February 21, 2013 · 4:37 PM

Dear editor,

In response to the editorial of Feb. 15, regarding test results in schools, please do not repeat claims that 50 years of educational research on effective schools have shown to be false.

I refer to the assertion that children from single-parent families do not do well on educational tests.

In fact, the research shows that the best single predictor of children’s school performance is the educational attainment of the mother. The claims that either income or single-parent family status accurately predicts performance are false, and result from misreadings of the statistical analyses involved.

It is true that income and education level tend to be correlated, which has lead to the very common false claim about income and attainment. Since single parents are often poor, it is easy to draw the inference you repeat, about the school performance of their children.

With regard to the use of tests to assess the performance of schools generally, the only really useful assessments are those which measure student performance growth from Time One to Time Two, perhaps from one grade to the next.

Even these require so much interpretation that they are only crude measures of school performance, which need to be supplemented with very careful observational and interview information to make accurate judgements.

On the other hand, your headline, Student assessment tricky, is absolutely correct. I spent 20 years as a researcher trying to do school assessment, a related task which it is even harder to get right.

Dr. Peter Coleman,

Comox Valley

Editor's note: Dr. Peter Coleman is a School District 71 trustee and Professor Emeritus at Simon Fraser University.