I've developed a work personality inventory and want to make sure my questions are valid. For each question, there is a choice of 8 responses corresponding to the different work personality types. I have asked a pilot sample to select their first, second and third most preferred answers but now I'm stuck. ts similar to a mostly A's are a certain type, mostly B's are another type and mostly C's are another type ...x 8. But now I want the responses to be mixed up. How do I assign who fits into which personality type? Please refer me to any articles, references or guides? Any advise would be so appreciated
Jeanine Blumenau you seem to have an interesting study, about work personalities. Keep in mind there is a lot of published studies about personality theories and a priori instruments. You ought to review those and learn how an instrument is constructed and also to ensure your effort is unique (not already accomplished). That being said, I would have to know your RQ and hypotheses to help in detail. Generically I can offer similar comments as another poster (see https://www.researchgate.net/post/Gathering_data_concerning_online_education_management_practices_of_instructors_technology_and_learners?_ec=topicPostOverviewQuestionRecommendations&_sg=gE6txnxZIoDqKPWriTwsrFZpkcl7E_EiVywIyMLwfU9Fx5xpTnMSVFo1xZR-oprRM6ZihuqjeoIe2aO1)
The instrument construct (factors and relationships) dictate what and how the anlaysis should be done, what technique etc. based on the RQ and hypotheses. Each technique requires assumptions be met, and certain data types are needed. For example letters (nominal data types) as scales are usually not good for quantitative analysis - although you could use a program in R or alternative to convert those to metric. Depending on your research design, maybe you are exploring an unknown area and machine learning or nonparametric technique could work. Not sure. Here is a model from Strang (2015) which shows some of the statistical techniques and how they match to research design strategies. Let me know if you have more questions. Ken
Strang, K. D. (2015). Selecting Research Techniques for a Method and Strategy. In K. D. Strang (Ed.), Palgrave Handbook of Research Design in Business and Management (pp. 63-80). New York: Palgrave Macmillan. Retrieved from https://www.palgrave.com/us/book/9781137379924. doi:10.1057/9781137484956
I believe you are using a "multiple response" format for your items, which is very difficult to analyze statistically.
Hi Jeanine, Your survey design is pretty complicated. Maybe make each of your questions into 8 questions, so item 1 becomes 1a, 1b, ..., 1h.. In cells put each participant's choice: 1 for 1st, 2 for 2nd, 3 for 3rd, and 4 for the rest. You can do a factor analysis and you're hypothesizing all the a's load together, all the b's together, etc. Hopefully you have 8 reliable scales. If people actually are types, you can make a histogram of each scale and they should be bimodal, so you have one group of low scorers (if 1 is 1st choice) and then a much larger group of high scorers (with almost all 4's). Hope this helps. ~ Katie
Dear Jeanine,
I agree with the other respondents. My contribution is to state that questionnaires are used to establish and confirm theories of personality, not just to measure predetermined theories. How do you know that these 8 personalities exist, are measurable, are mutually distinct, and each question you have asked is equally important?
I suggest you study the literature on the validation and reliability of high quality psychological personality type measures, such as 16PF which use oblique multimensional analysis.
- Badges
- Science topic