Hello everyone,
I've got a question regarding deciding whether the short vs full questionnaire is better at predicting or showing an association with a variable of interest. I've used the full and the short form of the Five Facets of Mindfulness (FFMQ) to look into its association with various variables of interest (ranging from ordinal to continuous). I get not only different magnitudes of associations (e.g., rho = .3 vs rho = .5) but I also get different number of significant factors (there are five factors in the FFMQ) depending on whether the short of the full version of the questionnaire was used. This suggests that the addition or removal of some of the questions in having quite a significant difference. What tests would you propose I use to get to the bottom of this? Is there a test that can tell me with some assurance whether the short or the full FFMQ is more appropriate?
P.S To complicate things further, the full FFMQ questionnaire has around 500 participants while the short form has around 1000.
Regards,
Achilleas
Any formal scale / test / questionnaire is designed after months or even years of testing. Validity and reliability is tested rigorously so that it can withstand the test of times for a long period of time. Such scales etc made at one point in time may not give the same result after say 15-20 years primarily due to various and obvious changes in the society and the way people think and behave. If you are using the FFMQ ( I have not used it so far though) for a certain specific purpose ( I assume that is why you want to use a shorter version) then you need to use only that portion which is relevant for your research. If you want to measure all the factors / parameters that the full test is designed to do through a shorter version you may encounter anomalies as you have mentioned. I recommend using the full version and if it is being used for measuring any specific attribute then only that relevant portion to be administered to the subjects.
Hello Kamal Karamchandani ,
Thank you for your response. My results make more sense to me when using the full FFMQ questionnaire than the short FFMQ. If anything, I am hoping to prove that the full FFMQ should be chosen as it is more relevant to the variable I am trying to predict. To note, the short version of FFMQ has been validated and used a couple of times in the literature, I am not the one who created it.
I'm looking into lucid dream frequency and mindfulness and I've collected FFMQ and LDF in various questionnaires that I've put out over the years. One of the questionnaires I gave out used the short version of FFMQ (24 questions) while all the rest had the full FFMQ version (39 questions). Hence now have around 1000 participants with short FFMQ consisting of the questionnaire that used only the short FFMQ form + taking only the 24 questions of the full FFMQ version from the other 500 participants who completed the full version.
My question is whether there are some statistical analyses that I can do to support using the full FFMQ questionnaire over the short FFMQ, despite the short FFMQ having a higher number of participants.
Thank you for your help in the matter.
So basically is there such a statistical analysis I could do or should I just discard the short version of FFMQ results out of principle that using the full version is always more valid?
That you need to decide as per your research design, hypotheses, aim, and objectives. If you think the longer version is appropriate to use, then apply the statistical analysis on the results so generated.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs