By Stuart A. Singer, Author of The Algebra Miracle
Perhaps because I am a former math teacher I cannot help myself. Maybe my extensive coaching background is what makes it even more compelling. Regardless of the precise motivator the fundamental conclusion seems so obvious.
Education needs to utilize data analysis more effectively.
Stats are bursting out all over
Data analysis can be a powerful tool for innovation in a multitude of endeavors. It can illuminate the path to better outcomes and accurately affirm success and failure. It is not, however, a static process. In order to maximize its effectiveness constant reevaluation is required. Otherwise conclusions made based on statistics can quickly become inaccurate and irrelevant.
One powerful example of such numerical evolution was evident in the aftermath of the gubernatorial election in Virginia. When the final results had been tabulated several newscasts explained the victory in these terms—the Democrat won the women’s vote by a larger margin than the Republican won the men’s. From there the speculation became focused on what specific issues had caused this “gender gap”.
But a day later another set of numbers presented a significantly different perspective. When one statistician divided the same voters into the category of either “married” or “unmarried” new conclusions emerged. A majority of married men and married women favored the GOP; unmarried men and women did not. Suddenly, because of these numbers the conversation and potential suppositions veered in a very different direction.
Similar numerical adjustments are occurring in the world of sports. A recent article in the Washington Post explained that the Nationals new baseball manager Matt Williams based a large portion of his improvement plan on the introduction of something new to the organization—data analysis. The plan is basic. An individual will be hired who will filter through the statistics provided by advance scouts (individuals who attend other team’s games to acquire data) and their video counterparts (watch tape of future opponents with the same goal). Then the overall analysis will be implemented as strategy in the Nationals games.
Statistics are not always convenient
Failure to correctly interpret data can be highly misleading and detrimental in education. During the last four years of my teaching career I was regularly at odds with district policymakers. They had a very simple prescription for overcoming any academic problems being encountered in math—introduce the curriculum sooner and move through it faster. To that end the district embarked on a campaign to have 100% of the students take Algebra 1 by grade eight. Another component was to move more topics into Algebra 2 making the course more rigorous. It was a wonderful approach in terms of public relations. Much like the fictional Lake Woebegone (“where all of the children are above average”) a bumper sticker stating my eighth-grader has mastered Algebra 1 is wonderful if the data had given validity to such an outcome.
It did not.
For several years in an effort to reach that district goal our feeder middle school placed the top half of its eighth grade math students into Honors Algebra 1; the remainder was relegated to taking a “regular” version of the course when they were in high school. Based on these circumstances, it was not surprising that few if any of the honors students received a grade below “B”; the report cards at the high school were not nearly as glowing. But there was one pesky piece of information that sullied the euphoria. Both groups took the exact same state mandated end-of-course exam. The results were revealing. The honors students had an average of 472; the less historically successful math group at the high school came in at a virtually identical 469. More statistically important was that when the top 10% of each group was removed the regular students scored significantly higher.
The implications of such analysis were profound. It clarified why so many of the middle school Algebra 1 students struggled in Honors Geometry and large portions of this population did poorly in the remainder of their math careers. It was not unusual to find “regular” Algebra 1 students out performing their “honors” counterparts in Algebra 2 and Pre-calculus. In fact these outcomes were also easily predicted if attention was paid to the data. Over a period of several years it was determined that all but one of the honor students who scored below 480 made a “D” or “F” in the subsequent Advanced Geometry course. Some trends are impossible to ignore.
When our school introduced double-block (DB) Algebra 2, loud complaints emanated from the district leaders. The word “incomprehensible” was uttered by the system’s math coordinator when describing the concept. The results of this experiment proved the inaccuracy of that pronouncement. In the initial year of implementation one teacher was assigned a schedule consisting of the two pilot double block Algebra 2 sections and one regular. The DB classes were filled exclusively with students with extremely weak math resumes—poor grades and dire recommendations from their prior math classes. All three groups, regular and double-block, were given the same assignments, quizzes and tests. The only difference was the amount of class time allotted. At the end of the year the results were stunning. Although far weaker academically by virtually every previous measure, the final grade averages of the DB students were only one point lower. Their pass rates on the state end-of-course exam were almost identical. As the program expanded the data continued to reflect this success.
The danger is that if these students had been placed among the general population and done poorly, their failure might well have been attributed to a weak background or poor teaching rather than a need for more time. That general analysis may have been substantiated by numbers but would have missed the more important point.
Think locally not globally
This school which had a free or reduced lunch rate of 54% was located in one of the three wealthiest counties in the country. Consequently, policies that might be effective for much of the district were not necessarily best for this particular building. Even within the same geographical area every student body is unique. As a result of that reality data analysis should be school based and constantly evolving. It is a trend that is changing politics and baseball. It needs to find an equally comfortable home in education.