Skip to main content
Cross Validated

You are not logged in. Your edit will be placed in a queue until it is peer reviewed.

We welcome edits that make the post easier to understand and more valuable for readers. Because community members review edits, please try to make the post substantially better than how you found it, for example, by fixing grammar or adding additional resources and hyperlinks.

Required fields*

4
  • 2
    $\begingroup$ Why exactly is the analogy nonsense? Because categorizing continuous variables never produces significantly worse models? Or because using a significantly worse model never has any practical consequences? $\endgroup$ Commented Sep 4, 2013 at 11:22
  • 13
    $\begingroup$ That is simply not the case @Roland. Estimates obtained from cutoffs are only simple because people do not understand what the estimates estimate. That is because they do not estimate a scientific quantity, i.e., a quantity that has meaning outside the sample or experiment. For example the high:low odds ratio or mean difference will increase if you add patients with ultra-high or ultra-low values to the dataset. Also, the use of cutoffs implies that biology is discontinuous, which is not the case. $\endgroup$ Commented Sep 4, 2013 at 11:41
  • $\begingroup$ @Scortchi Changing from medical to surgical treatment because it is easier to explain (is it really?) would be like replacing age with height as explanatory variable. $\endgroup$ Commented Sep 5, 2013 at 7:57
  • $\begingroup$ I agree about avoiding dichotomised variables. Clinical medicine is not a rocke science where the last decimal is important. In the models I work with the results only change at the last decimal if I use categories of age vs age as continous and squared variables but increases the understanding and communicability of the associations enormously. $\endgroup$ Commented Sep 5, 2013 at 8:12