I've recently read and heard a presentation about an article published in the Academy of Management Journal this past year (Pil & Leana, 2009) concerning human and social capital and public schools. I thought it was pertinent to the issue of education currently permeating our politics.
In this article, the authors were able to analyze the human capital (experience teaching in a subject like math, and ability to teach math using a skill inventory) and social capital (how often teachers talk about subjects like math with one another in their grade and with their principal). The analysis allowed for explanation of changes in student test performance from one grade to another by controlling for student-level info (like economic status of the child, race), teacher info (human capital, see above) and team info (grade-level teams--social capital). For those stats buffs among us, she used a three-level HLM approach to analyze the data. Here's the abstract:
We found that teacher human capital that is specific to a setting and task, and some indicators of teacher social capital, predicted student performance improvement. At the team level, average educational attainment and horizontal tie strength were significant predictors of student improvement. We provide some evidence that team horizontal tie strength and density moderate the relationship between teacher ability and student performance.
In other words, here's what's up. Teachers with high ability in teaching math and whose grade-level team shared a lot of information amongst themselves about math teaching had the highest performing students. Teachers with high ability didn't have nearly the same effect without the social ties amongst the teachers! Teachers who have low ability in teaching math benefit greatly from their teams having a lot of contact with their principals regarding math. Thus, the more time spent building social capital with teachers in the same grade, the better the students do. (these were 3rd, 4th, and 5th graders in urban schools with varying demographic distributions in NYC)
Remember, this is AFTER controlling for the poverty-level of the student, which research has shown to be one of the most robust predictors of student performance.
Cool, huh? I thought all my teacher pals out there would like the info.
As a researcher/prof studying organizations, it's amazing to me how many of the problems in our education system can be attributed to poor HR practices. Between staffing teachers based on knowledge, skills, ability and motivation, compensating them using a motivational system, designing teaching jobs to be motivational in and of themselves, to evaluating performance and training teachers appropriately--seems like most of the major issues could use a good HR academic to take a stab at them. I'd love to see the DOE hire some big time HR folks to tackle these issues.
Examples I'm thinking of off the top of my head:
- Work design. We know that autonomy (control) over one's work is one of the most motivational components of job design. Yet school systems are trying their darnedest to centralize decisions on teaching, taking autonomy away from them. Increase turnover, anyone?
- We know that incentive pay (paying bonuses based on perf) only works if one has control over their performance. In theory, this is great for every job. However, if you can't quantify performance and/or the employee can't control how it is measured, it won't work. The idea of pay-for-perf based on student test scores assumes that teachers control student test scores and that they have this secret ability that they are holding back that they could release given the right amount of money. Not realistic.
- Performance evaluations. As I teach my master's students, perf appraisal has four purposes--strategic, legal, developmental and motivational. A good system accomplishes all four purposes. Additionally, three types of performance should be assessed--task perf (e.g., student achievement), contextual perf (e.g., helping other teachers, etc.), and counterproductive behaviors (e.g., badmouthing others, stealing, absenteeism). A good system counts all of these things as important. And a good system links rewards with performance. But you have to be careful not to just reward one prong (i.e., perf based pay determined by test scores).
Anyhow, just thought having a discussion about these issues was pertinent.
(also see TeacherKen's diary on this topic--it's excellent!)