In an ecologically diverse society (the Philippines), measurement equivalence across ecological settings was investigated with two culture-specific intelligence tests. Comparative reliability and factor analyses indicated that the tests (one for preschoolers, one for adults) were measuring the same ability dimensions equally reliably within each setting. Across settings, however, delta-plot analyses revealed considerable bias at the item level in tests of crystallized abilities, and the preschool tests tended to favor children from the setting where the tests were developed. The results are interpreted in terms of an ecocultural framework in which group differences in intellectual performance can reflect not only actual ability differences but differences in practice with, and exposure to, particular environmental stimuli and/or cognitive skills. The study demonstrates: (a) the need, particularly in ecologically diverse and segmented societies, to inspect tests for item bias; (b) the difficulty of culturally “equating” cognitive measures, even for very similar cultural subgroups who differ primarily in ecological setting.