Teaching to the test

From Learning
Revision as of 23:30, 24 May 2014 by Vipul (talk | contribs) (For a criterion-referenced assessment)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Definition

Teaching to the test refers to a phenomenon where a teacher or instructor narrowly focuses the content and method of teaching with the goal of optimizing for performance on a particular test or tests. The term is typically used in the context where the test is externally conducted. The test may be a widely used standardized test or it may be a common test between different sections for a course in a school or college setting (an example outside academics is students being taught how to get a driving license by being made to extensively practice the specific driving route and maneuvers tested for acquiring the driving license).

Evaluation of teaching to the test

For a criterion-referenced assessment

For a criterion-referenced assessment, i.e., a test whose goal is to evaluate whether an individual has acquired a particular set of knowledge and skills, teaching to the test is:

  • not a problem if the test is designed so that one can do well on the test only if one acquires the underlying knowledge and skills.
  • a problem if there are strategies to improve performance on the test that do not reflect improvement on the underlying knowledge and skills.

In the latter situation, teaching to the test can be viewed as an example of overfitting.

For a norm-referenced assessment

For a norm-referenced assessment (i.e., a test where the goal is to rank or sort people), teaching to the test is a mixed bag:

  • If some people receive teaching to the test and others don't, then the norm-referenced assessment distorts the relative ranking of individuals.
  • Even if everybody received teaching to the test in the same measure, the effect of too much teaching to the test might be to compress the range of scores, and therefore increase the relative importance of measurement error and distort the rankings.
  • Teaching to the test can increase the reliability of the assessment by reducing the incidence of people underperforming their potential because of unfamiliarity with the test environment.