Putting rubrics to the test

Andrade, H. L., Du, Y., & Wang, X. (2008). Putting rubrics to the test: The effect of a model, criteria generation, and rubric-referenced self-assessment on elementary school students' writing. Educational Measurement: Issues and Practice, 27(2), 3-13.

Background

There is limited evidence to support that rubrics can promote learning and achievement.

Aims

This study investigates the effects of reading a model assignment, generating a list of criteria for the assignment, and self-assessing according to a rubric, as well as gender, time spent writing, prior rubric use, and previous achievement on primary school students' scores for a written assignment.

Sample

The participants were 116 students in one class in each of seven public schools in northeastern United States. The table below summarizes the distribution of the classes in the experimental and control groups:

Method

Students in the experimental condition used a model paper to generate a list of criteria for an effective story or essay before receiving a written rubric, and using the rubric to self-assess their first drafts by underlining key phrases in the rubric, then by underlining or circling in their drafts the evidence of having met the standard articulated by the phrase. They were asked to write themselves a reminder to make improvements when writing their final drafts. Students in the comparison group were asked to generate a list of criteria for an effective story or essay and review their first drafts. They did not receive a model paper nor a written rubric, and did not self-assess according to a given rubric.

Findings

This study provides support for the hypothesis that having students use model papers to generate criteria for a writing assignment and using a rubric to self-assess their drafts is positively related to the quality of their writing. The essay score was not significantly correlated with writing time. Prior rubric use was not correlated with essay score. Controlling for previous achievement, a General Linear Model two-way ANOVA was used to analyze the main effect of treatment and gender. Results showed that the treatment group's writing scores were significantly higher than the comparison group's scores. Girls tended to have higher scores than boys. The interaction between treatment and gender was not significant.

Implications

Primary school students should actively engage in critiquing sample pieces of writing, in thinking about the criteria that are in the rubrics by which their writing will be evaluated and in self-assessment of their works in progress.

Original article retrieved from EBSCOHost.