Hi there I have a very important question for my statistical analysis which needs to be finished in a few days.

First of all I have to say that I have not much knowledge of statistics and therefore have problems to understand most of the answers for similar questions.

My question is:

I have conducted a survey with nine questions grouped into 3 variables (3 questions for each variable). Now I would like to have the score my variable which results from the 3 questions in each variable. So now I'm wondering how can I get my scores for the single variables? Can I calculate the mean of the scores from every respondent and then use the mean? Would that be okay?

I would be happy if I got a quick (and easy :D) answer. Thank you!!

  • $\begingroup$ The quick and easy answer is to add the values. I would probably do that if I were you. It is not perfect but it provides very meaningful results most of the time and a majority of researchers do it this way. $\endgroup$ – Behacad Sep 9 '13 at 4:18

This isn't a question where there is universal agreement. Some people say you cannot add Likert items together, because adding them implies that the individual items are interval scaled rather than ordinal scaled. E.g., suppose you just had two questions, each with 5 points from "strongly disagree" to "strongly agree". Suppose one person answered "Agree" to both and another answered "Neutral" and "Strongly Agree".

Now, if you coded the scale 1 2 3 4 5, the two would both add to the same sum. But are they the same? Technically, an ordinal scale could also be coded 1 4 5 100 101 102, then your two people would get different totals.

Lots of people ignore this and just assume that the adding is OK.

And yet other people do factor analysis, where there are some methods for dealing with ordinal data, but this is complex and also somewhat controversial.

  • $\begingroup$ Thanks Peter. You explained it very good and even for me it was understandable. But I still have that universal problem: "To mean or not to mean". I understand the the reasoning why taking the mean is not the best idea, however, factor analysis (which is controversial as well) is way too complicated for me and I really lack time. $\endgroup$ – Fiona Sep 8 '13 at 20:47
  • $\begingroup$ If it's for a dissertation, ask your mentor. If it's for a paper, try taking the mean and see what reviewers say. $\endgroup$ – Peter Flom - Reinstate Monica Sep 8 '13 at 22:25

The question's not whether you can do it (you can), but what the answer means.

If your items are "I sometimes enjoy a salad for lunch", & "Meat is murder", what does the average score measure? Propensity to vegetarianism? Probably not: a score of 4 for the former & 2 for the latter will be a much weaker predictor than vice versa.

So the difficulty is is in convincing people (including yourself) that the average means what you say it does. If you're not using formal scale construction techniques as mentioned by @Peter, then the least you can do is examine the plausibility of the average as a measure of another unmeasured variable: consider whether each item is similar in terms of relevance, direction, & weight; & whether the ranking of response combinations by average looks right.


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