These programs for PCs accompany the paper: Crawford, J. R. & Garthwaite, P.H. (2006). Comparing patients’ predicted test scores from a regression equation with their obtained scores: a significance test and point estimate of abnormality with accompanying confidence limits. Neuropsychology, 20, 259-271.
Regression equations are widely used in neuropsychology to draw inferences concerning the cognitive status of individual patients. For example, an equation predicting retest scores from scores at original testing can be used to test whether there has been change in a patient's level of functioning. Equations can also be used as an alternative to conventional normative data by providing "continuous norms" as when a patient's score on a neuropsychological test is compared to the score predicted by their age (see paper for further examples).
These programs test if there is a significant discrepancy between an individual's obtained and predicted score (one- and two-tailed p values are provided). They also provide a point estimate of the abnormality of the discrepancy (i.e., a point estimate of the percentage of the population exhibiting a larger discrepancy) and accompanying confidence limits on this quantitiy.
A commonly used alternative to these methods of analyzing the discrepancy between obtained and predicted scores involves dividing the discrepancy by the equation's standard error of estimate and treating the result as a standard normal deviate (the p value for this z is then obtained from a table of areas under the normal curve). Monte Carlo simulations show that, unlike the method implemented in these programs, the latter method does not control Type I errors and overestimates the abnormality of an individual's discrepancy. In addition, because it does not acknowledge the uncertainty associated with sample regression statistics, it cannot provide confidence limits on the abnormality of the discrepancy.
The program regdiscl.exe is for use with bivariate regression equations (ie single predictor). Click here to download the program (462Kb)
The program regdisclv_V2.exe is for use with multiple regression equations and requires that you have access to the correlation matrix for the sample used to build the equation. (The program has the suffix V2 as this is an updated version to fix a bug in the original) . Click here to download the program (463Kb).
Alternatively both files can be downloaded as a zip file; click here (465 Kb)
Note: These programs assume that an existing regression equation is available; click here for details of programs that build equations from summary dataYour web browser is most probably configured to recognise that the files are executable. If you have any problems (i.e. the browser treats them as text files), hold down the shift key when clicking.
Once downloaded, the programs can be run by any of the normal Windows procedures i.e. by clicking on file in File Manager or by placing on desktop etc.
The author of this software (John R Crawford) and the University of Aberdeen make no representations about the suitability of the software or about any content or information made accessible by the software, for any purpose.
The software is provided 'as is' without express or implied warranties, including warranties of merchantability and fitness for a particular purpose or noninfringement.
The software is provided gratuitously and, accordingly, the author shall not be liable under any theory or any damages suffered by you or any user of the software.
If there are any problems please e-mail me at j.crawford@abdn.ac.uk. Further contact details are available in the footer of this page.
