This investigation compares child language obtained with standardized tests and samples of spontaneous speech obtained in natural settings. It was hypothesized that differences would exist between social class and racial groups on the unfamiliar standard tests, but such differences would not be evident on spontaneous speech measures. Also, higher correlations between test and spontaneous scores were expected when subject-examiner similarity was maximized. Subjects were 40 4-year-old preschoolers. They were tested with subscales of the Wechsler Preschool and Primary Scale of Intelligence (WPPSI) and the Illinois Test of Psycholinguistic Ability (ITPA). Spontaneous speech was recorded by inserting a small transmitter and microphone into vests worn by the children. The transmissions were recorded and scored, using Hunt's T-units, a vocabulary range measure, and a count of concepts used in speech. Analysis of results indicate that both hypotheses were partially supported. Discussion focuses on methodological issues which may have contributed to the unclear data. (DP)