ADMINISTRATION OF GROUP INTELLIGENCE TESTS TO EAST AFRICAN PUPILS

S ummary . Intelligence tests could be even more valuable for educational predictions in the developing, than in Western, countries, because of the large variations in efficiency and linguistic background in different primary schools. But there may be serious problems in getting them across to unsophisticated pupils, especially with printed instructions only. An experiment was carried out with fourteen top primary classes in Tanzanian schools, where a short non‐verbal battery was given with printed, or minimal oral, instructions in English or Swahili. A second parallel form was taken after this practice or after various types of coaching. The mean improvement varied considerably in different classes and different subtests, and was affected by certain motivational or attitude factors. Oral instructions gave better mean scores than printed, and Swahili better than English. The overall rise with practice averaged 3 to 4 points in deviation I.Q. units, with coaching about 7 points. These figures are not unusually large, possibly because the pupils were already sophisticated to objective attainment tests. Thus they suggest that intelligence tests could be introduced with the same kind of familiarization procedures common in England, though the actual items should be constructed locally. Comparison of item difficulties with those found in England showed that abstract non‐verbal items may be perceived differently in different cultural groups.