What Evidence Will Support My Claim? Evidence Selection Among Young Adolescent Science Students

• In Event: Focusing the Analytic Lens on Evidence-Based Reasoning: Where We Are and Where We Need to Go

Abstract

Schooling traditionally affords students more experience in learning and practicing procedures than in identifying what a situation calls for. When asked to choose appropriate numerical data to support causal claims, college students in an earlier study performed surprisingly poorly. Almost all chose limited, inconclusive data as sufficient evidence, despite having available more comprehensive data needed to support their claim.

Objectives

We highlight this weakness to make the case that choosing one’s evidence constitutes an important metacognitive intellectual skill and educational objective, one within and beyond the science curriculum. Here we examine the extent to which the same concerns hold true in a K-12 population. We find they do, even among students at a high-performing school at which argument skills are emphasized and regularly practiced.

Theoretical Framework

Disposition, as well as competence, become critical when people have a largely free hand in deciding whether and how to bring to bear evidence to support a claim. Yet it has been the subject of much less investigation, except perhaps in research on epistemological understanding, where people are queried regarding how they know what to accept as true.
One of us has written at length about mature epistemological understanding as a developmental achievement (Kuhn, 1991), one foundational to the disposition, as distinguished from competence, to engage in what has been coined epistemic vigilance. The marshaling of evidence thus assumes its key role as an essential tool in distinguishing among conflicting claims.

Methods

The main problem introduced in writing to a class of middle-school science students was this: A National Health Data Bank have identified children’s obesity (excess weight) as a concern and want to determine the cause. They have collected information from a broad, representative sample of 200 four-year-olds. These are the numbers they came up with. Each card shows the children they identified who fall into the category shown on that card. The Data Bank researchers now need someone to analyze the results and to write a summary report of what these findings show about the causes of young children’s being overweight. [bold text in original].
Twelve cards introduced numerical data – just a single number, identifying the number of children who did/didn’t possess one of three binary attributes (diet quality, exercise, and parent weight) and who were/weren’t overweight (e.g., “This card lists cases of CHILDREN WHO DO NO EXERCISE and ARE NOT OVERWEIGHT”).

Results and Discussion

Younger students were no more cautious than college students in inhibiting the assertion of causal claims without providing sufficient supporting evidence. The task affords students considerable freedom to organize a response by choosing their evidence. The problem thus becomes one of epistemological standards: What constitutes relevant evidence and sufficient evidence? This challenge is infrequent in both research tasks and classrooms.
The epistemological failing exhibited raises concern about susceptibility to accepting unsubstantiated claims encountered in academic work and elsewhere. The incomplete data these students chose as good evidence for a causal claim are ubiquitous in media. Consequences for both personal and public discourse have become painfully evident.

Authors

• Daniel Lerman

• Anahid S. Modrek, Thomas Jefferson University

• Deanna Kuhn, Columbia University