Dear all
I would value some advice on the apparent difference between the Chi-square test for a linear trend (as provided under the SPSS crosstabs option when such a trend is detected) and the linear trend test forthcoming from a Kaplan Meier analysis on SPSS.
I wish to test for the monotonicity of a relationship for mortalities and note that the two tests produce different p-values. The latter of these two tests may be selected as the 'linear tend for factor levels' option in SPSS only once one has selected the log-rank test option. If these tests are truly different, could someone kindly provide me with a reference for the 'log-rank test for a linear trend' and recommend which is the better when looking for a monotonic relationship in conjunction with a Kaplan Meier analysis. I am aware that the Chi-square test for a linear trend is referred to in published work as having being used for this purpose but if there is a better option, I would like to learn about it. Further, it would appear that the 'log-rank test for a linear trend' carries the assumption of equal 'distance' between levels but I am not aware that this is an assumption for the other test.
Sorry if there is no such thing as a 'log-rank test for a linear trend' but I am taking SPSS as read.
Many thanks
Best wishes
Margaret
---------------------------------
How much free photo storage do you get? Store your holiday snaps for FREE with Yahoo! Photos. Get Yahoo! Photos
|