SPSS: Fornell-Larcker condition

[@acerrman]

You are a life saving GEM!!! Thank you!!!

[@educationalinsightskarachi4969]

[@Adsila88]

She made things easy, I have been searching and searching

[@toobakhalid9236]

maam, your student from cbm, love and prayers for this helpful video!

[@educationalinsightskarachi4969]

Wow... that's great. Thank you so much for your kind words

[@saharmona785]

Well Explained.. ..

[@davidy6577]

Thank you for your helpful video! If one variable does not pass the discrimant validity test (like you showed in this video), then what options do we have? 1) Should we drop them? 2) if so, how do we determine which variable to drop? 3) if not, how do we report this in the paper in APA format? Would you please provide some examples of how to report this? Thank you

[@educationalinsightskarachi4969]

Discriminant validity is assessed pair-wise. As per rule, if two variables are not Discriminant, they should be merged.

[@educationalinsightskarachi4969]

Alternatively, we canncheck the correlation of the items. Only drop the items that have high correlation with the items of the other construct.

[@abdullahkhali]

Thank you so much. That’s really informative! I am interested to use this technique to calculate the convergen and discriminate validity of within the same assessment A. My question is that ok to do that or I have to the compare and calculate convergent validity to assessment B and compare the discriminante validity to assessment C.

[@ArminMoradi-k6x]

if we have two numbers in row of rotated compnent matrix, which of these numbers should place in firs column?

[@stelladeluz2746]

Thank you so much should all items be correlated at a significant level? Or is it ok if some items do not fulfil this condition?

[@educationalinsightskarachi4969]

Generally speaking, all constructs should correlate

[@educationalinsightskarachi4969]

However, there can be exceptions. Independent variables might not correlate. Moderator might not correlate with independent variables.

[@mihithapiumal3512]

Thanks a lot

[@davidy6577]

Could you share 1) the reference for Fornell-Larcker condition 2) examples of how to report the result (whether valid or not) in the paper in APA format? What information should we provide int eh paper itself?

[@educationalinsightskarachi4969]

I will share a written sample with you soon.

[@davidy6577]

@@educationalinsightskarachi4969 Thank you SO much!

[@Tom13961876]

Hello, I did the factor analysis and for the rotated component matrix I got variables without any component value, some variables with more than one component value and also variables with a negative component value. Is it still possible to test for discriminant validity and how can I continue?

[@educationalinsightskarachi4969]

Negative component loadings occur when a variable is inversely related to a factor or component. While they can be somewhat counterintuitive, they are not uncommon in factor analysis. Negative loadings suggest that the variable is conceptually opposite to the direction of the factor it loads on. Please also check if these items need reverse coding.

[@Tom13961876]

@educationalinsightskarachi4969 thank you for your answer! Is it still possible to test for discriminant validity with negative loadings or is it only possible with positive loadings?