How to analyze two different research data measurement scale? For example: I am in condition quantitative research, I found 2 variables. But the problem they have different indicators, the one maybe measure by likert scale (1, 2, 3, ...) and the other maybe measure by nominal scale (yes-no scale). So, in this case how to data analyze?
I suggest you to use NetSys software, very easy tool for combined analysis of two different data
The answer to your question hinges on what you want to know. Do you want to analyze the association between these two measures? Test the mean difference of your continuous variable between your nominal categories? Test the difference that increasing a unit on your continuous measure has on the likelihood of being in one or the other group of your nominal variable? Each of these involves a different approach.
May i add converting the outcomes of both varibales to percentage mean scores where you will present both in mean standerd deviation and analyze by t-test.
You can then reconvert them to qualitative measures(ex. 50 and above good perception or knowledge or attitude) and present them in categorical values.
Hi Reza, Julia, Jan, Philip and Mahmoud
Thanks for your answer, actually I just want to know the relation of the two variables such as Julia said. Base on questionnaire that I found from the past research I conclude that they have different different research data measurement scale such as my question. How I code it at SPSS if the measuring scale different, because I use SPSS to analyze it. Is it no matter?
It matters. What is the scale of your predictor (X) and what is the scale of your outcome (Y)? There are different ways to run a correlation in SPSS depending on your scales. (Note, I am assuming that you are in fact wanting to assess the association between two variables and not the other possibilities I listed?)
Julia, Yes I want to assess the association between two variables, But the fact: the variable (X) using Likert Scale, at the questionnaire the respondents will be given about 4 answer options, for example: very agree (scores=4), agree (scores=3), disagree (scores=2), and very disagree (scores=1). And the variable (Y) using nominal scale, maybe at the questionnaire only there are 2 answer options such as yes (scores=1) and no (scores=0). In this case, how to analyze the data?
You may consider Kruskal-Wallis. (Analyze > Non-parametric > Legacy dialog > K-independent samples. Put the Likert variables into Test variable list and put the nominal variable into Grouping variable).
- Badges
- Science topic
- Similar topics
- Computer Communications (Networks)
- Simulators