Let me first start by saying I'm dyslexic, two universities were unable to teach me research methods and statistics in a way that worked with my dyslexic limitations, and I've continuingly failed to find a way I could teach myself. So bare with me and treat me more like you're explaining it to a child.
I want to conduct a survey/questionnaire to assess the needs off a subculture to find out what areas they could do with additional support. My first draft of questions is based on multiple-choice questions, Linkert scale questions, GAD-7, PHQ-9, and some qualitative open-ended questions.
What would be the best way to statistically analyse the data?
I read somewhere that my Likert scale questions can be analysed by doing T-Test's or the Mann-Whitney test, is that right? But what about the multiple-choice questions, and questions where you can pick multiple answers?
Also, would it be better to split the survey/questionnaire into topic-based sections and give each section their own set of IVs and DVs? For example, a section on discrimination and a section on parenting support
As long as your questionnaires evaluate a particular variable, you can divide the two sections mentioned above into subscales, such as discrimination and parental support. These instruments should be standardized for their application (evidence of reliability and validity), and it is recommended that the likert questions have the same number of options.
How large is N ?
for small sizes Excel is still Ok
but you can also use tools on the internet like survey monkey or google docs
kind regards,
jan
You can divide the questionnaire and find out the cronbach alpha for subsection for reliability and consistency. For ordinal data and if your sample is not normally distributed. Then run non parametric Kruskal Wallis Test.
The multiple-choice question in statistic is a matter of coding. Treat each statement from the question as a separate question and write using the 0-1 method, 0 no circling, 1 singing. Then sum up the results. You can treat individual statements as separate questions or after adding up as a whole, i.e. one question from the questionnaire (this is a multiple choice, one your question)
regards
Slavek
- Badges
- Science topic
- Similar topics
- Higher Education
- Universities