Multinominal logistic regression, Part 1: Introduction

This video introduces the method and when it should be used. It shows a simple example with one explanatory variable to illustrate how the method works and how the results can be interpreted using either odds ratios or predicted probabilities.
This video is part of NCRM Online Resource on Multinominal logistic regression by Dr Dr Heini Väisänen. To view the resource (which includes, slides, worksheet, data and reading list) visit www.ncrm.ac.uk/resources/online/
Please note: we may be unable to respond to individual questions on this video.
The National Centre for Research Methods (NCRM) delivers research methods training through short courses and free online resources.
- Visit the NCRM website: www.ncrm.ac.uk
- Browse our short courses: www.ncrm.ac.uk/training/ncrm-...
- Find online resources: www.ncrm.ac.uk/resources
Follow NCRM on social media:
- Twitter: / ncrmuk
- LinkedIn: / ncrmuk
- Facebook: / ncrmuk
- KZbin: / ncrmuk
Subscribe to the NCRM monthly newsletter: www.ncrm.ac.uk/news/subscribe

[@drisselghoufi6728]

Thank you🙏

[@tatianabarcenasbarreto6967]

Thanks for sharing this valuable knowledge with your clear and fantastic explanations.

[@Robin-zc2iw]

I'm doing a master's program in the US and my professor just explained this concept and I was so confused. Today's my test and this video makes my understanding of MN logistic regression so much better than it was. Thank you!

[@kaynkayn9870]

"Today's my test" - certified uni student moment

[@Sanjeevpudasaini]

Thank you so much for such a good explanation!

[@f.181]

Thank you very much for the excellent presentation. Very good video! I have a question. At 13:37: shouldn't it be "The odds of being *unemployed* rather than in employment are 42% lower for women than for men"?

[@asadyare3268]

You are right.

[@Willi_Zhang]

Thank you for this very helpful video!

[@samnangum2750]

Thank for sharing Dr

[@KayYesYouTuber]

Good explanation of multinomial logistic regression.

[@jakobforslin6301]

Thank you very much!

[@kaushikjayaram8556]

Thanks for sharing

[@JiyongKim-et9sw]

Thanks !

[@bismarkanloadey289]

hello, what if, instead of the dependent variable being more than 2, you have the explanatory variable rather to be more than 2. example; how sitting technique (upright, bent and curled) impacts the shape of the spinal cord. can you help with the impact model that'll be ideal for this analysis?

[@KyambaddeFrancis-ih8uk]

Thanks for the presentation, which values of x did you use

[@clausvind8010]

The sliste at arounnd 13:22 have the same text for both bullets: I believe the second bullet should read "The odds of being unemployed rather than in employment are 42% lower for women than for men"

[@clausvind8010]

slide not sliste

[@prempant6428]

What's the explanation for that equation on slide 13:26 ? The logit scale which is used first is ln(x/(1-x)) = y, if I am not wrong so x = e^y / (1 + e^y), you say that you've used the odd scale values but you used the logit scale values, during the calculation of the percentages ?

[@shashidharchavan3895]

@time line 13.23 the second interpretation should be unemployment rather than in employment.

[@dentoda]

Mlogit depvar indepvar, rrr gives OR output instead of coefficients

[@babaabba9348]

Mycket bra!

[@littlenuts779]

Why the numerator of pi3 is 1 as the reference category?