Probing for hypotheses with adults and children: Blank trials and introtacts.

Abstract

This paper discusses two probe techniques (blank trials and introtacts) in explorations of problem solving by adults and children. Originally, the blank-trial probe was developed by Levine (1966) to detect the nature of adults' processing in discrimination-learning tasks; it consists of a series of no-feedback ("blank") trials which are inserted between each feedback trial. This probe yielded extremely orderly data and directly validated an emerging theory of human discrimination learning, hypothesis (H) theory. Blank-trial probes clearly demonstrated that adults do employ and test Hs in a systematic fashion.The blank-trial probe was then employed to investigate developmental aspects of H theory. Gholson, Levine, and Phillips (1972) discovered that most elementary-school children actively employ and test Hs in discrimination-learning problems. With a slight modification of Levine's (1966) stimuli, blank trials also enabled the detection of systematic response patterns in kindergarten children which were not predictions about the task. Analysis of developmental blank-trial data yielded a comprehensive expansion of H theory. It is clear that people select H& according to some general plan (a System). Examining the Systems employed at different age levels indicated a developmental continuum of sophistication. Both qualitative and quantitative changes occur in problem-solving styles.Although the blank-trial probe has been a useful tool, it is methodologically cumbersome. The search for an alternative probe technique led to the introtact, a verbal report of an H. Karpf and Levine (1971) compared the two probe conditions in adults. They demonstrated that introtacts were valid probes for Hs and that neither type of probe affected the problem-solving process.Phillips (1974) explored similar questions with children. She developed a procedure, appropriate for use with children, which yielded valid H data via introtacts. Phillips' results suggested that H theory is basically correct in its portrayal of the problem-solving process. However, she found that the use of probes does affect younger (second-grade) children's processing. Blank trials produce somewhat more stereotypic responding by these children, whereas introtacts facilitate problem solving (both relative to a control condition). Sixthgraders' performance is unaffected by probe condition. The implications of these findings for H theory, developmental theory, and developmental research are discussed.

Interactions between children’s metacognitive abilities, working memory capacity, strategies and performance during problem-solving

“…Within each problem the pattern of presentation of the clothing items was standardised to follow the kind of "orthogonal" sequence originally devised by Levine (1966) Early studies of hypothesis-testing deduced subjects' hypotheses on an MDL task simply from their pattern of choices, using Levine's (1966) blank-trial procedure. More recently, Phillips and Levine (1975) and Kemler (1978) demonstrated that this procedure led to an underestimation of children's problem-solving abilities, and that supplementing this by asking children for verbal "hypotheses" (sometimes referred to as "introtacts") provided a more accurate picture. This procedure was, therefore, adopted within the present study, with the children being asked on each trial which card they thought represented Anna (or for "negative information" type problems which twin the one card shown represented).…”

Distinct discrimination learning strategies and their relation with spatial memory and attentional control in 4- to 14-year-olds

“…However, the first explanation seems to fit the results of the slow learning children, whereas the second and third explanations seem more likely in the nonlearning children, as judged by the estimated accuracies. Latent learning modeling of performance on learning tasks that involve blank trials may help to distinguish between these strategies (Phillips & Levine, 1975). However, performance is expected to suffer in these tasks because blank trial tasks are more demanding.…”

“…Studies of the hypothesis testing behavior of children suggest that young children might not apply the strict win-stay and lose-shift substrategies (Gholson, Levine, & Phillips, 1972;Kemler, 1978;Phillips & Levine, 1975). However, these studies arrived at different conclusions, specifically concerning the use of the lose-shift substrategy in the age range of 4 to 10.5 years.…”

Computer Programming and Generalized Problem-Solving Skills:

“…In other words, first graders cannot demonstrate the strategies in the context in which they were taught, let alone transfer such strategies (see also, e.g., Canton & Spiker, 1978;Phillips & Levine, 1975).…”