Dear all and any,
I have run a mediation analysis on SPSS using the PROCESS extension (model 4). My results don't really make any intuitive sense, where my ID (availability of healthy fast foods, has a negative association between psychosocial determinants in regards to eating healthy fast food options, where theoretically it should either have no association or a positive one.
Im wondering therefore
1) is this an inconsistent mediation / do my mediators act as suppression variables?
2) how would i interpret these in my results section?
3) what I can actually conclude from this?
independent variable: availability of healthy fast foods
dependent variable: intentions to consume healthy fast food options
I have 4 mediators: Cognitive Attitudes (CA), Affective Attitudes (AA) , Capacity PBC, Subjective norms (SN)
RESULTS
my total effect (path c) : negative and significant
path a's (x on m) : CA = negative , non-significant / AA, Capacity PBC, SN = negative & significant
path b's (m on y) : all positive and signifcant
direct effect (path c'): positive and significant
inidrect effects: CA: negative / non sig AA, SN, Capacity PBC : negative, significant
My thesis is due tomorrow, I'm so desperate, PLEASE HELP!!!!!!!!!!!!!!! I would really appreciate it
Hello Iara,
Yes, it can be frustrating when results don't go to plan. However, that can often open doors for understanding influences that have previously gone unidentified. Based on what you posted:
1. The direct effect (path c') makes sense as a positive relationship;
2. Direct links of HFFs availability to CA, AA, PGC, and SN are consistently negative; doesn't that make sense?;
3. Direct links of CA, AA, PBC, SN to Intention to consume are all positive...are these the unexpected relationships (since the product of the terms for estimating indirect effects will therefore all be negative)? If so, I would have to know more about how you quantified each of these 5 variables (CA, AA, PBC, SN, Intention).
Without any other information, all I can suggest is that, for your sample, the measures you used (and their quantification), these are the results you obtained. I doubt that suppression would be the answer for all 4 indirect relationships being negative in sign.
Good luck with your work.
Hi David Morse , thank you so much for your response.
1. yes, this would theoretically make sense, though what does it actually mean when comparing it to the total effect?
2. not so much, i may have to think of an interpretation for this. though statistically, it does make sense considering the total effect is is negative
3. no these are expected, at the cognitive factors relate to attitudes, etc etc in regards to eating healthier fast food options (should have mentioned that before, sorry)
Thank you for your help!!!
- Badges
- Science topic
- Similar topics
- Thesis Research