One of the best explanations I've seen and yet is significance of any relevance, it is not mentioned. Can we draw the same conclusions on the odds ratio from an independent variable that is not significant or should it simply be ignored?
Relevance of the significance depends somewhat on the situation in which you are trying to apply the prediction. The odds ratios can be used to determine the relevance of the prediction. OR above 2 are usually considered clinically/practically relevant but they should be interpreted with the CIs.
Hi, I'd like to know if running a "binary logistic regression" is the same as a "multivariate logistic regression"? I am looking to assess the impact of different characteristics of a population on an outcome (age, gender, smoking status and others, on whether they develop kidney failure after an operation). The dependent factor is a categorical variable (kidney failure = yes / no).
Those terms mean the same analysis. "binary logistic regression" refers to the outcome you want to predict and "multivariate logistic regression" refers to the multiple possible predictors you can use.
thanks a lot, your video is so helpful! and it would be superb if you could make tutorial on what are the functions in every "options" tab mean, which to tick, when, and why? but thanks again!
My supervisor adviced to make dummies for each variable (ordinal variables). What is the purpose of a dummy? And what does is do with the regression? And is it true that it isn't worth to put a variable in the regression if the correlation is weak? What are you using as cut-off? Thank you very much!
Hi, Really great videos on RUclips! Can you please help me with 2 questions (both to do with dummy variables) - 1) If I have 2 sets of nominal variables (eg ethnicity [white, Indian, African] and smoking status [current smoker, ex-smoker, non-smoker]) and I turn each set into 2 dummy variables (+1 baseline reference variable for each set) for entry into a multiple linear regression in SPSS. How does SPSS know which baseline reference variable to use for which set? At least in logistic regression you have to option to label each variable as categorical (so there seems to be some way of grouping them) - this does not exist for linear regression. For my example given above, does it mean that the multiple regression would produce 2 constants? 2) If I recode ethnicity (White, Indian, African) into 3 dummy variables, why can't these be treated as 3 separate binary predictor variables and entered into the model? Ie why can't I enter 'White / Non-White', 'Indian / Non-Indian' and 'African / Non-African' : and thus why can't I interpret the resultant coefficients as White compared to Non-white, Indian compared to Non-Indian and African compared to Non-African?
Hi, thanks for the video. I am trying to get the right model to predict firm's default with means that my variables vary in time . Can you please tell me if there is a way to use binary logit model for this type of analysis. I found some comments that it may be possible with Logit model which includes eash firm year as a SEPARATE OBSERVATION. But I a little bit confuded how to do this. Would appreaciate if you can explane it.
If your outcome is default/ not default, then logistic regression would be appropriate. If a understand your question, if you are trying to predict default for each year, then you would need a regression equation for each year of data.
Hi thank you for this it is very clear. Just wondering how to calculate odds ratio using a column that uses three different variables in terms of then predicting an outcome i.e one column that records if a pt is low,medium,high risk and then incorporating this to calculate odds of failing?
2) If I recode ethnicity (White, Indian, African) into 3 dummy variables, why can't these be treated as 3 separate binary predictor variables and entered into the model? Ie why can't I enter 'White / Non-White', 'Indian / Non-Indian' and 'African / Non-African' : and thus why can't I interpret the resultant coefficients as White compared to Non-white, Indian compared to Non-Indian and African compared to Non-African?
Hey, great video! what about individual significances? Not all of the variables seem to be significant indivually at the end of the video? leave them out?
Thank you very much!! I turned all ordinal variables into dummies. Thing I noticed is that some of the dummies aren't sign in the coefficients table. As example I made education level (low, mid, high) a dummy (mid is the baseline). The dummy education_low has a sign of 0.255. The baseline and dummy education_high are
How can i get averaged marginal effects instead of odds ratios in a binary logistic regression with SPSS? i need them to compare two different points in time like 2002 and 2010 and odds ratios are not adequate for this.
Hi, thanks for your video. What if I had categorical variables, for example people with long hair, people with short hair, and people with curly hair, to try and predict which of these are likely to pass the exam? How would input this data in spss? Would you give them a code "0", "1" and "2" in another column?
Great video! I noticed in the correlation matrix that there seems to be a fairly high positive correlation between overall GPA and Major GPA (> .8). The results of the regression are a little bit confusing, because the Beta weights (and exponentiated Betas) of these two IVs are oppositely related to passing the exam. Is multicollinearity an issue for logistic regression?
hey, great video, can you tell me how to interpret the Exp(B) when my variable is in the form of a ratio between 0 and 1? what is the unit of measure in that case?
There likely wouldn't be a unit of measure if your variable is dichotomous. That is one of the limitations when you have a predictor variable that is in the nominal scale because you cannot interpret B the same way you would for a continuous variable.
noussa taec I would suggest the following resources:SPSS Survival Manual by J PallantAdvanced and Multivariate Statistical Methods by Mertler & Vannatta
Finally, I finished my statistical analysis. You helped me a lot with problems. I'm very grateful for your help.
fabulous video, highly recommended for statisticians, students, etc. thank you very much
Thank you very much. Your video is superbly clear. I have learnt a lot. God bless.Thks again.
A really informative and well presented video. I learnt a lot about using logistic regression to analyse my own data. Excellent!
It was really a complicated topic which was made too easy to understand.
Thank you so much!
Many thanks man, it is about a week don't know what to do with my data. great video.
Thank you so much, this video made the topic much more understandable.
You're very welcome!
Good presentation, the lecture cleared by concept how to apply Logit
Excellent video. Thanks for posting!
You have used the correlation for the nominal dichotomous variable? Is pearson r works for this data?
Vincent Bongolan Technically it does not. Multiple regression traditionally is reserved for numeric predictors.
One of the best explanations I've seen and yet is significance of any relevance, it is not mentioned. Can we draw the same conclusions on the odds ratio from an independent variable that is not significant or should it simply be ignored?
Relevance of the significance depends somewhat on the situation in which you are trying to apply the prediction. The odds ratios can be used to determine the relevance of the prediction. OR above 2 are usually considered clinically/practically relevant but they should be interpreted with the CIs.
Your Videos do miracle! Thankssssssssss a lot.
Hi, I'd like to know if running a "binary logistic regression" is the same as a "multivariate logistic regression"? I am looking to assess the impact of different characteristics of a population on an outcome (age, gender, smoking status and others, on whether they develop kidney failure after an operation). The dependent factor is a categorical variable (kidney failure = yes / no).
Those terms mean the same analysis. "binary logistic regression" refers to the outcome you want to predict and "multivariate logistic regression" refers to the multiple possible predictors you can use.
thank you so much for making my study easy....
thanks a lot, your video is so helpful! and it would be superb if you could make tutorial on what are the functions in every "options" tab mean, which to tick, when, and why?
but thanks again!
Thank you for this video, was very helpful.
Best video
Thank you!
thank you very infornative video
My supervisor adviced to make dummies for each variable (ordinal variables). What is the purpose of a dummy? And what does is do with the regression?
And is it true that it isn't worth to put a variable in the regression if the correlation is weak? What are you using as cut-off?
Thank you very much!
Hi,
Really great videos on RUclips!
Can you please help me with 2 questions (both to do with dummy variables) -
1) If I have 2 sets of nominal variables (eg ethnicity [white, Indian, African] and smoking status [current smoker, ex-smoker, non-smoker]) and I turn each set into 2 dummy variables (+1 baseline reference variable for each set) for entry into a multiple linear regression in SPSS. How does SPSS know which baseline reference variable to use for which set?
At least in logistic regression you have to option to label each variable as categorical (so there seems to be some way of grouping them) - this does not exist for linear regression.
For my example given above, does it mean that the multiple regression would produce 2 constants?
2) If I recode ethnicity (White, Indian, African) into 3 dummy variables, why can't these be treated as 3 separate binary predictor variables and entered into the model? Ie why can't I enter 'White / Non-White', 'Indian / Non-Indian' and 'African / Non-African' : and thus why can't I interpret the resultant coefficients as White compared to Non-white, Indian compared to Non-Indian and African compared to Non-African?
Hi, thanks for the video. I am trying to get the right model to predict firm's default with means that my variables vary in time . Can you please tell me if there is a way to use binary logit model for this type of analysis. I found some comments that it may be possible with Logit model which includes eash firm year as a SEPARATE OBSERVATION. But I a little bit confuded how to do this. Would appreaciate if you can explane it.
If your outcome is default/ not default, then logistic regression would be appropriate. If a understand your question, if you are trying to predict default for each year, then you would need a regression equation for each year of data.
Hi thank you for this it is very clear. Just wondering how to calculate odds ratio using a column that uses three different variables in terms of then predicting an outcome i.e one column that records if a pt is low,medium,high risk and then incorporating this to calculate odds of failing?
2) If I recode ethnicity (White, Indian, African) into 3 dummy variables, why can't these be treated as 3 separate binary predictor variables and entered into the model? Ie why can't I enter 'White / Non-White', 'Indian / Non-Indian' and 'African / Non-African' : and thus why can't I interpret the resultant coefficients as White compared to Non-white, Indian compared to Non-Indian and African compared to Non-African?
Hey, great video!
what about individual significances? Not all of the variables seem to be significant indivually at the end of the video? leave them out?
Yes, that would either strengthen the prediction model or least make it more efficient.
Thank you very much!! I turned all ordinal variables into dummies. Thing I noticed is that some of the dummies aren't sign in the coefficients table. As example I made education level (low, mid, high) a dummy (mid is the baseline). The dummy education_low has a sign of 0.255. The baseline and dummy education_high are
How can i get averaged marginal effects instead of odds ratios in a binary logistic regression with SPSS? i need them to compare two different points in time like 2002 and 2010 and odds ratios are not adequate for this.
Hi, thanks for your video. What if I had categorical variables, for example people with long hair, people with short hair, and people with curly hair, to try and predict which of these are likely to pass the exam? How would input this data in spss? Would you give them a code "0", "1" and "2" in another column?
Yes, that is correct.
Great video! I noticed in the correlation matrix that there seems to be a fairly high positive correlation between overall GPA and Major GPA (> .8). The results of the regression are a little bit confusing, because the Beta weights (and exponentiated Betas) of these two IVs are oppositely related to passing the exam. Is multicollinearity an issue for logistic regression?
+Jesse Grabman Yes, mulicollinearity can be an issue.
Which one is the adjusted odds ratio?
hey, great video, can you tell me how to interpret the Exp(B) when my variable is in the form of a ratio between 0 and 1? what is the unit of measure in that case?
There likely wouldn't be a unit of measure if your variable is dichotomous. That is one of the limitations when you have a predictor variable that is in the nominal scale because you cannot interpret B the same way you would for a continuous variable.
the outcome in your video is pass or fail, if mine is poor, average and good, can i use logistic regression?
Yes, you can.
great video. can you please upload the data file?
Hi could you share a data set that you used in your video example of logistic regression.
Shahriar Shamil Uulu I would be happy to. Please send me a private message with your email address.
hi, plz I want to know how to interpret our result?? if there is any WORD Document plz give to me!!!
noussa taec I would suggest the following resources:SPSS Survival Manual by J PallantAdvanced and Multivariate Statistical Methods by Mertler & Vannatta