For the research purpose I developed 4 surveys, where each of them measures the same constructs/items, but under different requirements (treatments). To distribute the survey links, I visited classes (1-4 semester) of the faculty I belong to and then randomly assigned students in to the groups.
I aim to analyse the gap between the 4 groups, to find differences I assumed in my Hypothesis. Tukey? LSD? Bonferroni?
Im a bit confused, which post hoc will be the best for my research and hope you can help me!
Hello Christoph,
thank you so much for your valuable answer!
may I introduce a bit more in detail into the research purpose:
I want to figure out, whether the brand itself or the sourcing practices of the brand have greater impact on the consumers` perception of a company. The consumers perception is specified through 4 constructs (items), adapted from past literature. Therefore I developed four surveys and selected as the sample for the study a well-known retailer and its sourcing practices for the comparison.
Survey 1: Treatment ---> during the entire survey the brand name (of the well-known retailer) is given, as well as the actual sourcing chart of the brand (company)
Survey 2: Treatment ---> during the entire survey the brand name (of the well-known retailer) is given, but this time a self-faked sourcing chart is provided.
Survey 3: Treatment ---> during the entire survey an imaginary "Brand X" is given, as well as the actual sourcing chart (same as in survey 1) of the above mentioned sample brand (well-known retailer)
Survey 4: Treatment ----> during the entire survey an imaginary "Brand X" is given, as well as the self-faked sourcing chart (same as in survey 2).
All 4 surveys are measuring the same constructs, which are representing the consumers` perception of a company. The idea is, as already mentioned above, to investigate the gap between the groups, thus be able to highlight where a company should put its focus to generate an improved or more positive consumer perception.
The ANOVA shows that there is a statistical significance between the groups (in every case), what indicates that there is a difference due to the treatment.
Now i want to know where the differences are that i am able to show whether the brand or the sourcing practices have more impact on the perception. May you can give me some food for my thoughts, if its better to create two groups (eg. survey 1 and survey 3 - survey 2 and survey 4) and compare them, or to compare all to a reference group (eg. reference group = survey 1) to answer my research question.
I hope this gives a better insight into my work.
Thank you in advance!!!
Not sure if you are talking about continuous data or what, but if you want to compare means between two groups (at a time?), instead of hypothesis testing, where there can be confusion - see the link below - could you try a confidence interval about the difference in means to see if a reasonable confidence interval includes a difference in means of zero? If these are data that lend themselves to confidence intervals about the difference between two means, then such results would be far more practically interpretable than hypothesis testing.
Article Practical Interpretation of Hypothesis Tests - letter to the...
- Badges
- Science method
- Similar topics
- Mathematical Sciences
- Statistics
- Applied Statistics
- Survey Design