“Look on the back of their baseball card.”

That’s the common reply from baseball managers and general managers around the major leagues when asked about what kind of production they expect from a player in an upcoming season.

It is a reference to the career stats of baseball players, which appear on the back of their baseball cards. Most often, at the end of the season a player’s stats will be more or less as their career stats.

This basic assumption is so accepted in Major League Baseball that teams offer tens of millions of dollars to those players with the strongest career stats, while players with the weakest stats tend to disappear after a few years.  The stakes don’t get much higher than that.

But in education, it’s still rare for similar high stakes decisions about school personnel (teachers, principals, and other administrators) to be made using data such as student test scores. Critics argue that evaluating educators primarily based on student test scores would be unfair. That even the most complex statistical models designed to isolate educators’ impact, such as value-added growth models, are not reliable enough to base high-stakes decisions such as salaries or tenure.

Now, these critics argue correctly that there is only a moderate correlation between a teacher’s past performance, as measured by value-added models, and their future effectiveness. For example, a Brookings report on value-added models stated that the correlation between test-based measures of teacher effectiveness between one school year and the next is between .30 and .40 (1.0 would mean a perfect correlation and 0.0 would be no correlation). This is a low to moderate correlation in statistical terms. Sounds like the critics have a strong point.

However, Brookings also points out when comparing that correlation to statistics used in other professions, such as Major League Baseball, the correlation doesn’t sound very weak.

Take batting average—one of the most widely used statistics to evaluate a player. The between-season correlation for batting averages for Major League Baseball players is just .36. Yet, a team will pay a player millions for hitting .300 and cut a player for hitting .250.  

Of course batting average is just one statistic teams use to evaluate a player and value-added scores are just one statistic that could be used to evaluate educators, but both can be key measures to determine how effective they will be in the future if used correctly.  

Do teams sometimes make the wrong decision by paying millions for the player who hit .300 last season while cutting the player who hit .250? Yes, it does happen. But overall, the teams are better off signing the players who had the strongest stats on back of his baseball card. Our schools could be more effective if they kept the educators with the strongest stats as well.– Jim Hull