Logistic Regression Basics Explained: Probabilities, Odds, Odds-Ratios and Log-Odds-Ratios (4K) 

Thanks a lot for the guidance! The most understandable explanations of statistics.

Thanks. Your videos and perspective are unique.

Well explained, thank you!

Sweet problem while watching your videow is: I try to link it with all other videos you've uploaded and I get lost 😆 Solution: I again watch that specific video of yours where I think I got lost and get back on track. the best thing about your videos is every second is informative. Thank you very much.

Thanks. I watched every video.

Again, great video!!

Many thanks. Could you please also do something about probit regression (and as compared to logit regression), as well as ordinal probit regression? Many thanks for your effort!

How I understood what odds are, odds ratios. Using your example: the survival rate of females on a ship is 70 to 30, males 20 to 80. If we convert it into probability, then for females it is 0.8, and for males 0.2. How to calculate the odds based on this probability: chance (to survive) = probability (of survival) / (1 - probability (of survival)) = 0.7 / (1-0.7) = 2.33. It turns out that females have a chance of survival that is 2.33 times greater than the chance of dying. According to this formula, men have a chance to survive = 0.25, that is, the chance to survive is 4 times less than the chance to die. Or if you can interpret the odds this way - for every 233 surviving females, there are 100 dead. For every 25 surviving males, there are 100 dead. Odds ratio, as I understand it, calculates the odds ratio for two events, let’s assume event A - female survival rate is 70% and event B - male survival rate is 20%. Formula Odds ratio = Odds females (2.33) / Odds males (0.25) = 9.32. OR = 9.32 means that the chances of a female surviving a given event are 9.32 times greater than the chances of a male surviving the same event.

Excellent video

Thanks a lot

Prof. Please increase the frequency of videos. All forms of Probit Model, Impact evaluation model like Switching regression, DID, and so on...