I run repeated measure ANOVA with between subjects (gender) as independent variable to test its effect on the willingness to pay (WTP) for two conditions A and B (within-subjects) where the same participants provided their willingness to pay for each condition.
The result of ANOVA was as following:
WTP
Hello Areej,
Since (a) you've judged the gender x WTP interaction to be non-significant, and (b) observed a significant main effect for WTP, there really was no need for simple effects tests of WTP by level of gender...the results turned out just as one would predict from (a) and (b).
However, your questions appear to overlook the chief point (at least, in my opinion): What was the hypothesis/es you were intending to test in the first place? That, coupled with the nature of the scores and the data collection method, is what should govern your analysis method.
As you indicate that WTP level comparisons within gender levels is a principal focus, I'd suggest you:
1. Report and interpret the overall tests (e.g., interaction, then main effects of gender, WTP).
2. Report and interpret the first two simple effects tests if you wish. Perhaps you find them useful for purposes of comparing relative effect sizes. Note that these just affirm the (a) and (b) points above.
The simple effect tests you describe in Question #3 are actually interaction tests...yet you've already established you believe that there was no noteworthy interaction. So, these wouldn't add to the results.
Finally, I'm not sure it makes the most sense to ignore other salient IVs in your model (e.g., to run multiple anovas, sorting out IV inclusion in each).
Good luck with your work.
Thank you very much David Morse , that was helpful.
Its an exploratory study, we do not have a hypothesis. however, we want to examine if there are differences between condition A and condition B and whether these differences hold across gender, education level, income level and age groups. At the same time, examine possible association between these demographic factors and WTP for the two condition (clearly no association considering the interaction result).
ANOVA was the best choice -so far- to answer the above research questions. As mentioned in the main question, including other IVs in the same model was not possible due to sample size (250 samples). including some of them in one model would also result in serious ANOVAs to find possible interaction (when included gender, education level and income level in one model; gender interaction became significant).
Running a repeated measures ANOVA with gender as the between-subjects independent variable and willingness to pay (WTP) for conditions A and B as the within-subjects variable is an appropriate approach to analyze your data. Here's a general outline of the steps involved in conducting this analysis:
Remember, it's crucial to consult a statistical expert or refer to appropriate statistical resources to ensure the accuracy and validity of your analysis.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics