Hello
I am investigating the effect between emotional state and free sampling product evaluation and am currently at a standstill in terms of my statistical analysis.
Data was collected through an experiment. Participants were shown one of three possible videos: a positive one, neutral one or negative one in order to influence their emotional state. This was measured using the PAD (pleasure, arousal & dominance)-scale.
Participants were asked to taste a sample and evaluate it.
Video = one of three, nominal variable
Emotional state = 6 questions, each likert scale 7-items
Product evaluation = 11 questions, each likert scale 7 items
ANOVA (+ contrast testing) between video and product evaluation was not significant for none of the tests. ANOVA (+ contrast testing) between video and emotional state was significant for all tests. Correlation between emotional state (sum of question values) and product evaluation (sum of question values) was also significant.
In short, I did not find a significant direct relationship (video seen and product evaluation), but did find significant intermediate relationships (video => emotional state => product evaluation).
What might be the cause of this? Did I choose my statistical tests incorrectly? Should I be looking for other variables I have not controlled for?
Thank you in advance
(This is the first ever question I put on this forum. Apologies if it is unclear or lacks details. I am more than happy to provide whatever is necessary.)
Kind regards