The middle school at Ruffing has two levels, one for sixth graders and one for seventh and eighth graders. (The latter I guess follows Montessori's philosophy of multi-age classrooms.) Wikipedia says Montessori schools generally don't assign homework, but this school clearly does. Heck, Caroline has even been given some homework. They also have separate teachers for Spanish, band and gym, which is a good sign. The school has been around since 1957 and teaching middle school since 1977, which is also a good sign. Looking for a certain amount of longevity in your alternative educational institution rules out the freakier places. Ruffing doesn't have much else on line, and this is not so good. "Montessori" isn't governed by
As I understand it, the centerpiece of Montessori education is giving children a large block of time to explore at their own direction a prepared environment filled with activities that allow the children to learn concepts or skills. This explains the conversations I had with Caroline picking her up from school. Generally, when I asked her what she did at school that day, she would say "activities."
"Well what activities"
"Oh, just activities."
Later I was able to figure out that "activities" included something called "nuts and bolts" and something called "braiding." At the end of this year, Caroline came back with a "name tracing" book, in which she had traced her name a couple dozen times. We were told this is quite an accomplishment. This page emphasizes that the materials for the activities are self correcting. The child can perceive the problem without being corrected by the teacher, and thus is drawn into solving it on her own.
So will this work with a 12 year old learning algebra? Well, there was this fascinating study in Science which took advantage of a nice experiment in nature. A school district in Milwaukee servicing mostly poor and minority students had some Montessori schools and some regular schools. Since more people wanted to go to the Montessori schools than they could handle, students were admitted by lottery. So we have effectively a randomized trail. Students who applied to the Montessori schools but didn't get in are the control group; the students who did get in are the experimental group. Notice that both groups have already been selected to have parents that take an active interest in the kids education.
Results: the Montessori schools did better. They did better on academic skills, and a lot better on interpersonal skills. At age 5, children were told stories about kids behaving badly and were far more likely (43% versus 18%) to discuss these using concepts of justice and fairness. More fundamentally, the Montessori kids did better at tests of their concept of other minds. Eighty percent of the kids passed a test where they had to recognize that another person had a false belief. The kids from the other schools performed at chance level (50%).
But what about the older kids? Well for the 12 year olds studied, the academic advantage disappears. Kids in both groups performed similarly on basic skills. The Montessori kids, though still performed better on community and social skills, and also, interestingly, used more sophisticated sentence structures and creative stories.
The charts for the science article are below. Note, though the strong and somewhat misleading graphic rhetoric of the charts. They converted all the test scores to a single measure (z) where 0 is the statistical mean for each test, and then represented the scores using thick bars to represent deviation from the mean. Thus even if the Montessori kids are only scoring slightly better than the regular kids, they are still represented by a bright line going up rather than down.
Also note that the chart for the 12 year olds simply omits the data where the two groups were comparable.
The first graph has 16 data points, and the second has 8. Tufte would claim that you shouldn't even use a graph here, just a chart.
I'm going to have to return to this later. Here is a collection of links to empirical research.