The Center for Public School Renewal
NOTE: Published in slightly different form by Mathematics In Michigan, Summer, 1995.
MEAP: Mathematics and the Reading Connection
Most mathematics teachers who have looked at the post-1990 MEAP mathematics tests are aware that there is considerably greater emphasis on reading mathematics problems than there was on the pre-1991 state tests. In fact, it is possible to use statistical techniques to assess the degree of the relationship between students' abilities in these two areas. The first analysis reported below, although not based on random sampling techniques, does represent more than half of the students in the state.
The scattergram below shows the relationship between average 1994 mathematics and reading scores for 139 districts, mostly in southeastern Michigan(1). As you can see, the graphs show very high statistically significant (p = .001) correlations between the two tests. If we assume that success on the mathematics tests depends to a degree on a student's reading ability, we can see that upwards of 71% of the variation in mathematics scores depends on reading ability as assessed by the reading test.
The Relationship Between Mathematics and Reading
Since the data is not randomly sampled we can't make as definitive a statement as we might like about the relation between reading skill and mathematics achievement. However, it does appear that some relationship exists. In order to probe this relationship further, I created a similar analysis for the 1994 4th grade test results with elementary schools in my own district.
Some problems exist with these data. First, with only 32 elementary schools in the district, the size of the sample is almost too small for standard statistical techniques. Also, there is the problem of non-randomness. In any event, the school-level data for these 32 schools do bear a resemblance to the district level data, although the correlation between reading and mathematics (r = .75, p = .001) is not quite as high. Presumably, a random sample of schools from around the state would greatly clarify this situation.
It was also possible for me to analyze individual student data in my district. See the table below.
In the table above "r/S" stands for the correlation between the scaled scores for mathematics and the story selection portion of the reading test. Similarly, "r/I" stands for the correlation of mathematics with the scaled score on the information selection portion of the reading test. These correlations are statistically significant at the p = .001 level. Notice that the "r/I" correlations show the stronger relationship, which conforms to the expectation that reading for information is the kind of reading one does on a mathematics test.
The data reported here, while not conclusive, do indicate a fairly reliable relationship between students reading ability and their achievement on the MEAP mathematics tests. A statewide sampling of MEAP data, perhaps one that would permit aggregating the same universe of students to different levels of aggregation, would be a worthwhile project for the mathematics education community to pursue. A larger study would also allow for the control of other variables, such as district size, level of poverty, and per-pupil expenditures, which may be related to achievement. Even without a more rigorous study it seems clear that we need to become more aware of the importance of reading to students' success with the type of mathematics fostered by Michigan's Essential Goals and Objectives, the NCTM Standards, and other new directions that are being developed.
printed in The Detroit Free Press, The Flint Journal and a Middle
Cities Education Association report. If resources become available a random sample of all
Michigan districts should also be analyzed to see if this relation holds up. These 139 districts
represent 58% of state students.