Time cluster analysis using lmer does not work with factor predictors

[brockf]

Currently, the example here doesn't work:

http://www.eyetracking-r.com/docs/analyze_time_clusters

To get it to work, you must recode Sex (factor) to a numeric variable, and then pass by shuffle_by=Sex to analyze_time_clusters. This is because of a mismatch between the auto-factored variable in lmer() (which is named SexM automatically) — it doesn't match a column name.