[@educationalinsightskarachi4969]

Thanks for your suggestion. I definitely do it soon

[@educationalinsightskarachi4969]

kzbin.info/www/bejne/q3Kck2pvjtpgY8U I guess you would like this video on HTMT ratios for the assessment of discriminant validity.

[@juliagabriel7452]

Is the validity also affected by the size of sample? I use the sample size of 30 and it shows less correlation among several indicators in some variables.

[@educationalinsightskarachi4969]

It is slightly improved. But it is mainly affected by wording of questions

[@shafacindanusabila5229]

Hello, thank you for the video. I have a question. I saw your another video, it explains how to assessing discriminant validity using HTMT or Fornell-Larcker. Are we just choose one between this method, HTMT, or Fornell-Lacker? Or, is there some condition when we can use this or that? Can I just use this method for assessing convergent and discriminant validity? Thank you in advance.

[@educationalinsightskarachi4969]

Fornell Larcker condition was developed first. Later, in 2015, HTMT ratios approaxh emeregd and refuted the use of Fornell Larcker. However, some journals might requie one particular assessment.

[@shafacindanusabila5229]

@@educationalinsightskarachi4969 what about using the pearson's correlation?

[@educationalinsightskarachi4969]

@@shafacindanusabila5229 ​This is still a valid approach. The HTMT ratio assessment is an advancement in the approach of assessing the correlation of items within a construct and outside the construct. John, O. P., and Benet-Martínez, V. (2000) proposed comparing items correlation of the items. Later, Henseler et al. (2015) developed the HTMT ratio which is basically a ratio between the correlation of items within the construct and the correlation of items with the items of other constructs. John, O. P., & Benet-Martínez, V. (2000). Measurement: reliability, construct validation, and scale construction. In H. T. Reis & C. M. Judd (Eds.), Handbook of research methods in social and personality psychology (pp. 339-369). Cambridge: Cambridge University Press. Henseler, J., Ringle, C.M. & Sarstedt, M. A new criterion for assessing discriminant validity in variance-based structural equation modeling. J. of the Acad. Mark. Sci. 43, 115-135 (2015). doi.org/10.1007/s11747-014-0403-8

[@educationalinsightskarachi4969]

@@shafacindanusabila5229 Probably, you can first write an interpretation of Fornell Larcker's condition. Then, compare the correlation of items. Finally, close with HTMT ratios.

[@chitrasingh9800]

nice

[@educationalinsightskarachi4969]

Thanks

[@gigi-km5tj]

Thanks but the video lacks clarity

[@educationalinsightskarachi4969]

I will enhance video quality in future