As I wrote yesterday, it wasn’t long after the Bluegrass Institute released our new report on “Blacks Continue Falling Through the Gaps in Louisville’s Schools, the 2016 Update” before the Jefferson County Public School District (JCPS) honored us with a response. In fact, the JCPS Response came out only a few hours after our report release news conference ended last Monday.
That JCPS Response did get us thinking. That response focused on KPREP score trends, which we avoided discussing in our report because recent evidence assembled at the institute reveals KPREP is showing signs of score inflation (see here, here, here, here and here).
Still, since the JCPS brought the KPREP subject up, we decided to take a look. Yesterday’s blog shows that even if we look at KPREP, the white minus black achievement gaps for most subjects and most school levels in Jefferson County deteriorated between 2011-12 and 2014-15.
I also realized we could look at something else, namely the differences in the trends between KPREP eighth grade math and reading and the EXPLORE tests.
To give you some perspective, the EXPLORE was created by the ACT, Inc., the same people who create the ACT college entrance test, to show if students are on track to be ready for college as of the eighth grade. EXPLORE is basically a pre-ACT test with Benchmark Scores for readiness that indicate students are indeed on track.
Since Common Core is supposed to be all about college and career readiness, EXPLORE should provide a useful indication if the promise really is being met.
Table 1 reports what the EXPLORE test shows us about Jefferson County white and black student math and reading performance between 2011-12 and 2014-15. Note: sharp readers who have looked at our blacks falling through gaps 2016 update report will note that Table 1 below is an extract from Table 1 in that report.
On EXPLORE, not only did the JCPS white minus black achievement gaps grow in both math and reading between 2011-12 and 2014-15, but the actual scores also declined for both races, as shown in the two right-most columns. The gaps, shown in the middle section of the table, grew because black scores dropped even more than white scores did.
Table 2 shows how the KPREP results look.
In all but one case, the KPREP score changes for whites and blacks, shown in the two far right columns, increased.
When we compare Table 1 to Table 2:
- For math for whites, EXPLORE dropped 0.4 points while KPREP says proficiency improved by 1.4 points. Scores from KPREP and EXPLORE went in opposite directions.
- For math for blacks, scores dropped on both tests, but the drop on EXPLORE was more than double the drop on KPREP.
- For reading for whites, EXPLORE dropped by 3.5 points while KPREP says reading improved by 7.1 points. The EXPLORE and KPREP scores not only moved in opposite directions, but they moved rather strongly in those opposite directions.
- For reading for blacks, EXPLORE dropped 3.8 points while KPREP says reading for blacks improved by 6.5 points. Just like the white reading case, the strongly different trends on EXPLORE versus KPREP are very problematic.
So, there are several messages here.
- When we are talking about the trends in the achievement gaps, whether we look at KPREP or the better established EXPLORE results, those white minus black achievement gaps increased in JCPS schools.
- When we are talking about possible inflation in the KPREP, especially in reading, the results shown above are highly disturbing. Not only do the trends generally move in opposite directions, but they do so rather strongly for reading.
Bottom line: my initial hesitancy to examine KPREP trends over time has been substantially supported, at least for the case of middle school students where an alternative test with a much longer track record is available to evaluate what KPREP is telling us. I continue to be very dubious of the validity of trend information from the obviously problematic KPREP scores.
Furthermore, I continue to be very concerned about the obvious adverse trends in achievement gaps, and even actual scores, in Louisville’s schools. Where black gains seem to be present, they are found on tests of questionable validity and are almost universally smaller than white gains. Thus, JCPS blacks generally trail farther behind today than they did four years ago.