For a newly-developed scale for a construct similar to cronyism, I conducted a survey questionnaire of the new scale items and included additional items to ascertain convergent, divergent (discriminant), and predictive validity. I administered the questionnaire in two waves, and am currently working on cleaning the responses and matching them.
What is the best method to analyse the data and confirm each validity?
Convergent and Discriminant validity are subsets of construct validity can be established using correlation coefficients. You may want to read about multitrait–multimethod matrix, nomological network, and pattern matching. These are some techniques for convergent and discriminant validity. Predictive validity requires the administration of a gold standard and the developed measure at two different point in time and you would also need an outcome/criterion measure.
Hi Mohammad Abdullah ,
Here is a summary answering your question from my experience:
- Convergent validity is assessed by calculating AVEs, i.e., Average Variance Extracted for each construct and should be > .50. Further, CRs, i.e., composite reliability can support the evidence of convergent validity.
- Discriminant validity is assessed by whether or not the square root of AVE is larger than the correlations amongest constructs. If yes, the measurement model is discriminantly valid.
- Predictive validity is assessed via the use of the new measurement model test hypotheses backed by relevant literature.
Best,
A
-------------------
Recent publications:
Article Email is Evil! Behavioural Responses towards Permission-base...
Article Validating a New TQM-Benchmarking Measurement Model in an In...
Article Total Quality Management Boosters and Blockers in a Humanita...
Article Political Advertising Effectiveness in Wartime Syria
Article When empathy hurts: Modelling university students’ word of m...
- Badges
- Science topic
- Similar topics
- Publishing