You can find the spreadsheets for this video and some additional materials here: drive.google.com/drive/folders/1sP40IW0p0w5IETCgo464uhDFfdyR6rh7 Please consider supporting NEDL on Patreon: www.patreon.com/NEDLeducation
I love how you so eloquently not only the essence of a chi-squared test but also the foundation of statistical concepts like degrees of freedom. You’ve done a far better job than the university courses and textbooks I’ve come across. Well done!
You legit just saved my life! This has been so beyond helpful, easy to understand, and engaging. Working on my thesis now doesn't seem so bleak anyone. Thank you thank you thank you :D
Hi Zam, and glad the video helped! As for your question, I just wanted to show the process the easiest way possible just using basic functions and not relying on advanced tools. So if you want to compile the contingency tables using pivots, it is absolutely fine.
Excellent video as always! One of the requirements for using the t-test is that the sample be independent. (I think that is correct, and I stated it properly.) Should I be applying Chi-squared test to evaluate independence? Is this similar to testing for autocorrelation? Thanks!
Hi and many thanks for the feedback! As for your question, it needn't be that the samples are independent for a t-test. For example, a paired t-test can be used to evaluate the significance of differences on a matched sample (whether the return of one portfolio is greater than the return of another portfolio on a day-by-day basis, for example). So t-test is perfectly applicable even when samples are not independent, you just have to adjust for dependence when calculating the variance. Chi-squared test is mainly used to test independence for categorical variables, however it can also be applied to quantitative data if you sort it into quartiles (for example, are returns below zero and above zero of two portfolios are associated). Hope it helps!
Thank you for the presentation. I have both nominal and ordinal independent variables (4) with (2) ordinal dependent variables . Please which one should i choose between Spearman correlation and Chi-square test to analyze the association or relationship between the variables ? Thanks
Hi, and glad you enjoyed the video! It seems that Chi-squared test is more appropriate in your case. I have got a video on Spearman correlation as well, however, so do check it out if you are interested: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-chgijGUVN7g.html
@@NEDLeducation Thank you for your reply. I will check the video about Spearman correlation. Is it possible to have the two analysis in my work? i mean doing chi-square to see the relationship between the 2 nominal independent variables and the dependent variables and use the spearman correlation to see the relationship between the 2 ordinal independent variable and the dependent variable.
Hi! What if I want to determine the relationship between a categorical variable and a numerical variable? I was tasked to determine the relationship between proportion of times I correctly judged that the test probe was on the list (dependent variable) and the type of cue at study / test cue (independent variable). The data I have are literally just on the proportion of times I guessed correctly and the categories (Weak/Weak, Weak/Strong, Strong/Weak, Strong/Strong, and Lure). How will I compute for the correlation?
Hi Hannah and many thanks for the question! There is a number of ways to approach this. If you want to apply the Chi-squared test, you can separate your numerical variable into quintiles and test for association between quintiles and categorical variables. However, for it to work the best, you need a relatively high number of observations. Alternatively, you can use the one-way ANOVA to compare variance of the numerical variable between groups and within groups. If the former is sufficiently high, the effect of the categorical variable can be considered significant. Lastly, you can code a set of dummy variables based on your categories and run a simple regression. If the F-stat of the regression is significant, you then can infer that the categorical variable is meaningful in terms of explaining variation. Hope it helps!
