I have data on two countries where I have evaluated 3 types of brands in each. Same questionnaire based on Likert scale. There are 10 items that were evaluated, that can be grouped into 2 independent and 1 dependent variables.
If possible I want to find a better tool than t-test and ANOVA.
Hello Zamira. I didn't think the t test was used for this, I would have used chi squared but I'm rusty on statistics at present. You could check Armitage and Berry statistical methods in medical research if it is in your library. I have it, but the binding has perished, so all the pages are falling out. I know where it is if you don't have it. I'm not 100% sure about my answer, wait to see what others say.
Hi, Zamira Ataniyazova . Your dependent variable based on Likert items ? When you say "scale" do you mean that you are combining several individual Likert items into a scale ? And is that you dependent variable ?
Sal Mangiafico yes, to all questions. and a similar approach is done with independent variables. In short, I have 10 facets that I have used to measure potential brand success, they can be devided (grouped) into emotional, rational and action-oriented. Rational and emotional are my independent (attitudes), Action-oriented is dependent variable. (behavior)
So, it sounds like you have one dependent variable and two independent variables, and all of them can be treated as quasi-continuous. The obvious approach would be multiple linear regression. Which is a form of general linear model, as is anova. How appropriate this is may depend on the (conditional) distribution of your measured variables, and other considerations. There are other approaches, like quantile regression, that may be better suited to your situation.
Is the difference between countries important to you? The problem is that you will have a sample of 2 and while you may be able conclude that there is a difference you will not be able to put into the model any country variables that may be able to account for that difference.
Analysing and Designing Extensive Studies in Which Place Matters: A Comment on Neighbourhood Incivilities and the Study of Crime in Place
Kelvyn Jones
Area, Vol. 26, No. 1 (Mar., 1994), pp. 79-81
https://www.jstor.org/stable/20003374
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire