Population differences in constructive memory

An experiment was conducted to evaluate constructive memory in fourth-grade low-SES black and middle-SES white children. The children were read nine short passages; each passage was composed of two premise statements and a filler sentence. The children were tested by a recognition procedure. Thirty-six test sentences were presented randomly. Four sentences were associated with each passage (true premise, false premise, true inference, and false inference). Subjects were asked to make yes/no recognition decisions based on whether the test sentence was identical to a sentence presented at study. Middle-SES white children made few errors, except to true inference test sentences. This pattern of performance is consistent with the constructive view of memory. Low-SES black children also had a high error rate to true inference items. However, this outcome may not reflect constructive memory because they also had a high error rate to some of the other types of test sentences.