Re: coefficient in logistic regression
Assuming you have coded everything correctly, I would look at something called complete or quasi-complete separation. These conditions often lead to grossly inflated coefficients. Claudiu D. Tufis wrote: Hi, I have a multiple logistic regression. Among the predictors, I have 6 variables that represent the dummies for an interaction term (the seventh is the reference category and is not included in analysis). I have obtained for five of these variables extremely large coefficients: exp(b) ranges from 90,000 to 166,000. Could you please tell me if it is normal to have such values for exp(b)? Do you think it is something wrong? Thank you very much -- Tim Victor Policy Research, Evaluation, and Measurement Psychology in Education Division Graduate School of Education University of Pennsylvania = Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at http://jse.stat.ncsu.edu/ =
Re: coefficient in logistic regression
Something from Greenland and Rothman's book Modern Epidemiology (page 258) may apply here: there is a hallmark sysmptom of the bias that arises when stratification has exceeded the limits of the data: The exposure effect estimates begin to get further and further from the null as more variables are added to the stratification or regression model. [...] This inflation is sometimes mistakenly interpreted as evidence of confounding, but in our experience is more often bias due to applying large-sample methods to excessively sparse data. how many variables are in your model? maybe you're stratifying so much that you're ending up with near-empty cells. KH At 10:44 AM 1/26/2002 -0500, Timothy W. Victor wrote: Assuming you have coded everything correctly, I would look at something called complete or quasi-complete separation. These conditions often lead to grossly inflated coefficients. Claudiu D. Tufis wrote: Hi, I have a multiple logistic regression. Among the predictors, I have 6 variables that represent the dummies for an interaction term (the seventh is the reference category and is not included in analysis). I have obtained for five of these variables extremely large coefficients: exp(b) ranges from 90,000 to 166,000. Could you please tell me if it is normal to have such values for exp(b)? Do you think it is something wrong? Thank you very much -- Tim Victor Policy Research, Evaluation, and Measurement Psychology in Education Division Graduate School of Education University of Pennsylvania = Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at http://jse.stat.ncsu.edu/ = = Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at http://jse.stat.ncsu.edu/ =
coefficient in logistic regression
Hi, I have a multiple logistic regression. Among the predictors, I have 6 variables that represent the dummies for an interaction term (the seventh is the reference category and is not included in analysis). I have obtained for five of these variables extremely large coefficients: exp(b) ranges from 90,000 to 166,000. Could you please tell me if it is normal to have such values for exp(b)? Do you think it is something wrong? Thank you very much
Re: coefficient in logistic regression
On 25 Jan 2002 08:13:41 -0800, [EMAIL PROTECTED] (Claudiu D. Tufis) wrote: Hi, I have a multiple logistic regression. Among the predictors, I have 6 variables that represent the dummies for an interaction term (the seventh is the reference category and is not included in analysis). I have obtained for five of these variables extremely large coefficients: exp(b) ranges from 90,000 to 166,000. Could you please tell me if it is normal to have such values for exp(b)? Do you think it is something wrong? No, it is not normal. Yes, something is wrong, if that is really what you have. If you have more question, copy some output for us; send some set-up lines; and mention what program. Also, you posted many, many lines of HTML; please turn off that option if you can figure out how. -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html = Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at http://jse.stat.ncsu.edu/ =