Do Higher Science Standards Lead to Higher Achievement?

In a recent article in Scientific American, it was suggested that the U.S. should adopt higher standards in science, and that all 50 states should adopt them.

When you check the literature on science standards, the main reason for aiming for higher standards (raising the bar) is because in the “Olympics” of international academic test taking, the U.S. never takes home the gold.  In fact, according the tests results reported by the Program for International Student Assessment (PISA), U.S. students never score high enough to even merit a bronze medal.  In the last PISA Science Olympics, Shanghai-China (population 23 million) took home the Gold, Finland (population 5.4 million) the Silver, and Hong Kong-China (population 7 million, the Bronze.  The United States (population 314 million) average score positioned them 22nd on the leaderboard of 65 countries that participated in the PISA 2009 testing.

Some would argue that comparing scores across countries that vary so much in population, ethnic groups, poverty, health care, and housing is not a valid enterprise.  We’ll take that into consideration as we explore the relationship of standards to student achievement.

Its assumed that there is a connection or correlation between the quality of the standards in a particular discipline such as science, and the achievement levels of students as measured by tests.  So the argument is promoted that because U.S. students score near the bottom of the top third of countries that took the PISA test in 2009, then the U.S. science education standards need to be ramped up.  If we ramp up the standards, that is to say, make them more rigorous and at a higher level, then we should see a movement upwards for U.S. students on future PISA tests.  It seems like a reasonable assumption, and one that has driven the U.S. education system toward a single set of standards in mathematics and reading/language arts (Common Core State Standards-CCSS), and very soon, there will be a single set of science standards.

There is a real problem here

There is no research to support the contention that higher standards mean higher student achievement.  In fact there are very few facts to show that standards make a difference in student achievement.  It could be that standards, per se, act as barriers to learning, not bridges to the world of science.

Barriers to Learning

I’ve reported on this blog research published in the Journal of Research in Science Teaching by professor Carolyn Wallace of Indiana State University that indicates that the science standards in Georgia actually present barriers to teaching and learning. Wallace analyzed the effects of authoritarian standards language on science  classroom teaching.  She argues that curriculum standards based on a content and product model of education are “incongruent” with research in science education, cognitive psychology, language use, and science as inquiry.  The Next Generation Science Standards is based on a content and product model of teaching, and in fact, has not deviated from the earlier National Science Education Standards.

Over the past three decades, researchers from around the world have shown that students prior knowledge and the context of how science is learned are significant factors in helping students learn science.  Instead of starting with the prior experiences and interests of students, the standards are used to determine what students learn.  Even the standards in the NGSS, or the CCSS are lists of objectives defining a body of knowledge to be learned by all learners.  As Wallace shows, its the individuals in charge of curriculum (read standards) that determine the lists of standards to be learned. Science content to be learned exists without a context, and without any knowledge of the students who are required to master this stuff, and teachers who plan and carry out the instruction.

An important point that Wallace highlights is that teachers (and students) are recipients of the standards, rather than having been a part of the process in creating the standards. By and large teachers are nonparticipants in the design and writing of standards. But more importantly, teachers were not part of the decision to use standards to drive school science, in the first place. That was done by élite groups of scientists, consultants, and educators.

The Brown Center Report

According to the 2012 Brown Center Report on American Education, the Common Core State Standards will have little to no effect on student achievement. Author Tom Loveless explains that neither the quality nor the rigor of state standards is related to state NAEP scores. Loveless suggests that if there was an effect, we would have seen it since all states had standards since 2003.

For example in the Brown Center study, it was reported (in a separate 2009 study by Whitehurst), that there was no correlation of NAEP scores with the quality ratings of state standards. Whitehurst studied scores from 2000 to 2007, and found that NAEP scores did not depend upon the “quality of the standards,” and he reported that this was true for both white and black students (The Brown Center Report on American Education, p.9). The correlation coefficients ranged from -0.6 to 0.08.

The higher a “cut score” that a state established for difficulty of performance can be used to define the rigor or expectations of standards. One would expect that over time, achievement scores in states that have more rigorous and higher expectations, would trend upwards. The Brown study reported it this way:

States with higher, more rigorous cut points did not have stronger NAEP scores than states with less rigorous cut points.

The researchers found that it did not matter if states raised the bar, or lowered the bar on NAEP scores. The only positive and significant correlations reported between raising and lowering the bar were in 4th grade math and reading. One can not decide causality using simple correlations, but we can say there is some relationship here.

When researchers looked at facts to find out if standardization would cut the variation of scores between states, they found that the variation was relatively small compared to looking at the variation within states. The researchers put it this way (The Brown Center Report on American Education, p. 12): The findings are clear.

Most variation on NAEP occurs within states not between them. The variation within states is four to five times larger than the variation between states.

According to the Brown Report, the Common Core will have very little impact on national achievement (Brown Report, p. 12).  There is no reason to believe that won’t be true for science.

The researchers concluded that we should not expect much from the Common Core. In an interesting discussion of the implications of their findings, Tom Loveless, the author of the report, cautions us to be careful about not being drawn into thinking that standards represent a kind of system of “weights and measures.” Loveless tells us that standards’ reformers use the word—benchmarks—as a synonym for standards. And he says that they use it too often. In science education, we’ve had a long history of using the word benchmarks, and Loveless reminds us that there are not real, or measured benchmarks in any content area. Yet, when you read the standards—common core or science—there is the implication we really know–almost in a measured way–what standards should be met at a particular grade level.

Loveless also makes a strong point when he says the entire system of education is “teeming with variation.” To think that creating a set of common core standards will cut this variation between states or within a state simply will not succeed. As he puts it, the common core (a kind of intended curriculum) sits on top of the implemented and achieved curriculum. The implemented curriculum is what teachers do with their students day-to-day. It is full of variation within a school. Two biology teachers in the same school will get very different results for a lot of different factors. But as far as the state is concerned, the achieved curriculum is all that matters. The state uses high-stakes tests to decide whether schools met Adequate Yearly Progress (AYP).

Now What?

If standards do not result in improved learning as measured by achievement tests, what should we be doing to improve schools?

Over on Anthony Cody’s blog on Education Week, we might find some answers to this question.  Cody has begun a series of dialogs with the Gates Foundation on educational reform by bringing together discussions between opposing views to uncover some common ground. Cody has already broken new ground because the Gates Foundation is not only participating with him on his website, but Gates is publishing everything on their own site: Impatient Optimists blog. Three of the five dialog posts have been written, and it is the third one written by Anthony Cody that I want to bring in here.

In his post, Can Schools Defeat Poverty by Ignoring it?, Cody reminds us that the U.S. Department of Education (through the Race to the Top and NCLB Flexibility Requests) is unwavering in its promotion of data-driven education, using student test scores to rate and evaluate teachers and administrators.  Cody believes that the Gates Foundation has used its political influence to support this.  There is also an alliance between the ED, and PARCC which is developing assessments to be aligned to the Common Core Standards.  The Gates Foundation is a financial contributor to Achieve, which oversees the Common Core State Standards, the Next Generation Science Standards, and PARCC.

There is a “no excuses” attitude suggesting that students from impoverished backgrounds should do just as well as students from enriched communities.  The idea here is that teachers make the difference in student learning, and if this is true, then it is the “quality” of the teacher that will decide whether students do well on academic tests.

Anthony Cody says this is a huge error.  In his post, he says, and later in the post uses research to tell us:

In the US, the linchpin for education is not teacher effectiveness or data-driven management systems. It is the effects of poverty and racial isolation on our children.

As he points out, teachers account for only 20% of the variance in student test scores.  More than 60% of score variance on achievement tests correlates to out-of-school factors.  Out-of-school factors vary a great deal.  However, as Cody points out, the impact of violence, health, housing, and child development in poverty are factors that far out weigh the effect of teacher on a test given in the spring to students whose attendance is attendance, interest, and acceptance is poor.

In the Scientific American article I referenced at the beginning of this post, the author cites research from the Fordham Foundation that scores most state science standards as poor to mediocre.  We debunked the Fordham “research” here, and showed that its research method was unreliable, and invalid.  Unfortunately, various groups, even Scientific American, accept Fordham’s findings, and use in articles and papers as if it a valid assessment of science education standards.  It is not.

It’s not that we don’t have adequate science standards.  It’s that if we ignore the most important and significant factors that affect the life of students in and out of school, then standards of any quality won’t make a difference.

What is your view on the effect of changing the science standards on student achievement.  Are we heading in the wrong direction?  If so, which way should we go?