Preschoolers’ understanding of the control of variables strategy

• In Event: 1-029 - The emergence of scientific reasoning skills in early childhood

Integrative Statement

The Control of Variables Strategy (CVS), a strategy for determining cause–effect relations between variables by designing unconfounded experiments, which manipulate the variable in question while keeping all others constant (Chen & Klahr, 1999), has traditionally been shown to develop late, only in adolescence, with even many adults failing to produce it spontaneously (Kuhn, Garcia-Mila, Zohar, & Andersen, 1995; Bullock & Ziegler, 1999; Schauble, 1996). However, research has also shown that spontaneously producing controlled experiments can be much more difficult than selecting controlled experiments. In fact, in choice tasks, even third graders prefer controlled over confounded experiments and by the sixth grade, children reach adult-like performance. It has also been shown that elementary school children produce and understand contrastive tests before they begin to select controlled tests (Bullock & Ziegler, 1999). Preschoolers’ abilities in CVS have received little attention, but there is evidence of a beginning understanding of CVS when provided with support for producing controlled experiments (van der Graaf, Segers, & Verhoeven, 2015).

In the present study, we systematically investigated preschoolers’ abilities in CVS with novel, knowledge-lean choice tasks using the blicket detector paradigm (Gopnik & Sobel, 2000). Children must select a “good” test for a hypothesis about the cause of a light effect. The 2-choice task represented a contrastive test and the 3-choice task represented a controlled test. Each child performed each task twice. The experimenter placed a stick of two or three differently colored bricks on a “light box” and the box lit up. Children were instructed to find out if the X brick (e.g. yellow brick) was a lighter. To do so, they could pick one stick from two or three options, to place on the box. In each task, there is one correct stick that varies the X brick and keeps the other bricks the same. The other options vary either two or three bricks.

Results showed that preschoolers (n = 110, MageYoung = 4;8, MageOld = 6;1, range: 3;5 - 6;9) selected the correct option in at least one of two trials, more often than to be expected by chance (2-choice: 62%; 3-choice: 46%). Across two trials, 40% of children were consistently correct in the 2-choice task, but only 20% were in the 3-choice task, suggesting that the 3-choice task is more difficult. This is further supported by the difference in performance between younger and older children: more older children selected the correct choice twice in the 3-choice task (26% vs 15%).

Van der Graaf et al. (2015) found that, after repeated feedback on experimental design, 47% of preschoolers could produce a controlled experiment with three variables. Our results show that about the same proportion of preschoolers could spontaneously select a controlled test without support or training. Further, our results indicate that preschoolers also find contrastive testing easier to understand than CVS. These findings indicate genuine scientific reasoning abilities in preschoolers. Age-related developmental change and individual differences remain to be explored in depth in future research.

Authors

• April C Moeller, Ludwig-Maximilians-Universität München
  Presenting Author

• Beate Sodian, LMU Munich
  Non-Presenting Author