when we do Factor analysis, the response of the Likert scale types of questionnaire would be 1,2,3,4,5. so when we feed the response the the answer will be either 1,2,3,4,5. but some times, these response might have weighted scores. example in SGRQ , This has weighted score for each response. so what score we have to use for factor analysis?
For factor analysis, you typically use the raw Likert scale scores (e.g., 1, 2, 3, 4, 5) unless the specific instrument (like the SGRQ) provides a weighted scoring system. If weighted scores are part of the validated scoring procedure of the instrument, use those weighted scores.
References:
1. Field, A. (2018). Discovering Statistics Using IBM SPSS Statistics. Sage Publications.
2. Costello, A. B., & Osborne, J. W. (2005). Best practices in exploratory factor analysis: Four recommendations for getting the most from your analysis. Practical Assessment, Research, and Evaluation.
You mention that the scoring of your questionnaire is 1-5. this is often indicated as numbering of the questions on the answer sheet.
However, if the scale rates something which goes for example from 'not at all' to 'very much', the lowest category logically shoud be 0, not 1. This does not influence the EFA, but it is easier for interpretation.
Like if you compute a sum score from at 32 item quetionnalire, the lowest score would be 32 with the 1-5 scoring, but the more interpretable 0 with a 0-4 scoring.
I agreed Ahmed Shiyam If your scale rates from 'not at all' to 'very much', using 0-4 instead of 1-5 can make the interpretation of scores more intuitive. For instance, a 32-item questionnaire with a 1-5 scale has a lowest possible sum score of 32, whereas with a 0-4 scale, the lowest score is 0, which directly reflects the absence of the measured attribute. This adjustment simplifies understanding and communicating results without affecting the Exploratory Factor Analysis (EFA).
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics