Greetings,
I am a DBA student conducting a study about "Factors Impacting Employee Turnover in the Medical Device Industry in the UAE."
My research model consists of 7 variables, out of which:
- 5 Variables measured using multi-item scales adapted from literature ex. Perceived External Prestige (6 items), Location (4 items), Flextime (4 items),.. etc.
- 2 are nominal variables
I want to conduct a reliability analysis using SPSS & I thought I need to do the below?
Am I thinking right? Attached are my results up to now..
Thank you
Hi Ali Mossarah
you are on the right track and your results look well-organised. Let me address your questions one at a time:
Conduct reliability test using SPSS Cronbach's alpha for each construct (except for nominal variables)
--> This a certainly a viable option although there are better alternatives to Cronbach's Alpha, McDonald's Omega for example (available in SPSS 28). See:
McNeish, D. (2018). Thanks coefficient alpha, we'll take it from here. Psychological Methods, 23(3), 412–433.
Schmitt, N. (1996). Uses and abuses of coefficient alpha. Psychological Assessment, 8(4), 350–353.
Deal with low alpha coefficients (how to do so?)
--> "low" alpha values (typically < .7) are not always bad. Check what values are to be expected in your line of research and how high the items within each factor are correlated. Ultimately, you may remove certain items to boost the alpha value, but do not do so just data-driven.
Conduct Exploratory Factor Analysis to test for discriminant validity
--> You have got it the wrong way round. If you want to include EFA, do so before estimating Cronbach's Alpha. The EFA would probably have shown that the items you designed for "Location" and "Job Satisfaction" do not load onto their intended factors, hence the low alpha values. Naturally, once you have the factor structure straightened out, high reliability coefficients are more likely, but EFA and reliabilty coefficients support rather than substitute each other (i. e. a low alpha value alone does not tell you anything about the distribution of items to factors, it just tells you that the internal consistency is insufficient).
Hulin, C., & Cudeck, R. (2001). Why conduct a factor analysis and then compute an alpha? Journal of Consumer Psychology, 10(1/2), 58–59.
What is more, the EFA's main objective is not to test for discriminant validity, but to identify constructs among your items and have those constructs account for variance on the manifest indicators.
Best of luck
Marcel
People use the word "reliability" in lots of different ways, so if you say precisely what you mean it will help in addressing your question.
First, you can run a reliability analysis on these multi-item scales including all items at the same time and checking for Cronbach's Alpha coefficient. This value should be greater than 0,70 for a scale to be reliable.
In case you face low reliability values, you can run some descriptive statistics to check for the "Crobach's Alpha if Item Deleted". If there is any item that decreases the reliability, you can decide to withdraw that item and not include in the analyses.
You conduct Exploratory Factor Analysis to reduce the number of items into factors and to reveal the structure among items. Since it will be difficult to work with a big number of items, EFA will group items with similar characteristics under a factor and after that, you can work with factors' means.
Meanwile, you will need to check the KMO and Barletts tests and Total Explained Variance.
WoW Marcel Grieger, thanks a lot for your detailed answer, please check the summary below.
Daniel Wright, I am conducting my reliability analysis based on the questionnaire that I sent to 23 participants that made up my pilot study.
Kujtim Hameli, thanks, and please check below too.
So, I conducted my pilot study and got these results with the 2 low Cronbach's Alpha coefficients I mentioned above. I have checked if I can remove some items to boost the alpha value, but unfortunately, none does.
The question is now: Do I edit the questions for these 2 variables and run the pilot study on a different group and check if I get a good Cronbach's Alpha coefficient? or do I keep them and just mention in my research that they had a low coefficient and will run descriptive statistics for them only?
Thanks all,
Ali
I personally use AMOS add on for spss.
My workflow is as follows.
1. EFA followed by deletion of low loading scores.
2. CFA followed by reliability test and model fit
3. Pathway analysis and model fit.
Regards
Tariq Kadhim
please stop spamming the same non-answer under every question. You are wasting everyone's time.
Tariq Kadhim
most hilarious explanation for spam I have ever heard. Congratulations, you have made my day.
Sorry, i have to go back to this.
As you can see in the picture, there are sub-factors that make up the factors.
As i stated above, i conducted a reliability test using SPSS Cronbach's alpha for each construct (except for nominal variables), but these were for the sub-factors.
Do you think i should run the same for the factors as a whole?
Thanks,
Ali
you have computed cronbach alpha for each item and c.alpha more than 0.7 indicated their is internal consistency in each item of the questionnaire and can be used as a validated tool provided if you have assessed validity.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS