Hi Folks,
1. All things being equal, What is the best way to combine ordinal, binomial and interval items into a composite score/index as one of many other variables for later use in a multivariate correlation and regression analysis.
Options 1
1. Convert items into the lowest level of measurement i.e. Interval and Binomial into ordinal levels of measurement
2. Re-scale all items by normalizing using feature scaling into a range between 0 & 1
3. Find the average to generate a score
Options 2
1. Standardize [z-score] all the items
2. Sum [or average ?] them up to generate a overall score
Other questions
1. What informs the choice to either sum or average when generating an overall score
2. What is a relative score and how exactly does one generate a relative score for items as above i.e. ordinal, binomial, interval items
Läs pallant's survival manual från kapitel 8 framåt. Där finns alla råd som du behöver.
Dear George Tetteh Hadjor,
Currently, I am facing the same difficulty as you described in your question above (combining ordinal, dichotomous, and ratio items in a composite score) and am unsure of how to deal with this. Did you find an answer to your question? If so, would you be able to share that with me?
Thank you in advance and have a lovely day!
With kind regards,
Nina
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling