Sorry, but I'm confused about when to categorize within and between groups variables. Update: I think that I was confused by the "within" and "between" terms. What I really want to do is confirm if the scenario described below constitutes a repeated measures design or not.

The scenario:
I conduct a survey where I have participants rate their subjective taste of three different juices: apple, mango, and kiwi.
Information on gender is also collected.
For the data analysis I want to:
1) Check for significant differences in overall mean juice ratings.
2) Check for significant differences in individual mean juice ratings for a subset of the population (e.g. just men, just women, etc.)

So in this instance, would 1) be:
A repeated measures experiment with:
Dependent variable = juice rating
Within subjects factor = category of juice
Between subjects factor = gender

...And 2) would be:
A between subjects experiment with:
Dependent variables = apple juice rating, mango juice rating, kiwi juice rating
Within subjects factor = none
Between subjects factor = gender


I have seen the terms between groups and within groups used in the context of one-way ANOVA tests. For example, if you were interested in determining if juice rating was independent of juice flavor then you can calculate an f-statistic and perform a hypothesis test for independence. In this test you are trying to see if the mean juice rating is independent from the juice flavor category by studying the variance between groups and within groups (i.e. levels). You could also study how juice rating is affected by gender where you look at how mean juice varies across the genders male and female (i.e. between groups) and what the variance of juice rating (i.e. within group) is for each gender.

Depending on what relationships you are interested in investigating can you establish whether it is a purely within group or between group experiment.

  • 1
    $\begingroup$ You are right that because each person makes an evaluation for each juice flavor that this may have an effect on the experiment you carry out. I will think about your question some more and get back to you with a better answer if I can. As for your option “A”, I don’t see how you can have Independent variable = juice rating since in all likelihood juice rating will be affected by either juice flavor, gender, or age. Also, I am having a hard time understanding the terms “within subject” and “between subjects”. But this may due to my unfamiliarity with the field that the questions comes from. $\endgroup$ – ecampana Dec 21 '17 at 8:50
  • 1
    $\begingroup$ In addition, from my understanding juice flavor and gender would each be consider category variables/features. Category features have levels, which can also be referred to as groups. Juice flavor therefore has levels/groups of apple, mango, kiwi, etc... and gender has levels male and female. On the other hand juice rating and age are numerical features. So if you are studying juice rating vs juice flavor you can calculate the “within group” variance and “between group” variance. In your case it is a bit more difficult because each subject evaluates each juice flavor. $\endgroup$ – ecampana Dec 21 '17 at 9:00
  • 1
    $\begingroup$ Lastly, I am not sure if a subject will necessarily have a biased juice rating depending on the juice flavor, or if it can be assume that he will remain objective as possible as he evaluates each flavor. If this is the case then maybe it may simplify the experiment. $\endgroup$ – ecampana Dec 21 '17 at 9:04
  • 1
    $\begingroup$ You may find these links useful en.m.wikipedia.org/wiki/Categorical_variable and en.m.wikipedia.org/wiki/F-test $\endgroup$ – ecampana Dec 21 '17 at 9:10
  • 1
    $\begingroup$ If I understand you correctly when you say “within subject” variable you are referring to juice flavor since each person will evaluate each flavor and that is what motivates you to say “within” because they all come from the same source so to speak. Is this correct? Either way I don’t think this is what is meant by “within group”. I would say “between groups”. To clarify, I don’t use the term juice category since I reserve “category” to mean a type of variable that has different levels which are the flavors. $\endgroup$ – ecampana Dec 21 '17 at 9:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.