No, it is probably better to use and keep all data. Repeated-measures methods as implemented in functions like lmer from R package lme4, and gam and s from mgcv do not require higher-level units to have the same number of lower-level units.

Yes. Cross-validated accuracy is likely less optimistic when you sample at the higher than at the lower level.

No, it is probably better to use and keep all data. Repeated-measures methods as implemented in functions like lmer from R package lme4, and gam and s from mgcv do not require higher-level units to have the same number of lower-level units.

Yes. Cross-validated accuracy measures are likely less optimistic when you sample at the higher than at the lower level.

No, it is probably better to use and keep all data. Repeated-measures methods as implements in functions like lmer from R package lme4, and gam and s from mgcv do not require higher-level units to have the same number of lower-level units.

Yes. Cross-validated accuracy is likely less optimistic when you sample at the higher than at the lower level.

No, it is probably better to use and keep all data. Repeated-measures methods as implemented in functions like lmer from R package lme4, and gam and s from mgcv do not require higher-level units to have the same number of lower-level units.

Yes. Cross-validated accuracy is likely less optimistic when you sample at the higher than at the lower level.