I'm currently analysing data to write up my dissertation however I've found myself stuck. I have a data set based on a questionnaire I conducted investigating people's belief in epiphany as well as their belief in religion, magic ideation and existential thinking. The problem is that I'm unsure on what statistical analysis to use to find out the following;
1). Does an increased religious belief, existential thinking and magic ideation score signify an increased epiphany belief score?
I've been notified that multiple regression would be suited for this however I'm unsure and was also told about using ANOVA.
I hope the detail I've gone into is enough
Before we can determine the right analysis for your data, we need to know what research design you used and how the data were collected. Did you pre-define a population (people you're interested in)? And if so, what sampling method was used? If it's convenience sampling, a lot of statistical tools (based on probability sampling) have to be ruled out.
Hey @Antoon, the data had no specific demographic; it was a volunteer sample for anyone who felt like applying. It was collected using an online questionnaire on Qualtrics. The participants were given multiple choice questions on their views (testing for beliefs in epiphany etc. as aforementioned).
@Ingrid I've used a regression analysis for 3 predictor variables (experiential thinking, magical ideation and religion) against 1 criterion variable (perceived authenticity of epiphany) and reported this. The multiple regression I assume is all that's necessary
Though it's done a lot, you cannot use regression analysis when working with a non-probability sample.
Public online surveys are also tricky because of self-selection and the volunteer bias.
Leonardo Jordan I would honestly start over (too many issues already) and re-angle the whole study towards a mixed-methods research design, with the dataset you have as indicative of some people's viewpoints but not allowing a thick description.
By the way, how many people filled in the questionnaire? Note that I do not use "sample size" here because it's not a sample drawn from a pre-defined population.
Also, what's your context? You mention "dissertation" so I assume there is a supervisor or supervisory panel, right? If you're meant to extract a publishable article from your work, that will also raise the bar considerably.
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Writing