It is my first time using SMARTpls and I stumbled upon a few issues that my books/readings do not seem to answer. The below are the issues. Thank you in advance for the guidance and advice. The below questions are for reflective measurement model.
1) According to literature, rho_a should be in between the values of Cronbach Alpha and rho_c composite reliability. However, when deleting some indicators, the Cronbach Alpha increased together with the rho_c Composite Reliability but the rho_a decreased. Why did this happen? Is there an explanation to this.
2) Occasionally deletion of an indicator leads to increase in composite reliability but decrease in Cronbach Alpha and rho_a. Is there a reason to why some test for estimated reliability increase while others decrease? does it indicate something about my data.
3) Regarding the process of deleting indicators. According to literature, the loadings/weights below 0.7 should be deleted. However, if such deletions decrease AVE or composite reliability, it should not be deleted. Is this correct? In that case, am i allowed to delete the next lowest (loading/weighted) indicator in the construct (meaning the second lowest) to increase AVE to above 0.5 threshold? Or am i able to delete each indicator to see which provides the best composite reliability and AVE. What other rules should i be aware of.
4) Am I allowed to switch the indicators to other constructs if it reduces discriminant validity and "makes sense". For example switching a conceptually similar indicator to the other construct when there is discriminant validity issues between the two constructs. After switching that one indicator, the discriminant validity improved quite a lot. Hence is this allowed?
5) Lastly, may i know how the weights and loadings are calculated for SMARTpls? How does the software calculate the weights? Does it measure the similarity of each participants answers? or is it based on scoring (1-5 likert scale, with an average of 5 being highest weight).
- Badges
- Science topic
- Similar topics
- Psychology
- Mental Processes
- Learning