I have a defined regression model for the healthy control (HC) group, with corresponding CIs of coefficients and of E(Y).

I would like to test whether individuals belonging to another population (patients), one by one individually, are fitting or not the healthy control regression line. I have both Xs and Y values for the new individuals.

I'm not sure about how to proceed: should I test the individual coefficients compared to the ones of the HC regression line, or should I predict the new Y values based on the HC model and test whether these Y values are significantly different from the actual ones?

  • $\begingroup$ How many predictors does the HC model have? $\endgroup$ Commented May 20, 2019 at 20:37
  • $\begingroup$ @JamesPhillips 4, corresponding to a cognitive ability measure (the same) across 4 timepoints close in time (we're looking for cognitive performance fluctuations, and whether these are different between healthy controls and patients, at the individual level). $\endgroup$ Commented May 22, 2019 at 14:22

1 Answer 1


I've found the answer to my own question, in case it'll be useful to someone I post it here.

Once the HC regression has been fitted, I estimated a regression line separately for each individual of the other group. I then used a F statistic to test, all together, whether the linear combination of the parameters of the new regressions are statistically different from the HC ones.

This can be done in R with the package car. Ad example, if we want to test whether a patient with coefficients 5.5, -1, -2 and -2 belongs to the HC population, we can use:

linearHypothesis(lm_healthycontrols,  c("(Intercept) = 5.5", "x1 = -1", "x2 = -2", "x3 = -2"))

If the p-value is significant (say p<0.05), the patient is considered as belonging to a different population (patients).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.