Hi,
I'm doing a pre-test to see whether my manipulative frames work as intended. I'm manipulating valence (loss versus gain) and temporal orientation (proximal versus distal) in a 2x2 design, and measure valence and temporal orientation with 3 semantic differential statements each. E.g. for valence: "this message emhasizes negative outcomes (-3) .... positive outcomes (+3)
Should I use a one sample t test to see whether the the means for the loss frames are below 0, and for gain frames above zero? And then again for temporal orientation?
Or should I use an ANOVA?
Glad with any help you can provide
If you have quantitative data, ANOVA is a good statistical technique for this case provided the data meets the assumptions of ANOVA. t-test can be used for post-hoc analysis if F-value is significant. Also, ANOVA yields interactional effect enabling you to infer more from data.
Thanks for all your answers. In similar studies, I found that ANOVA was used. However, my argumentation is that an ANOVA is used to test whether there is a difference between a loss and a gain frame, not whether a loss frame is actually perceived as negative. E.g. if gain frames score 2.5, and loss frames score 1.5 on the positivity scale, than ANOVA would find a significant difference, whereas a loss frame is not actually perceived as negative (not
- Badges
- Science topic
- Similar topics
- Condensed Matter Physics
- Zero