Statistics Help @ Talk Stats Forum - Regression Analysis
http://www.talkstats.com/
Linear regression, linear models, nonlinear regressionenTue, 17 Oct 2017 16:14:25 GMTvBulletin60http://www.talkstats.com/images/misc/rss.pngStatistics Help @ Talk Stats Forum - Regression Analysis
http://www.talkstats.com/
Ordinal logistic regression - How to get a single scoring function?
http://www.talkstats.com/showthread.php/70398-Ordinal-logistic-regression-How-to-get-a-single-scoring-function?goto=newpost
Tue, 17 Oct 2017 09:33:15 GMTHi all,

I'm running ordinal logistic regressions, with adjacent-category logits.
The dependent variable has four ordered levels.
So I have 3 logit equations, i.e. 3 linear predictors.
I'm wondering whether it is possible to get a single scoring function by aggregation of the 3 linear predictors. If yes, how to proceed to this aggregation in a rigorous way ?
Just to say, my eventual aim is to compute the Volume Under the ROC Surface (VUS), by using this single scoring function, so as to evaluate and compare my models' performance.

Thank you in advance,
Best regards.
]]>Regression Analysisatrollehttp://www.talkstats.com/showthread.php/70398-Ordinal-logistic-regression-How-to-get-a-single-scoring-functionConverting a continuous variable into a categorical variable (low, medium, high)
http://www.talkstats.com/showthread.php/70395-Converting-a-continuous-variable-into-a-categorical-variable-(low-medium-high)?goto=newpost
Tue, 17 Oct 2017 05:08:11 GMTI have a continuous dependent variable (scores on an exam) and a number of predictor variables that I'd like to use in a regression analysis, such as gender, age, and income. I'd like to convert age and income into categorical variables: low, medium, and high. I'd like to do this using percentiles, but am not sure if I should use tertiles (lower 33% = low, middle 33% = medium, upper 33% = high) or if I should divide the data into lower 25% (low), middle 50% (medium), and upper 25% (high).

Thoughts? Thank you!
]]>Regression Analysistess2http://www.talkstats.com/showthread.php/70395-Converting-a-continuous-variable-into-a-categorical-variable-(low-medium-high)Controlling for time
http://www.talkstats.com/showthread.php/70389-Controlling-for-time?goto=newpost
Mon, 16 Oct 2017 13:15:17 GMTHello,
Thank-you for taking the time to read this post. I have managed to get myself very confused about what is (hopefully) quite a simple...Hello,

Thank-you for taking the time to read this post. I have managed to get myself very confused about what is (hopefully) quite a simple problem.

I am aiming to look at the engagement on Twitter during election debates. I have collected tweets and have coded them in regards to their policy areas. I want to see whether Twitter users have a good understanding of devolved and non-devolved policy. Therefore, my DV is whether or not a particular policy area is mentioned (e.g. health mentioned/ not mentioned), and my IV is whether it is a general or local debate.

However, in order to get an accurate reading of the Twitter users understandings I need to control for the length of time the politicians spent discussing each policy area as this could bias the results (i.e. its not that they understand its a devolved area- they just are tweeting about it because the are talking about it on telly).

I have the amount of minutes each policy area was discussed, but I am unsure where to put this in the model. I have tried to create a 'time' variable which gives each tweet the time the politicians spent discussing it during the debates- but this is throwing up strange results in a logistic regression.

Any ideas of how to do this would be really beneficial! Thank-you very much!
]]>Regression AnalysisCathyKhttp://www.talkstats.com/showthread.php/70389-Controlling-for-time