Hi ALL,
I want to specify the associations of confidence with demographic variables, but I think depending on if confidence is measured using a single Likert-type item or a Likert scale, different statistical tests are appropriate.
I think if I use Likert scale, Pearson correlation would be the option, because Likert scale can be analyzed as a numeric variable (ie. mean is meaningful)
But if I go with A single Likert-type item, which test should he appropriate?
As we know A single Likert-type item is ordinal and can be presented as either a median, mode or as frequency of responses, but not as a mean.
Thanks,
Bahar
The statistical test that you choose largely depends on the following factors:
1. The hypothesis.
2. The conceptual framework of the study (dependent and independent variables).
3. Type of variables (categorical, scale, or ordinal variables).
4. Parametric vs. Non-parametric tests.
I have a YouTube video titled- Which statistical test should I perform? that explains this in detail- https://www.youtube.com/watch?v=F4AKZ9-u_ic
Thank you for sharing this link, and yes, but the aim is just to get the association for the Likert item.
my question focuses on Likert item, not the Likert scale
I was just asking to be sure, is that correct to say Spearman will be used for the Likert item, and Pearson will be used for the Likert scale?
@Bahar, yes, if you are comparing two ordinal variables (likert variable) then Spearman's correlation is ideal whereas Pearson's correlation is appropriate when you are comparing two variables measured as scales (continuous variables).