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.

3
  • $\begingroup$ First of all, appreciate your effort and time of your response. As far I can see the missingness is a strong problem of my database, in fact, seems to be a problem with one of the variables there, which in the long format process was distorted. But I need to point out some details of you comments. Which na.action would you recommend? na.omit /na.exclude, AFA I know lme don't work under the listwise deletion procedure losing statistical power. About the collinearity, didn't know about check_collinearity. Is it different from caret::findLinearCombos? $\endgroup$
    – Javier Hernando
    Commented Mar 26 at 9:20
  • $\begingroup$ The findLinearCombos function uses a QR decomposition to check for linear combinations and then removes them. This is not the same as VIF, which estimates how much the variance is inflated by the inclusion of certain predictors as a result of multi-collinearity. I don't promote any action for missingness before finding out why it exits. Filter the data for NA values, check their cause, figure out if it is due to completely random factors or because of observed data, and then act accordingly (per the article I referenced). Without knowing more about the NA values, I can't give concrete words. $\endgroup$
    – Shawn Hemelstrand
    Commented Mar 26 at 10:02
  • $\begingroup$ One should only delete missing values in any case if there are strong reasons for doing so (such as clearly erroneously coded values or an artefact of some pivoting of the data, which should anyway be fixed before moving forward). $\endgroup$
    – Shawn Hemelstrand
    Commented Mar 26 at 10:03