The Southern Educational Foundation just issued “A New Diverse Majority,” a report on demographic changes in the public school student body.
It has a surprise.
Kentucky isn’t as “poor” as our educators might like us to believe when they make excuses for performance.
This graphical map from the Southern Educational Foundation shows the percentage of the students in each state that qualify for the federal free and reduced cost school lunch program. The numbers are a commonly used indicator of poverty for school children.
Plenty of states on the map have higher poverty rates than we have. And, because the break point used to color code the map is a 50 percent rate, and because we barely exceed that, several of the states shown in green are actually within a few points of our poverty rate, as well.
So, when white students in every state on this map except Oklahoma (poorer), Tennessee (poorer), Hawaii, Alabama (poorer), Mississippi (MUCH poorer) and West Virginia (equally poor) outscored our whites on the 2009 National Assessment of Educational Progress Grade 8 Math Assessment, and when of that group of six states, Oklahoma (poorer), Tennessee (poorer) and Hawaii actually were able to tie us, we have just cause to worry.
The poverty excuse won’t fly any more in Kentucky. Almost every state in the South plus California and West Virginia have more poverty now than we do, but in most cases their kids now outperform our dominant racial group.
(Note: I ranked the NAEP scores as downloaded from the NAEP Data Explorer web tool)
Sunday, January 17, 2010
Is Kentucky’s poverty rate really exceptional?
Subscribe to:
Post Comments (Atom)
12 comments:
When one looks in the NAEP Data Explorer at the percent of students on the NAEP 2009 mathematics test at grade 4 who qualified for the National School Lunch Program (the NAEP poverty indicator), one notices that the percentages of students nationwide and of 11 states were not significantly different from Kentucky (i.e., CA, AZ, KS, OK, TN, AL, GA, FL, NC, SC, and NY). Six states (i.e., NM, TX, AR, LA, MS, and WV) and the District of Coumbia had significantly higher percentages on the poverty indicator than Kentucky.
Anonymous Jan 19 at 2010's comment actually helps point out a notable shortcoming in the National Assessment of Educational Progress (NAEP).
The results from NAEP cannot be as accurate as the data referenced on the map because the NAEP relies on a sampling system which inevitably introduces some notable sampling errors.
In contrast, the numbers on the map in the main blog item are based on the total headcount of students in the federal and reduced cost lunch program and total enrollment and therefore have no notable plus or minus sampling errors to blur the results.
NAEP simply can't show this level of detail. However, the data used here can be accepted as posted with high confidence that errors are small.
Thus, though you cannot show it with the NAEP, Kentucky has lower poverty numbers than a lot more than six states. You just have to use better data to show that.
I fear that you place to much confidence in headcounts for the school lunch program as a "true" indicator of poverty. The count is pretty good in the elementary grades, starts to fall apart in the middle grades, and becomes totally unreliable in the high school grades. School lunch counts, whether actual counts or statistical estimates have limited use as poverty indicators. I understand that NAEP is studying the use of US census codes to obtain poverty estimates based on the students' block and tract family income data. It works well for business studies; maybe it will work for NAEP.
Just for fun, I downloaded the USDA data that the Southern Educational Foundation cited (i.e., National School Lunch Program -- State-level Data, FFY 1989 to FFY 2008). One worksheet in the Excel file ranked states by the percentage of students in the school lunch program statewide at all grades combined. In 2008, Kentucky ranked 14th in the country in the percentage of students in the school lunch headcount. Here's the list (for what it's worth):
Mississippi (1)
Dist of Columbia (2)
Louisiana (3)
New Mexico (4)
Texas (5)
Oklahoma (6)
Arkansas (7)
Georgia (8)
Tennessee (9)
South Carolina (10)
California (11)
Alabama (12)
West Virginia (13)
KENTUCKY (14)
North Carolina (15)
New York (16)
Florida (17)
Illinois (18)
Arizona (19)
Oregon (20)
Idaho (21)
Missouri (22)
Alaska (23)
Nevada (24)
Nebraska (25)
Delaware (26)
Kansas (27)
Rhode Island (28)
Hawaii (29)
Washington (30)
Maine (31)
Indiana (32)
Michigan (33)
South Dakota (34)
Montana (35)
Pennsylvania (36)
Ohio (37)
Colorado (38)
Maryland (39)
Iowa (40)
New Jersey (41)
Wisconsin (42)
Virginia (43)
Minnesota (44)
Wyoming (45)
Utah (46)
Vermont (47)
Massachusetts (48)
Connecticut (49)
North Dakota (50)
New Hampshire (51)
RE: Anonymous Jan 20 at 1156 AM
Hmmm. I guess you also posted the earlier Anonymous comments, as well. So, lets see, you started out wanting to admit only six states had more poverty that Kentucky. Now, you point out that 13 states have more. That's a fairly big difference, don't you think?
Also, the comments about free/reduced lunch not being a great proxy are valid, but this is generally the best data currently available.
I have never said the quality of research data in education is good, far from it. I just try to use the best available.
Also, consider how some of those other states performed against us in the 2009 NAEP Grade 8 Math Assessment. Some did worse, but Texas and South Carolina outscored us for overall proficient or better score results and Arkansas, Georgia tied us.
When we look at disaggregated data just for whites (Kentucky hardly has any minorities except blacks), New Mexico, Texas, Arkansas, Georgia, Tennessee, South Carolina and California, outscored us for proficiency rates and Lousiana, Oklahoma and Alabama tied us. All of those states have more poverty, and they also have much higher proportions of minority students, as well. We should outscore them, but we don't.
So, forget the poverty excuse, Anonymous. It doesn't work for Kentucky anymore.
I admit freely that NAEP statistically significant estimates of poverty among 4th graders in all states in 2009 show that only six states have a higher percentage of poverty students than Kentucky. This NAEP statistic is more useful than a percentage based on the combined 2008 USDA K-12 headcounts for conducting analyses of 4th grade performance. The same may be said, but less so, for 8th grade analyses. Neither source can support useful 12th grade analyses.
"Six" is indeed different from "13" because, obviously, grade 4 data alone are different from K-12 data combined, expecially when the grade 4 data and the K-12 data come from different sources in different years.
Anonymous 20 Jan at 2:52 PM writes,
"This NAEP statistic is more useful than a percentage based on the combined 2008 USDA K-12 headcounts for conducting analyses of 4th grade performance."
NAEP does not report a "poverty indicator" per se. What people use as a poverty indicator in NAEP is the same school lunch eligibility data that goes to the US Department of Agriculture.
Unless there is a differential in poverty by grade level from state to state (unlikely), whether we use the K to 12 numbers or specific numbers for just one grade won't matter. It's all from the same data.
But, the data is more completely reported to the USDOE, so we can more accurately rank the states.
It looks like the number is 13 states, not six, more or less like the report mentioned in the main blog generally indicates.
Before students are sampled, NAEP obtains individual school lunch status data for all students in the sampled schools, whether grade 4, 8 or 12 (not the general K-12 counts submitted to USDA). After the student sample for the state has been selected, NAEP conducts quality assurance reviews with the states to determine whether the percentage of school lunch eligible students in the NAEP student sample adequately represents the percentage of school lunch eligible students in the state at that grade level.
This attention toward grade level estimates for poverty supports a claim that NAEP's estimates constitute the best available information about poverty in the states to conduct analyses at grades 4 and 8, and even 12.
For NAEP reading 2005 (the most recent year in which national NSLP eligible (i.e., poverty) percentages were available for grades 4, 8 and 12), the 4th grade poverty rate was 45%, the 8th grade poverty rate was 39%, and the 12th grade poverty rate was 25%. In that same year NAEP's "Large City" rates were also available for grade 4 (71%) and grade 8 (63%), but not for gradae 12.
For sure, each of these grade level poverty rates is different from a national K-12 poverty rate. Caution is advised when making claims about achievement when controlling for poverty.
RE: Anonymous 21 Jan at 10:54 AM
The NAEP statistics all have sampling errors. That introduces some plus/minus errors in the numbers that cannot be overcome.
When the entire community's statistics are compiled, as the federal lunch statistics do, there are no sampling errors.
I just ran a quick check with the NAEP Data Explorer for the sampling error in Kentucky's percent of grade 4 students in the 2009 NAEP Math assessment that were in the free/reduced lunch program. The standard error was 1.7. That means that while the published rate of students in the lunch program was 51%, the real rate could range (with a 95% level of confidence) anywhere from 47.6 percent to 54.4 percent - an indeterminancy span of nearly 7 points.
That isn't a really precise set of data, is it.
Kentucky would slide up or down a number of places in the ranking if its actual rate varied by that much.
We don't have that sort of error when all students are counted, as was done with the mapped data in the base blog.
Now, let's stop wasting time with the poverty excuse. If we dwell on that, our kids will never get the educations they need.
the only state out of the south that even comes close to kentucky's poverty is new york.
meanwhile, alabama, gerogia, south carolina and tennessee are within 2 percentage points.
you dont think poverty comes into play? take a basic sociology class
RE: Anonymous on February 6, 2011 11:53 PM
You need to click on the map to enlarge it.
The USDA-based map shows states outside of the South that have similar or even larger poverty rates to Kentucky's (based on school lunch data) also include: California, New Mexico, aside from New York. Two other states, Arizona and Illinois, have rates within five points of Kentucky's, as well.
Also, when we look at how Kentucky's dominant racial group, white students, who also qualify for the federal lunch program compare to their counterparts in other states on NAEP Grade 8 Math (see that map here: http://www.freedomkentucky.org/index.php?title=The_National_Assessment_of_Educational_Progress#Maps_for_Students_in_Poverty), we learn that only one state in the country scored statistically significantly lower -- West Virginia.
Furthermore, poor whites in Maryland, North Carolina, South Carolina and Texas statistically significantly outscored our poor whites, as well.
So, no sir, while poverty certainly is a player in poor school performance, it is not an excuse. It is possible to do better than we are doing in Kentucky.
Post a Comment