Hello everyone,
I'm writing because I'm in stuck with my master's degree thesis in Marketing, I would be so grateful if someone could help me to find out my problem.
I have the following situation in my study:
Dependent variable: Purchase intention (Mesaured with a Likert scale 1-7, 3 items)
Mediator: Perceived quality (Mesaured with a Likert scale 1-7, 3 items)
Independent variable: I have 2 independent variable, Saturation of packaging color, Packaging transparency
I submitted a questionnaire between subjects with 4 scenarios, every respondent replied only for a single scene. I had 4 visual stimulis (More saturated pack, less saturated pack. Transparent pack, no-transparent pack). The 2 independent variables don't talk to themselves, they're not related
These are my main hipothesis:
1) More saturated pack brings an higher purchase intention than less saturated pack
1a) More saturated pack brings an higher perceived quality than less saturated pack
2)Transparent pack brings an higher purchase intention than a no-transparent pack
2a)Transparenct pack bring anhigher purchase intention than a no-transparent pack
I have the idea to set a Independent samples T Test to study my hipothesis. I have the following doubts:
1)Should I create a dummy variable for my Independet variables (Example: 1=more saturated pack, 0=less saturated pack)?
2)What should I do with my Dependent variable and mediator? I have, for each Indep.var., a mean value for each scene (I have the Mean purchase intention for More Saturated pack respondents, and a Mean purchase intention for Less saturated pack respondents), what have I to do with these? Should I create a dummy variable for this too?
Please help me, I'm in trouble with this since late August and my thesis deadline is coming in few weeks.
Thank you to everyone who will help me
Dear Andrea Callegari ,
You are using the wrong terminology. Both purchase intention and perceived quality are dependent variables. Saturation and transparency are grouping variables. Yes, you can run for one sided independent t tests provided that their normality assumption is met. A Levene's test will indicate whether you can assume equal variances. You may need to manage your Type I error risk by reducing your significance threshold. Please see my study guides for more information.
In case you do not meet the normality test it is recommended to use Matt Wittney's U, you can also evaluate differences from the perspective of factor analysis of structural equations, where significant effects refer to specific differences.
Let me check if I understand. You had four groups defined by a 2x2 between subject design. You want to see if these predict purchase intent and you believe difference in purchase intent may be mediated by perceived quality. You would want to see the relationship between perceived quality and the 2x2, and then look for differentiating the 2x2 direct effect on purchase intent and the effect mediated by perceived quality. It is that a correct summary? Are the mediator and dv variables the same for all people? Have you created summary measures for these (from the 3 measured variables)? What are the summary measures distributed like?
Hi everyone, thank you for you help. @Daniel Wright my study is not a 2x2, my independent variables don't talk to themselves, I got 2 of them and for each independent I have 2 scenes. So they judged my image evaluating perceived quality and purchase intention. Every respondent answered to the same questions about all the variables. The mediator is the perceived quality and the main dependent variable is purchase intention. I created for each respondent a Mean Value about these variables (I had Likert scales, so I made the mean of their items for each respondent. Now I have the Mean Purchase Intention and the Mean Perceived quality for each respondent)
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS