I discovered the undiscriminant validity between the two constructs when I observed that the values of MSV are higher than AVE for each factor; also the factor analysis _reduction test (validity using SPSS) the output shows that the two constructs were found in one component and the loads are in one components as well.
How to settle this. Should I combine them ? or remove one of them ? and How?
I am using SPSS and the whole model would be submitted to SEM using AMOS 16.
Thanks
There are two ways of looking into the matter. First, the theory should support the discriminant nature of the factors. Secondly, if the theory supports the discriminant nature of the factor but your data reflects something else. It indicates the need to check your data for any biasness or errors.
NEXT, since you are using SEM using AMOS you can check for discriminant and convergent validity by performing CFA. Also, a plugin by James Gaskin can be added to AMOS which helps in checking validity and also suggests you possible cause of 'Undiscriminant validity'. Based on the version of AMOS you can download and use the plugin 'MASTER VALIDITY' from this link https://drive.google.com/drive/folders/0B3T1TGdHG9aEbFg1eEpqOWtrR3c
Discriminant validity :-
In psychology, discriminant validity or divergent validity tests whether concepts or measurements that are not supposed to be related are actually unrelated.[1]
Campbell and Fiske (1959) introduced the concept of discriminant validity within their discussion on evaluating test validity. They stressed the importance of using both discriminant and convergent validation techniques when assessing new tests. A successful evaluation of discriminant validity shows that a test of a concept is not highly correlated with other tests designed to measure theoretically different concepts.
In showing that two scales do not correlate, it is necessary to correct for attenuation in the correlation due to measurement error. It is possible to calculate the extent to which the two scales overlap by using the following formula where {\displaystyle r_{xy}}📷 is correlation between x and y, {\displaystyle r_{xx}}📷 is the reliability of x, and {\displaystyle r_{yy}}📷 is the reliability of y:
{\displaystyle {\cfrac {r_{xy}}{\sqrt {r_{xx}\cdot r_{yy}}}}}📷
Although there is no standard value for discriminant validity, a result less than .85 tells us that discriminant validity likely exists between the two scales. A result greater than .85, however, tells us that the two constructs overlap greatly and they are likely measuring the same thing. Therefore, we cannot claim discriminant validity between them.
Consider researchers developing a new scale designed to measure narcissism. They may want to show discriminant validity with a scale measuring self-esteem. Narcissism and Self-esteem are theoretically different concepts, and therefore it is important that the researchers show that their new scale measures Narcissism and not simply Self-esteem.
First, we can calculate the Average Inter-Item Correlations within and between the two scales:
Narcissism — Narcissism: 0.47Narcissism — Self-esteem: 0.30Self-esteem — Self-esteem: 0.52
We then use the correction for attenuation formula:
{\displaystyle {\cfrac {0.30}{\sqrt {0.47*0.52}}}=0.607}📷
Since 0.607 is less than 0.85, we can conclude that discriminant validity exists between the scale measuring narcissism and the scale measuring self-esteem. The two scales measure theoretically different constructs.
Hi Ayman,
It appears that your data points to existence of a second order construct, i.e., the two latent constructs you've measured actually belongs to one higher latent construct. Check that your theory actually supports this or any emerging models that explains this. To be safe, make sure you have followed each step in doing SEM to ensure this is not an error - commonalities, multicollinearity, etc..
Cheers.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS