Thursday, March 19, 2009

College Performance Points to Problems in Kentucky’s P to 12 Schools

Over at the Prichard Committee’s Blog, Susan Weston recently pointed at a really interesting “Bang for the Buck” graph that depicts the efficiency of turning out college degrees in Kentucky’s postsecondary institutions. This graph comes from a Kentucky Chamber of Commerce report that discusses a number of ways that the state’s public college system could improve its efficiency. The graph shows that college degree production in Kentucky is very inefficient compared to the programs in other states.

While her discussion touches on some things that the colleges can directly control to improve their efficiency, Ms. Weston omits any mention of the possibility that Kentucky’s relatively high college costs and comparatively low college degree output may be due in significant measure to the poorly educated product coming from our public school system.

This is an important issue, so the Bluegrass Institute decided to dig a bit deeper. To see what we found, read on.


Susan Weston said...

I agree that improved P-12 performance would lead to higher college performance. That seems likely to be true in every state.

Your argument sounds like you're claiming that unusual weakness in our P-12 explains unusual weakness in higher education productivity.

We're unusually weak in P-12 mathematics.

Got anything else?

Richard Innes said...


Of course I do.

For our other readers, your comments about Kentucky being weak in math refer to results from the National Assessment of Educational Progress, or NAEP. Let’s discuss NAEP for other subjects besides math.

Unfortunately, as I have to point out every time Susan and I discuss the NAEP, the analysis of scores from this assessment have gotten very confused recently. That is because of three factors that now vary widely from state to state. Those factors are clearly mentioned in all the latest NAEP reports and include:

(1) Different rates of exclusion of students with learning disabilities (SD) and English Language Learners (ELL),

(2) Different rates of offering testing accommodations to SD and ELL that do get to take the NAEP, and

(3) Different student demographics.

For a brief answer to your discussion, I’ll just deal with the performance of Kentucky’s overwhelming demographic element – White students – because they comprise the vast majority of students in Kentucky, both in public schools and colleges.

Since we are talking about college, I’ll stick to the eighth grade NAEP, as this is the highest grade level where state data is available.

The comments below are based on scores and percentages accessed from the NAEP Data Explorer on March 20, 2009.

READING: Let’s start with reading, an area where overall NAEP scores indicate Kentucky is roughly tied with the nation. We’ll see how the Kentucky picture changes dramatically after we look at NAEP scores broken out by demographic makeup.

The national NAEP student demographics in reading have shifted dramatically since NAEP started to test states in the subject in 1992.

In 2007, the national NAEP sample for Grade 8 reading was only 58 percent White. This was a significant drop from the national 71 percent figure found across the nation back in 1992 (Note: NAEP did perform a national Grade 8 reading assessment in 1992 but did not do a state level reading assessment that year).

In general, on all NAEP assessments Whites notably outscore all other minority groups reported except Asian/Pacific Islanders (who have always comprised a fairly small part of the NAEP samples and don’t exceed the national average by very much because high scoring Asians from certain countries are offset by much lower scoring Asians from other countries and Pacific Islanders).

Thus, the significant reduction in the proportion of Whites in the national sample and in many other states has acted to hold down overall average scores elsewhere to a very notable extent. The effect even has a related name, “Simpson’s Paradox,” because it is so well known in testing circles.

Kentucky is one of the “Whitest” states in the country. Our NAEP Grade 8 Reading student sample in 2007 was 84 percent White. That hasn’t changed much over the years. Kentucky’s kids are mostly native born, and White. It has consistently been that way since NAEP started testing at the state level in 1990. So, Kentucky does not suffer much from Simpson’s Paradox. Unlike other states, our demographics have not strongly acted to hold back our overall average performance on the NAEP.

So, it is important to look at NAEP data disaggregated by race. As soon as we do that, we find that on Grade 8 Reading, only four states had lower White NAEP Grade 8 Reading Scale Scores in 2007 than Kentucky did. In fact, if statistical sampling error were considered, we probably outscored Whites in only one or two other states. That’s all.

SCIENCE: Doing a similar examination for NAEP Grade 8 Science, where you like to point out that overall NAEP scores show us in about the middle of the pack, in 2005 our Whites only outscored Whites in eight other states. Again, if statistical sampling error were considered, our Whites probably outscored only about five or fewer states.

Whites in Kentucky comprised 87 percent of that 2005 NAEP science sample, while the national average sample was only 60 percent White. That difference explains the big shift in Kentucky’s rankings when you change from looking at overall scores to only those scores for our dominant demographic group.

WRITING: Looking at the 2007 NAEP Grade 8 writing results, Kentucky only outscored four other states, probably only one by a statistically significant amount.

Now, let’s see how that might impact our colleges.

Because Kentucky’s student demographics are overwhelmingly White, that has an impact on the demographics of the colleges in Kentucky, as well.

So, as a normal reflection of the makeup of Kentucky’s overall demographics, the overwhelming portion of our college student body is White. However, Kentucky’s college student body is overwhelmingly composed of students who score much lower than their peers on NAEP.

So, when you try to make a case that Kentucky only has troubles in math on NAEP, you are overlooking some pretty stark realities from Simpson’s Paradox. In fact, Kentucky’s public education has a lot of problems, and those problems definitely impact our university system. But, you have to look deeper than simple analysis of overall NAEP scores to learn that.

Returning to the point in the main blog, it is fair to say that some of the efficiency problems our colleges face are due to weak products from our high schools. Now, thanks to your question, we can add the NAEP evidence above to the earlier material you pointed out to support that position.

Susan Weston said...

Kentucky students, as a group, treating them all as simply human, are in the middle of the national pack in reading, writing, and science.

What I hear you saying is that those results don't count, because our kids should be ahead of the nation, and the reason they should be ahead is that so many of our kids are white. I hope I'm misunderstanding.

Why does having a large white population mean we should score above national average?

Richard Innes said...

Susan, you have this really wrong.

For one thing, it isn’t just me who cautions about doing the sort of simplistic NAEP comparisons you seem to prefer.

The people who create the NAEP say right up front in their report cards that such comparisons have to consider things like demographics and exclusions. Look at page 7 in any of the 2007 NAEP report cards.

In fact, I got the suggestion to look at disaggregated NAEP scores from a staff member at the National Assessment Governing Board. THEY tell us to do this because of known problems that can be hidden when you only look at overall scores.

You just can’t pretend that racial gaps don’t exist and that they don’t significantly impact any state to state NAEP comparisons. That amounts to saying that gaps don’t matter, and I know that isn’t the way you really think.

How can Kentucky comfortably claim it looks good when that impression is largely due to having a lot more White kids who, thanks to a number of historic advantages, score better on NAEP than kids from other racial groups?

Do you really think Kentucky can just lay back and say all is OK because we outscore a state like California? That state’s poverty rate now matches us thanks to the huge influx of non-English speaking Hispanics. Those low scoring Hispanics now form the majority in California’s public schools – the majority. Do you really think that gives us a comfortable comparison?

California is far from alone. While Kentucky has maintained an extremely high percentage of Whites in its student cohort, the rest of the country has experienced a heavy influx of lower scoring students – many Hispanic – a significant number of which have not even been in this country for a long time and don’t even speak English as their native language.

In sharp comparison, we have so few Hispanics in Kentucky that the NAEP generally won’t even report scores for them. The samples are just too small.

To sum up, are you really willing to totally ignore all the historic educational disadvantages that non-Whites have experienced across the United States just to claim victory for Kentucky? The high remediation rates in our colleges say that is a dangerous, and ill-informed, position.

Returning to the central theme in the main Blog, P to 12 education is an important part of our college efficiency problem, whether or not you want to see it.

One other item, as a point of correction, Kentucky did not score in the middle of the nation in the latest eighth grade writing results, not even for all students taken together. In fact, we only statistically significantly outscored five other states for proficiency rates in the latest NAEP Grade 8 writing results. We only had a statistically significantly higher NAEP scale score than eight states. That was a drop from statistically significantly outscoring 11 states’ scale scores in 2002. Kentucky’s most recent writing performance isn’t in the middle of the nation.

Oh, by the way, California – that state where lower scoring Hispanics are now the majority and where one in five kids who took NAEP writing in 2007 was still officially still learning English – California got a statistical tie with us for NAEP Grade 8 writing proficiency in 2007. And, the writing assessment is only given in English. Doesn’t that tell you something is wrong?

Anonymous said...


10602 movies of premium DVD dignity

We secure thousands of the hottest movies you've been waiting to supervise! And they're stylish available in Hi-Def!
These movies are playable on most viewing devices including iPod, PDA (HandHelds), PC, DVD & DivX players.
There are categorically no limits. Download as much as you want with incredibly gamy speeds. No additional software is required!
Only click on a tie, download a flick picture show and guard it on your favorite player.

[size=4][url=]Sign Up[/url] today and forget about buying or renting DVDs offline. [/size]


[url=][b]TOP 40 BEST Movies:[/b][/url]
Halloween: 25 Years of Terror
Strange Invaders
Fearless Vampire Killers, The
Plumm Summer, A
Oh, God! You Devil
Horsemen, The
Moulin Rouge!
Romancing the Stone
Instinct to Kill
Anchorman: The Legend of Ron Burgundy
Home Alone
Godzilla vs. Ghidorah the Three Headed Monster
Lucky Numbers
Into the Sun
Ralphie May: Austin-Tatious
Four Horsemen, The
Death Race
Bone Collector, The
Boogeymen: The Killer Compilation
Lodge, The
Momma's Man
Analyze This
Goonies, The
Kate & Leopold
Children Shouldn't Play with Dead Things

[url=]All Roads Lead Home [/url]
[url=]Going in Style [/url]
[url=]Stay [/url]
[url=]Lost in Translation [/url]
[url=]Husbands and Lovers [/url]
[url=]Sin City [/url]
[url=]Powder [/url]
[url=]Once Upon a Time in Mexico [/url]
[url=]Bad Boys [/url]
[url=]Maximum Overdrive [/url]