I'm using Leventhal's Common Sense Model of Illness Perception to explore the relationship between illness perceptions and fatigue severity outcomes in cancer patients, using selection of positive and negative coping strategies as mediators.
I'm having trouble interpreting the results, as the c and c' pathways are opposite in sign from the indirect effect (a*b pathway). Furthermore, neither the c or c' paths are significant, but there is a significant positive indirect effect for positive coping as a mediator.
So if I'm correct, this means that greater patient sense of personal control over their cancer experience is associated with greater fatigue, which is accounted for by greater use of positive coping coping strategies. But this doesn't make sense, and the c and c' pathways indicate that greater sense of personal control is associated with reduced fatigue.
Any advice is appreciated!
From a purely statistical perspective I would alter you description of your result slightly as follow:
There is an overall negative, but not statistically significant effect of a sense of control on the level of fatigue. This is a result of a positive (and statistically significant) indirect effect via Positive Coping and two negative (but not significant) effects, namely a negative indirect effect mediated by Negative Coping and negative direct effect. The overall (total) negative effect (-.12) results from the combined effects of the latter two latter two negative effects (-.02 and -.19) being greater than the single indirect positive effect (+.10), mediated by Positive Coping.
There is no contradiction in the fact that the combined effect of two negative non-significant effects can be stronger that the single statistically significant positive one.
From a more substantive perspective, the interpretation of the results are less clear. One could state that the positive indirect effect via Positive Coping acts to lessen the otherwise stronger negative effect of Personal Control on Fatigue. But I agree that the positive effect of Positive Coping on Fatigue is difficult to interpret.
Incidentally, the indirect effects should be equal to the product of the two component paths, but the numbers in the diagram differ from this, more than can be explained by round-off errors. But nevertheless the main issue is clear.
You may also try stratifying the sample by age, type of diagnosis/prognosis, or other factors depending on your sample size.
- Badges
- Science topic