I am doing a study on how programmers with different levels of expertise choose class names. A description was provided to the participants together with a list of possible class names for the described class (An example is attached). I need to statistically show that novices select the same class name as experts or otherwise. Which statistics can I use for this?
Your research queation sounds intriguing. Assuming, that your depwndent variable is on the nominal svale level and not ordinal, I guess a two-sample chi-square test should do the trick. One small anotation concerning your plot: Avoid line graphs for categorical values as they suggest that the categories are connected somehow and that there are values between the categories.
Sven Greving Thanks a lot for your answer. I have updated my graphs to scatter plots. Also, I looked at two-sample chi-square test. From the example given here: https://bookdown.org/ndphillips/YaRrr/chi-square-chsq-test.html , what I understood is, we can tell whether the answers selected by different levels of expertise differ from each other or not using the test. Is that correct? If so, the chi-square value tells us how different the answers of different expertise from each other? And I will have to calculate chi square value for each pair of expertise levels, is that right?
It looks like you already got the basics, I however am a bit confused about the explanations in your reference. I think the example is a bit misleading. The chisquare test basically tells you whether the proportions of categorical outcomes in a sample is unlikely enough to be considered different from (a) the assumed distribution in the population or (b) from another sample.
Check out this reference for another explanation using R code.
Concerning your plot, I think bars will be more informative
- Badges
- Science topic
- Similar topics
- Names