If you are not reading some of the online topical threads on current education topics, you are missing an opportunity to follow the thoughts of an interesting variety of people who are interested in education. One group I follow is … Continue reading
Category Archives: Assessment and Data
Non-cognitive skills, Part II
Reply
More than just test scores
Continuing the theme of my last post, I want to reference Henry M. Levin’s paper “More than just test scores,” which can be accessed at http://roundtheinkwell.files.wordpress.com/2012/09/more-than-just-test-scores-sept2012-2.pdf. I’ve followed Dr. Levin’s work for several years and I’ve enjoyed his contributions to the knowledge base on assessment issues. In this brief, published online on August 30, 2012, Levin provides an outstanding and thorough discussion, supported by extensive research citations, on the relationship between test scores and earnings. He notes “…the measured relationships between test scores and earnings or productivity are modest and explain a relatively small share of the larger link between educational attainment and economic outcomes.”
Non-cognitive skills mean career readiness
Much of his paper discusses “the effects that education has on the development of interpersonal and intrapersonal skills and capabilities that affect the quality and productivity of the labour force.” It was Levin’s work here that prompted me to write this post and my last post on non-cognitive skills, starting with the SCANS work from 13 years ago, which offers substantial background to the reader for Levin’s research analysis. Just reading Levin’s abstract should be enough to keep you reading to the end of the paper, and then checking out some of the references.
Nowhere does Levin suggest that cognitive skills, as measure by test scores, are not important contributors to economic success later in life. But throughout the research cited one reads repeatedly that academic achievement is not the only measure of economic success. One must conclude that focusing 100% on improving test scores as the be-all and end-all for education is a narrow and likely misguided agenda. Coupled with the SCANS report, it’s clear that a focus on test scores and the corresponding narrowing of curriculum to cognitive issues only is not the best means to improve America’s economic competitiveness or the earnings potential of American youth.
Noting that the discussion of world-class educational standards, and comparisons of nations on various international tests leaves out “other attributes of school or student performances,…” Levin argues that “…basing world-class standards on measures of student achievement is a very limited approach to evaluating any educational system. To meet the economic, political, social, and personal demand for competency, much more is required of students and adults than just cognitive proficiencies as measured by test scores.” And he echoes the SCANS work when he suggests that interpersonal and intrapersonal skills are dimensions that can’t be ignored.
Test scores don’t explain earnings
By reviewing numerous studies on the connection between academic achievement and later earnings. Levin adds depth to the common understanding that more schooling equates to higher incomes. This is correct, but the assumption that the higher earnings are associated with cognitive skills is overstated. He identifies numerous studies that include test scores in their analysis of schooling and earnings, and they find that the “…impact of schooling on earnings retains more than 80% of its estimated effect on earnings, even after accounting for student achievement.” (See Bowles, Gintis, and Osborne, 2001, “The determinants of earnings: A behavioral approach,” Journal of Economic Literature, 39(4), 137-1176.) In plain English, test scores account for “a minority fraction” of the relationship between earnings and schooling, and other elements of schooling “must be accounted for.” Taking on the claim that test scores have predictive validity for worker productivity, Levin cites research that shows scores account for only 6% of the variance in productivity, “leaving 94% to be explained by factors other than the test result.”
So, it seems to me, the research isn’t particularly supportive of pushing exclusively for high academic achievement to return America to world-class productivity. And besides, the US worker is already among the most productive anywhere, even when our test performances appear mid-range.
Graduation rates do equate with economic performance
The relationship between finishing high school with a diploma and increased lifetime earnings is well-established. Levin offers some depth to this discussion by pointing to research on five educational interventions to prevent dropouts, which offer net benefits to taxpayers through public heath savings, crime savings, and increased exhale revenues from earnings. Interestingly, he also notes that GED graduates do not show the same benefits–their test scores are similar to graduates, but their earnings were inferior.
Having shown that cognitive skills and tests scores are not very good predictors of economic performance, but high school completion is, there is something other than academics of value for students who stay in school and complete a diploma other than compiling academic skills. And these factors, not currently tested, play as important a role, or a greater role, in helping students become ready for the world of work. I strongly suggest you read the rest of Levin’s article, in which he reviews what those characteristics are, and how schools can improve the way they are supported. In fact, Levin identifies several programs that teach the social and emotional skills, and cites research supporting the efficacy of doing so. He spends some time on the SEL movement–social and emotional learning–and goes beyond identifying the employment benefits by identifying research that shows SEL program intervention “…were comparable to or exceeded the results found in the literature for improving student achievement.” This is a really significant finding–“…interventions to develop social and emotional skills have particular salience for improving student achievement.”
I’ll end with a few sentences from Levin that should inspire you to read this piece several times, and read some of the back-up research he has so capably assembled. “The almost singular focus on test score performance in educational assessment at both domestic and international levels does rest on some foundation…. But the evidence does not support the assumption that cognitive skills are all that counts and that they alone can produce healthy and productive adult personalities.” “Far from being harmless, the focus on test scores and the omission of the non-cognitive impact of schools can create far-reaching damage.” “Other goals may be as, or more, important in the long run in terms of creating productive, equitable, and socially cohesive societies and economic growth.” (See Gradstein and Justman, 2002, “Education, social cohesion, and economic growth.” American Economic Review, 92)4_ 1192-1204.
As educators and activists for meaningful educational policy, we have to stop simply looking at easy solutions and start listening to what employers say they want. Listening to employers was a novel idea when the SCANs report came out in 2000, and it’s likely that those who publish achievement tests and those who push them as easy measures of future productivity will continue to ignore employers in the future. Real educators should speak up.
Teachers: The Courts Might Not Care About Bad Evaluation Systems
Courts and Teacher Evaluation Systems
Writing in Education Week, September 18, 2013, James Popham and Marguerita DeSander suggest that “thousands of American teachers will lose their jobs in the next few years because of the recently designed, more demand evaluation systems now found in most of our states.” They point out these systems have been triggered by recent federal policies, and then note that most teachers think the dismissals can be reversed in court. The main point of their essay is that the courts, both state and federal, have “historically refused to substitute their judgement for that of a school board in cases where a teacher (whether tenured or probationary) has been terminated for substandard job performance. This has been true if the termination is based on even a scintilla of evidence.”
Popham is a well-know figure among education professors who are experts on evaluation and assessment, with several books, most recently (2013) Evaluating America’s Teachers: Mission Possible? from Corwin Press. DeSander is an ed admin professor and formerly was an attorney specializing in employment law. These are two folks who should be seen as authoritative figures on this topic.
New York Teachers More Protected?
When I first saw the outlines of New York’s response to Race To The Top teacher evaluation changes, I thought that it might be a good time for college grads to go into the practice of education law, particularly if they had any statistical talent and were willing to delve into the psychometrics of testing and its use in teacher evaluation. As New York’s teacher evaluation system, including the use of test scores in teacher evaluation rolled out, I continued to believe the system would collapse under the weight of court challenges to it’s fairness and it’s validity. I still believe that, but I’m not quite so confident with my opinion after reading this piece. If I look at New York’s system, which has a unique additional element–teachers identified as low performing have to be given opportunities for professional development by their school districts–perhaps that district responsibility is going to prevent willy-nilly teacher dismissals. With the greater than anticipated increased costs of testing, and the Governor’s imposition of a 2% tax cap on raises in school budgets, it’s already estimated that 40% of districts across the state will be in bankruptcy in the next few years. So districts won’t have the money to provide professional development to their low performing staff, meaning they might not be able to dismiss them regardless of evaluations. That’s not the case in other states, where these warnings may be far more telling.
Evaluation Systems Show No Evidence of Efficacy
Popham and DeSander recognize that across the nation, the new evaluation systems have serious flaws–relying “too heavily of traditional achievement tests…” They note that the tests “are unaccompanied by any evidence that they are able to distinguish between well-taught and badly taught students.” They also point out that courts, based on past patterns of refusing to rule on the merits or validity of evaluation systems, “will not rule on the appropriateness of a teacher-evaluation system, or the evidence-collection procedures incorporated in that system, if the procedures are applied in a fair and consistent manner to all teachers affected. Thus, even an inadequate evaluation system will avoid the rigor of court scrutiny when it is applied equally to all teachers.” (Emphasis mine.)
The authors also take issue with the validity of classroom observations when they suggest that rating teacher classroom performance on the bases of “a handful of 30-minute classroom visits..” and even when looking at 40 or 50 dimensions of classroom performance, “the resultant observation data are often of little value.”
Courts May Not Overturn Bad Evaluation Systems
So where does that leave teachers all over the nation? I agree with the authors that the evaluation systems lack validity as they are being designed–the reliance on testing is not warranted psychometrically at all, nor are the tests designed for the purposes being implemented. And if courts are likely to disregard expert opinion because they won’t decide on system appropriateness, there is no venue for expert testimony about why these systems should be disregarded. If the legitimacy of procedures is not about to be adjudicated, there’s no place for teachers to turn. One need only review the multiple interpretations of the evaluation designs across 696 districts currently approved by the New York State Education department to know that there is no consistency in their format across the state. If courts give deference to school boards, teachers might be in far more trouble than they currently expect. Finally, a good civics education will remind all of us that courts are not always about right and wrong–they interpret the law. Where courts defer to school boards and states who have implemented bad evaluation systems, teachers beware.
Bruce Baker on Value Added and Student Growth Percentiles
I’ve lauded Bruce Baker’s blog once already, but day by day I’m becoming an even more devoted follower of his wit and his brilliant exposés of fallacious research, bogus claims, and shallow thinking. I’ve learned/confirmed so much by subscribing to School Finance 101, particularly about value added models and student growth percentile models, as Dr. Baker reexamines the research on these topics and highlights the shortcomings and the unintended consequences of using these measures in states all over the country.
I’m very interested in his work on student growth percentiles, first created by Damien Betebrenner at the National Center for the Improvement of Educational Assessment, www.nciea.org. First used in Colorado, and now spreading to Massachusetts, New Jersey and New York, the proponents of SGPs suggest they fix the problems inherent in value added measures. Baker emphatically says not so, and uses the writings of the fathers of SGPs to highlight how SGPs are being misused.
To get up to speed on VAMs and SGPs and their problems as Baker sees them, go through the blog references found in his Value Added Teacher Evaluation category here, and as you read the articles, click on through to the links within each one. You’ll get quite a wonderful look at the limitations of these statistical models as they relate to teacher evaluation. Baker also takes on state education officials in several states, including Colorado, New Jersey and New York, with his irrevent commentary.
Of particular interest for New Yorkers is his analysis of the preliminary technical reports on the results of NY’s first SGP assessment results. This analysis can be seen in the entry entitled “AIR Pollution in NY State…” He offers graphs showing how factors that cannot be attributed to teachers have an effect on the patterns of SGP scores–students in low income schools generally underperform students in higher income schools, as one already knows, but the SGPs of these students also lag behind higher income schools. This demonstrates that SGPs do not, in fact, account for effects of peer groups, of school effects, or of poverty, as some suggest.
Read Baker and weep about how assessments are regularly being misused by people who should know better,