Full house or Pandora's box? The treatment of variability in post-Piagetian research.