Poster #34 - Scientific Reasoning Development: Children’s Evaluation of Experiments and Evidence

• In Event: 3-106 - Poster Session 11
  In Poster Session: PS 11 Section - Cognitive Processes

Integrative Statement

We investigated scientific reasoning among elementary school students by assessing experiment and evidence evaluation. Previous studies have either not fully distinguished these two abilities, or have assessed them with different kinds of problems. We directly compared experiment and evidence evaluation within the same task and also examined recognition of confounded variables.
Using a modified version of procedures used by Sodian et al. (1991) and Piekny and Maehler (2013), we presented students in first-, third-, and fifth-grade (N = 54) with two sets of problems: (a) 1-Variable problems examined recognition of the difference between conclusive and inconclusive tests, and (b) 2-Variable problems further assessed recognition that confounded variables make an experiment inconclusive. Each problem included an experiment evaluation question and an evidence evaluation question. In the 1-Variable condition, the experimenter explained that two siblings/friends disagreed about the size of a mouse in their home/school. For the inconclusive trial, the siblings placed food in a box with a big opening. Participants were asked (a) “If Callie and Andy put the cereal into the box with the big opening, will they know if the mouse is big or small?” and (b) “In the morning, the cereal is gone from the box with the big opening. Do they know if the mouse is big or small?” For a conclusive trial, the siblings placed food in a box with a small opening. In the 2-Variable condition, there were two types of cereal as well as two sizes of box openings. An unconfounded problem was “Ryan and Liz put the blue cereal in both mouse houses. If the cereal is gone from the small house but not the big house, would they know what size the mouse is?” and “In the morning, the food is gone from only the small house. Do they know for sure if the mouse were big or small?” A confounded problem was “Bailey and Alex put different types of food in the mouse houses. They put the yellow cereal in the big house and the green cereal in the small house. Etc.”
We predicted that (a) for 1-Variable problems all age groups would perform above chance for experiment and evidence evaluation, and (b) for 2-Variable problems older students would perform better than younger children.
For the 1-Variable condition, a 3 x 3 x 2 (Age x Gender x Judgment) ANOVA did not reveal any significant effects. All age groups performed significantly better than chance for evidence evaluation, but not for experiment evaluation. In the 2-Variable condition there was an Age effect: third- and fifth-grade students performed better than the first-graders, and also performed above chance for both confounded and unconfounded problems
The pattern of mixed performance across all age groups suggest that (a) basic scientific reasoning abilities emerge gradually during elementary school, (b) these abilities are not entirely robust even at fifth-grade, and (c) experiment and evidence evaluation may be distinct at least for some problems.

Authors

• Taneisha Vilma, Northern Illinois University
  Presenting Author

• Bradford H Pillow, Northern Illinois University
  Non-Presenting Author