Answers for "Comparing regression lines assuming equal intercepts"
https://developer.ibm.com/answers/questions/478221/comparing-regression-lines-assuming-equal-intercep.html
The latest answers for the question "Comparing regression lines assuming equal intercepts"Answer by jkpeck
https://developer.ibm.com/answers/answers/478651/view.html
@Karim.sarhane
I'm not sure what you are trying to do, but the R2 for the regression line tells you about the fit, and in a plot you can see whether the residuals - deviations from the line - are consistent with regression assumptions. If not, you might want to log your data or perfom some other transformation that linearizes the data.Sat, 03 Nov 2018 21:16:29 GMTjkpeckAnswer by Karim.sarhane@utoledo.edu
https://developer.ibm.com/answers/answers/478331/view.html
@jkpeck
Thanks very much again Dr. Peck for your answer.
Quick question. When I convert my data to a liner plot, how can I evaluate how much that linear plot is actually representing the data. The decay over time of the intensity of some of my compounds is somewhat linear but for others there is considerable scatter. How much scatter can I tolerate if I want to change the decay into a linear relationship?Thu, 01 Nov 2018 10:12:24 GMTKarim.sarhane@utoledo.eduAnswer by jkpeck
https://developer.ibm.com/answers/answers/478284/view.html
@karim.sarhane
While you could rig this up with the REGRESSION procedure, its better here to use UNIVARIATE.
Analyze > General Linear Model. Select y as dependent, x as covariate, and your group variable, say g, as a fixed factor. In the Model subdialog, put x, g, and the xd-g interaction as the model. Choose deviation as the contrast. In Options, check Parameter Estimates. That still leaves the intercept question, but this should get you started.Wed, 31 Oct 2018 23:34:08 GMTjkpeckAnswer by Karim.sarhane@utoledo.edu
https://developer.ibm.com/answers/answers/478262/view.html
Thanks very much Dr. Peck for your time.
I have different compounds (6 of them) that fluoresce. I am studying the rate of decay of that fluorescence over time. So my graph in excel has 6 slopes (X-axis being time and Y-axis being fluorescence intensity). All of the intensities decay over time. I want to see which one has the slowest rate of decay. Equations look like
y = -0.1271x + 1.8297, y = -0.0632x + 0.853, y = -0.023x + 0.7892, y = -0.0988x + 1.4117.
I would like to rank these slopes to see which one has the slowest, then 2nd slowest, then 3rd slowest until 6th slowest. And then, assess whether there is a significant difference between the slowest and 2nd slowest. What complicates the matter is these curves have slightly different starting intensities and I would like to assume equal y intercepts?
Thanks again!
karim.sarhane@utoledo.eduWed, 31 Oct 2018 20:08:03 GMTKarim.sarhane@utoledo.eduAnswer by jkpeck
https://developer.ibm.com/answers/answers/478257/view.html
@karim.sarhane
Please provide more detail on your model. What is the actual equation, and what variable(s) are you testing?Wed, 31 Oct 2018 19:37:11 GMTjkpeckAnswer by Karim.sarhane@utoledo.edu
https://developer.ibm.com/answers/answers/478235/view.html
@jkpeck
Thanks very much Dr. Peck for your answer. I am very new to SPSS, but very eager to learn its different functions. Would it be possible to show me (on Skype) how to do these computations you kindly described above ? If you have some time, please email me your preference to Karim.Sarhane@utoleo.edu. I am available anytime you are. Thanks again for your help!Wed, 31 Oct 2018 15:32:06 GMTKarim.sarhane@utoledo.eduAnswer by jkpeck
https://developer.ibm.com/answers/answers/478229/view.html
@karim.sarhane
The simplest way to compare slopes is to compute the difference of, say, X1 and X2. Then run the regression with one of X1 and X2 and the difference variable. This can be generalized in the obvious way to multiple sets of variables. For pairwise comparisons you would just use one difference at a time.Wed, 31 Oct 2018 14:59:49 GMTjkpeck