Friday, March 18, 2011

Comparing public and private high schools

Whenever I am with parents of middle school children, one of the topics that invariably dominates the discussion is about the choice of high school and which one is better and why. Although the actual selection often gets determined by where one lives, there is some amount of decision making involved as people do plan their home purchases or rentals based on choice of available schools. Sometimes the choice is between sending the child to a private school at a considerable annual expense, versus going to the home school, with zero or minimal tuition fee.

Admittedly, the notion of a good school is a very subjective one. Some people look at the campus, the courses offered, proximity to home, opinion of their friends, reputation for good or bad things and other such considerations. However, for majority, goodness simply means emphasis on academic excellence. But how to quantify this in an objective and reliable manner?

California schools are annually evaluated and assigned a performance index called Academic Performance Index or API, which is supposed to be an indicator of academic excellence, among other things. The chart below shows the Growth API of 5 different public high schools for last 3 years, 2010, 2009 and 2008. Two of the schools, Wilcox High and Santa Clara High, are in Santa Clara Unified School District, my home school district. Two high schools, Monta Vista High and Mission High, are other bay area high schools known for academic excellence. I also included Gretchen Whitney High, the school with highest API score in California, for comparison purposes.

Growth API of Selected California Public Schools (2008-2010)

This does give a fairly good idea of relative academic performance of these schools making it somewhat easy to compare them. But what about private schools? How do you compare a private school with another private school or even a public school?

Turns out that there is another metric that can be used for comparison purposes: Combined Average SAT scores, ie; the average of the sum of critical reading, math and writing scores of all the students who took SAT. These scores for public schools, along with API and other academic metrics, can be found at California Department of Education DataQuest website. I wasn't able to find an equivalent database for private schools but most do publish their scores for the latest available year at their websites and this is how I was able to get the scores for Archbishop Mitty, St. Francis, Bellarmine, Notre Dame and Harker, five well known private high schools in the area where I live.

Llet us look at the average SAT scores of these schools:

Average Combined SAT Score

What this shows is a ranking of both public and private high schools based on average SAT scores. Not surprisingly, the ranking of public schools based on average SAT score is not much different from the one based on API. However, what did surprise me is that the good public schools compare quite favorably with private schools.

A key difference between the API and the average SAT score is that the former is based on performance of all the students whereas the later is based on scores of only those who take the SAT test. This percentage can vary widely. For example, only 45% students of Santa Clara High took SAT in 2008-9 whereas this percentage for Gretchen Whitney was 100%. This is true in general -- low performing schools have lower percentage of students taking the SAT. As only academically inclined students are likely to take SAT, it can be argued that the low performing schools will have even lower average SAT score if all the students took the test.

At this point, a natural question to ask would be: how well the API correlate with the average SAT score? Let us look at a chart that plots both these metrics on the same chart for all the five public schools. Note that I have adjusted the intervals for API (left vertical axis) and Average SAT Score (right vertical axis) a bit to compensate for different ranges.

2010 Growth API and 2008-2009 Avg. SAT Score

Not surprisingly, the correlation seems to be pretty high, validating our choice of Average SAT score as a predictor of academics and also a metric to use for comparison purposes.

I should add that the choice of schools in this post is somewhat arbitrary, and though meaningful to me, is likely to be completely meaningless for your purposes. In that sense the actual ranking is not as useful as the general idea of using average SAT score to rank a collection of private and public schools.

Other thing to keep in mind that an academic ranking can only be just one input to your selection process. Only you can decide what other aspects, such as sports, extra-curricular activities, tuition fee, proximity to residence, friends, special facilities etc. are important to you.