This paper considers the technique of the exact partitioning of x 2 contingency tables. Methods are presented for partitioning contingency tables into components. A general equation for x 2 is derived. The equation may be used for the calculation of exact x 2 values for (a) nonexhaustive sets of categories, and (b) situations in which some cells have small expected frequencies.

Sixty first-grade children from two different socioeconomic levels were presented a concept-acquisition problem after either 0, 6, or 12 trials of random reinforcement. The overall performance of the high-socioeconomiclevel children was superior to the low-socioeconomic-level children, but the results also indicated a significant interaction between socioeconomic level and number of random reinforcements. With an increase in the number of random reinforcements, the behavior of the high-socioecenomic-level children deteriorated to the same level as the low-socioeconomic-level children. The experimental manipulation of prior reinforcements produced low-socioeconomic-level behavior in high-socioeconomic-level subjects. The data are interpreted as further evidence that the inconsistent reinforcement histories of low-socioeconomic-level children contribute to their typical ineffectual performance.

No abstract

The subjects. 108 4-year-old children from two socioeconomic levels. were trained on a three-phase simultaneous discrimination task. In Phase 1, stimulus element A was reinforced; stimulus element B was held constant. In Phase 2. element B was reinforced on all trials; element A was available and reinforced on 100o~, 75%, or 50% of the trials and held constant on the remaining trials. In Phase 3. the two elements were tested separately. conjunctively. and disjunctively. Lower socioeconomic subjects (a) performed well on element A and blocked element B if element A was available on 100% of the P~ase 2 trials. (b) performed poorly on both elements A and B if element A was 750~ available. and (c) performed poorly on element A and well on element B if element A was 50% available. Higher socioeconomic subjects performed well on both elements A and B under all conditions; however.

Sixty 4-year-old children from two socioeconomic levels were run in a three-phase experiment designed to investigate the effect of Phase 1 training upon the acquisition of stimulus control by two stimulus dimensions. Phase 1 consisted of three different training conditions; one stimulus dimension was correlated with reinforcement, and the second stimulus dimension was: (a) also correlated with reinforcement, (b) uncorrelated with reinforcement, or (c) held constant. In Phase 2, all subjects received reinforcement correlated with both dimensions. In Phase 3, the two dimensions were tested conjunctively, disjunctively, and separately. The results showed that stimulus control was more dependent upon Phase 1 training in children from the higher s!)Cioeconomic level and was more dependent upon a priori preferences in children from the lower socioeconomic level.