1

I'm writing my master thesis but I have some problems with the analysis of the data collected using an online survey. The aim of the research is to assess if there are differences between two groups of travellers: students and workers.

In the survey there are three types of questions:

  1. Multiple choice with the possibility to select only one option
  2. Multiple choice with the possibility to selecto more than one option
  3. Likert scale

My doubts are the following:

  1. I think I have to codify each question, but I can I do this by using Excel or SPSS?
  2. Which stat test should I use to assess the differences between the two groups?
Denisa
  • 25
  • 4

1 Answers1

0

The multiple choice questions where only 1 option can be chosen can be registered in 1 nominal variable. In SPSS it's usually better to use numbers to identify each category.

For the questions where people may choose more than 1 option, my advice is to split these questions, creating several dichotomous nominal variables in the database. Dichotomous means that each variable have two options ($1$ = yes, or selected; $0$ = no, or not selected).

Regarding the Likert variables, you may register directly the value in the variable of the database.

Regarding the tests, you may do chi-square tests to evaluate the dependency between the group and a nominal variable.

With respect to the Likert scale you may do either a Mann-Whitney test or a independent samples t-Student's test, depending on the distribution of the variable.

Ertxiem - reinstate Monica
  • 1,403
  • 4
  • 8
  • 17