I learned more from this 11 minute video than my two years of graduate school. You are a phenomenal instructor, you provide relatable examples and easy to remember rules and guidelines for statistical analyses. THANK. YOU! 🙏🏾👏🏾
Thank you so much for this video! I have a quick question: is it possible to do a spearman's rank correlation test with an ordinal variable (1-7 scale) and a nominal variable, for example age that have been sectioned into groups? If not, is there a way to see if there are any correlations between these two variables? Thanks again!
Hi! My pleasure. Nominal variables should not be used with Spearman. If age has been chunked into 3 or more groups, it would be an ordinal variable as well and Spearman could be used. If age is just two groups, you would want an independent samples t-test or a point-biserial correlation.
This is very helpful. Thank you so much. Is there any specific marker to check the distribution? Or is it just the histogram? How do we write that in the paper? Is it just by looking at the histogram?
Thanks a lot, Madam, for your lovely explanation. My question related to the normality: Can we run a normality test (Shapiro or k-s test) to make a decision rather than usage figures, because for beginners like me, I can not judge on the normality from figure or you must decided on figures. Thanks in advance
My pleasure! Absolutely a normality test is appropriate. I run out of time to also teach this during my class (which is the primary audience for this video), but if I were doing it myself for research, I would rely on a K-S (Komolgorov-Smirnov) test myself.
@Virginia Wickline So, if the normality is violated, then the spearmen coefficient is a suitable alternative for pearson to make the correlation or not mandatory
@@fadhilfp6 If normality is violated, Spearman is more appropriate. For more, see: stats.stackexchange.com/questions/3730/pearsons-or-spearmans-correlation-with-non-normal-data
if there are a lot of cases clumped together, what kind of graph can we use to show correlation? I have a statistically significant correlation between household income and credit debt, however the scatter plot looks clumped together...?
Hi! What is the best way to plot a spearman correlation so that we clearly see the information that it is giving? I think it's not useful to plot a regression line as it depends on outliers, while the spearman correlation doesn't.
Hi! I sometimes still plot a regression line or a Loess curve to see where the data change directions. I am open to options/suggestions from other viewers!