Could anyone help with a bit of statistics:
There are three coping styles in the research (emotion focused, problem focused and avoidance oriented) and three dimensions of burnout (emotional exhaustion, personal achievement, depersonalisation).
Should I use two-way or three way ANOVA to test the intterrelationships between coping styles and each scale of burnout?
When I intend to study e.g., the Emotion Focused x Problem Focused Coping vs Burnout relationship should I use two-way anova (considering all its requirements and assumptions)?
I recently found a study using three way anova to test the following relationships:
Problem Foc x Emotion Foc vs Burnout
Problem Focused x Avoid Foc vs Burnout
Avoid Foc x Emot Foc vs Burnout
but theese sound like the relationships that need only two-way anova, used each time for these relationships;
If an ANOVA is a possibly helpful tool at all depends on the quality of your response ("burnout"). How is it measured? Is it a metric variabe where you can expect symmetric deviations of individual data around the expected values?
From your description I think that the coping styles are not three independet variables. I think if a subject has a very strong emotional focused coping it can not have at the same time a very strong problem focused coping style, right?
A three-way ANOVA is better than other ANOVAs. But I would suggest using a three-way MANOVA.
The scale(s) of your response(s) is/are not metric (it is "only" ordinal!), and symmetric errors is not a sensible assumption (bounds/limits). Therefore, ANOVA will be not the right tool anyway. There are special procedures how to best analyze (7-point-)Likert scales. I am not an expert there.
- Badges
- Science topic