Teenagers Are More Likely to Falsely Think That Crime Pays

That’s according to a new study from Penn’s Elizabeth Shulman:

To what extent is criminal behavior in adolescence attributable to risk appraisal? Using two large cross-sectional samples (N = 929, age range: 10–30 years; and N = 1,357, age range: 12–24 years), we examine whether (a) reward bias in risk appraisal is more prominent in adolescence and (b) the association between risk appraisal and criminal behavior is stronger during adolescence than at other ages. In Study 1, criminal behavior was self-reported; in Study 2, it was defined by involvement with the court. Perceived chances of a negative outcome, seriousness of consequences, and benefits versus costs of various risky activities were assessed to gauge reward bias in risk appraisal. The findings indicate that reward bias is elevated during the adolescence years. Also, risk appraisal bears a stronger relation to self-reported crime in middle adolescence and to official law-breaking behavior in early adolescence than at other ages. The findings are consistent with a dual-systems model of adolescent development and align with recent U.S. Supreme Court decisions addressing juvenile offenders’ culpability.

In addition to being bad at understanding the risk involved in committing a crime, teenagers are also bad at predicting their risk of dropping out of school or getting pregnant. It’s all the more reason that we need better ways of teaching behavior, whether it’s through schools, communities, or technology.


The Trouble With Politicians Who Always Talk About Values

Last spring Paul Ryan spoke about the latest version of his budget, promising it “offers a better path, consistent with the timeless principles of our nation’s founding and, frankly, consistent with how I understand my Catholic faith.” According to Ryan, “We put our trust in people, not in government.” There was nothing new or surprising in Ryan’s words. At that point his language was already run-of-the-mill Republican rhetoric about the fight over government spending. What is notable about Ryan’s message is that its careful framing may help explain some of the gridlock plaguing the American political system.

Crucially, although Ryan is talking about the budget, he doesn’t reference any numbers or specific programs. Instead, he bring up his faith, the country’s founding, and the simple people vs. government dichotomy. In doing so he takes an argument about dollars or resources and turns it into an argument about values. The question is no longer whether taxes should be 21% or 19% of GDP, the question whether we want to live in a government-run dystopia.

What’s troubling about Ryan’s strategy is that people tend to be less willing to compromise when a conflict involves values. Studies have shown that relative to resource-conflicts, value-conflicts involve more anger, a greater inability to find common ground, and compromises that are less mutually beneficial. The reason for these differences is that a compromise involving resources can simply be seen as an isolated incident about something external to the self, but a compromise involving a value is a threat to your identity. For example, if you live your life based on the principle that small government is better, supporting a compromise that involves increasing taxes can be viewed as an admission that small government principles are not of paramount importance. But admitting that your core belief is not immaculate is an incredibly hurtful thing to do, and people will try to avoid that degree of painful self-doubt. After all, if you’re now unsure about the only thing you were sure about, you could be wrong about anything. Can you even be sure you’re still good person?

Research establishing the differences between value and resource conflicts has been fairly conclusive, but less is known about the specific mechanisms through which the different types of conflict influence behavior. What is it about the negotiations that changes, and how do these changes make compromise harder?

A new study led by Marina Kouzakova of Leiden University suggests a couple of answers. Via a webcam, participants in the experiment discussed with a confederate how they should travel to their vacation destination. Though the destination was accessible by many means of transportation, the discussion focused on whether or not to take an airplane. In the value-conflict condition, participants were told to base their decision on whether they thought air-travel was too damaging to the the environment to justify taking a plane. In the resource-conflict condition participants were told to base their decision on whether air travel was too expensive to justify taking a plane.

After the discussion the researchers measured two potential influences on decision making. First, they had participants fill out a survey that would measure their regulatory focus, a construct that describes the different ways people are motivated to pursue goals. In general, people fall on a continuum between prevention-focus and promotion-focus. People with a prevention-focus are motivated to maintain security and avoid losses, while people with a promotion-focus are motivated by gains and growth. These orientations can influence human motivation in a number of ways. For example, people with a promotion-focus will be more motivated to do something for a $5 reward, while people with a prevention-focus will be more motivated to do something in order to avoid a $5 loss (pdf).

People tend to chronically fall somewhere along the spectrum, but regulatory focus can also be influenced by situational factors. When the researchers looked at regulatory focus immediately after the discussion over air travel, they found that participants in the value-conflict condition were significantly more likely to be prevention-focused.

It’s not hard to see how these shifts in regulatory focus could affect the political process. For example, if a liberal senator is promotion-focused he’s more likely to be open to increasing spending on food stamps even if it means cutting spending on renewable energy programs. Alternatively, if the senator is prevention-focused, he’s more likely to be concerned with protecting the current level of green energy subsidy. Overall, the more politicians drift towards a prevention-focus the more gridlock there will be. Even when there’s not a specific decision to be made, a prevention-focus will make politicians less likely to take risks and more likely to focus on not jeopardizing their position by angering of their base.

The second thing the researchers measured was the cardiovascular (CV) reposes of participants. Models of arousal regulation have shown that people tend to have different CV responses in a “threat motivational state,” which results when a person feels they have insufficient resources to deal with a situation, and a “challenge motivational state,” which results when a person feels they do have the necessary resources. When the researchers looked at the CV responses after the negotiation they found that participants in the value-conflict condition had CV responses that were more in-line with the threat state. The finding provides further support that value-conflicts pose a threat to people’s identities, and that these threats lead to physiological changes as well as psychological changes.

The bad news is that neither altered CV responses nor swings in regulatory focus seem to be things that politicians are uniquely equipped to avoid. In fact, given that politicians have worked very hard to get where they are and have a lot to lose, they may even be more susceptible to taking up a prevention-focus than the participants in the experiments.

In the end the study does a nice job illuminating one of many bad cycles that have engulfed the American political system. The propensity to frame things in terms of values makes compromise something that’s more likely to hurt a politician than help him. Horse-trading is allowed by the electorate; betraying a key principle is not. But framing disputes as value-conflicts also helps consolidate support and strengthen the resolve of those who are on your side, even as it makes it more difficult to find common ground with those who aren’t on your side. And as long as framing disputes as value-conflicts helps politicians get re-elected, that’s the way we’ll continue to talk about them, even if it continues to make compromise more difficult.

(cross-posted from The Inertia Trap)
Kouzakova, M., Harinck, F., Ellemers, N., & Scheepers, D. (2013). At the Heart of a Conflict: Cardiovascular and Self-Regulation Responses to Value Versus Resource Conflicts Social Psychological and Personality Science DOI: 10.1177/1948550613486673

Education Policy Needs Two Offenses (Or What I Learned From Albert Shanker)

I’ve been reading Richard Kahlenberg’s excellent biography of Albert Shanker (here’s some mid-book blogging on teacher qualifications and LIFO), but now that I’ve finally finished I wanted to take a moment to wax philosophically about what made Shanker special and why I believe that quality is missing from the current policy environment.

More than anything else, the reason Shanker was so respected by people across the political and ideological spectrum is that he was a man of ideas and alternatives rather than opposition. For example, when the issue of teacher quality started to gain prominence in the 1980’s teachers were faced with the prospect of pay cuts or decreased job security. While the NEA remained tethered to the position that there were no bad teachers, only bad circumstances in which teachers had to teach, Shanker eschewed political expediency and admitted that yes, there were in fact bad teachers. The NEA continued to fight the idea of differentiating between teachers, but Shanker was focused on working to identify great teachers. His efforts ultimately helped lead to the creation of the National Board For Professional Teaching Standards. Has the Board solved the every single teacher quality issue? Of course not. But it serves a purpose, and the reason it exists is because rather than simply opposing his opponents or denying a problem existed, Shanker countered with proposals of his own.

That’s not to say Shanker couldn’t be a fierce opponent. He turned friends into enemies with his steadfast opposition to race-based policies (e.g. quotas), and many of his initiatives, such as charter schools and strong standards, were indirect ways of fighting the school voucher movement. But that doesn’t take away from the fact that he consistently put his considerable muscle behind legitimate alternatives to policies he opposed. The result is that there were essentially two sides playing offense, and I think this created a positive environment for creativity, cooperation, and policy development.

Unfortunately, it now seems as though instead of two offenses, we have a situation more akin to an offense and a defense. Those who oppose contemporary systems of accountability or school choice are unified in their steadfast opposition, but unlike Shanker they haven’t built a consensus behind specific alternative solutions to the problems the policies they oppose ate attempting to address. I’ve talked about this before, but for all the invective lobbed at standardized testing it’s surprising how little you hear about legitimate alternatives. Sure, there are op-eds from time to time and classrooms or whole schools will occasionally try something out. But you don’t have anybody crafting a realistic large scale pilot project that could show whether an alternative system of assessment is effective or affordable.

Similarly, you don’t have powerful stakeholders (e.g. the labor movement) putting their weight behind any kind of school choice initiatives that are better aligned with their priorities. Yesterday Deborah Meier once again reminded people of these possibilities, but the events she references happened 20 years ago. What school-choice alternatives to the contemporary charter movement have been proposed since then? Obviously it’s extremely difficult to fund these kinds of projects at the necessary scale, and it can also be impossible to get the politics right. But it’s not as if the labor movement and its allies are some two-bit operation.

The single-minded opposition is understandable to a point. When everything being done is the antithesis of what you know should be done, your priority is just to stop everything and not worry about what come next. There are also exceptions that are more in line with how I think Shanker would have approached things. For example, the UFT seems very serious about experimenting with the community schools model, and I applaud their efforts. But for the most part one side of the policy debate seems to prize fervent opposition over policy ingenuity. Shanker showed that we’re better off when the opposite occurs.

What if 501(c)4 Organizations Had to Disclose Their Donors?

The IRS kerfuffle has increased interest in the tax code by about 5700%, and one outcome is that people are starting to take a closer look who is being granted exemptions. Dylan Matthews has thoughtful piece on 501(c)4 organizations, the groups at the center of the scandal, but instead of focusing on taxes, Matthews thinks the real issue is disclosure:

501(c)4s can accept unlimited donations and don’t have to tell a soul from whence they came. 527s, including super-PACs, have to file quarterly reports disclosing donors. That’s why so many super-PACs have attached 501(c)4s, which can collect unlimited donations and then donate them in turn to the super-PAC, as Fred Wertheimer explained to me last week.

The key question, then, in considering what should come next for 501(c)4s, is not “should groups like this have to pay taxes.” They’re never going to have to pay taxes. It’s whether they should have to disclose their donors.

Mathews goes on to lay out the different arguments. One on hand, it doesn’t seem like the 501(c)4 designation serves a particular purpose, and the ability of 527s to use them to effectively launder donations should be stopped. On the other hand, you to also need to be really really careful about the potential to stifle free speech in a detrimental way.

As Neil Irwin explained last week, there’s a long history of IRS persecution of LGBT-oriented charity groups, including denying one group tax-exempt status and making another demonstrate that it would not “encourage or facilitate homosexual practices or encourage the development of homosexual attitudes and propensities by minor individuals.” Making all charitable groups 501(c)3s could open them to that kind of scrutiny, particularly for any that deal with controversial topics, which could be far worse than anything that’s happened in the current scandal.

Matthews chooses to focus on the broader philosophical conflict and thus he takes a long-run view of the issue. He essentially wants to know which policy will leave the country in the best shape 30 years from now. But there’s also the question of the immediate practical impact of increased disclosure. Does giving people information about donors actually impact voting behavior?

A new study from Conor Dowling of the University of Mississippi provides one data point. Dowling’s study was based on a 2010 ad by American Crossroads, an organization founded by Karl Rove, that aired in the Missouri senate race. The ad attacked Democratic candidate Robin Carnahan by attempting to link her to President Obama, Nancy Pelosi, and “failed” Democratic policies.

After participants viewed the ad, Dowling and his co-author, Amber Wichowsky, placed them in one of five experimental conditions. The first group was shown the names, occupations, and donation amounts for American Crossroads’ five biggest donors. Groups 2-4 were shown an article about the large independent expenditures being made in the race, but Group 3 read additional information about the financial interests of American Crossroads donors (e.g. $25 million from “corporate executives” in the oil industry), and Group 4 read additional information about how American Crossroads didn’t have to disclose donors because of their 501(c)4 status. The 5th group only watched the ad, and there was also a control group that never watched the ad in the first place. At the end of the experiment, all participants rated how likely they were to vote for Carnahan.

The researchers found that the disclosure with the largest impact was the article describing why there was no disclosure. Participants in this condition gave Carnahan support that was significantly stronger than the support given by participants who only watched the attack ad, and their overall level of support was about the same as participants who never saw the ad. In other words, being told about how 501(c)4s didn’t have to disclose their donors completely negated the effect of the ad. The disclosure involving the top 5 donors also had a relatively strong impact, although the difference in support between participants who saw the 5 donors and participants who only saw the ad was not statistically significant (p =.64). A follow up experiment using an ad in which a Republican candidate was attacked largely replicated the findings, although in the thins experiment learing that disclosure is not required led Democrats to increase support for the Democratic candidate even when the Democrat was the one on the attack.

The straightforward explanation for the findings is that disclosure decreases support because people have negative associations with the specific people or groups who tend to fund these organizations. But the fact that the legality of non-disclosure may play a key role suggests that people are generally disgusted with the system, and when something reminds them of it — for example, an ad — they will develop a negative attitude toward that thing. People just don’t want to think about the fact that much of what we do on a daily basis is for money.

To get back to Matthews’ original discussion, I think “free speech vs. money-in-politics” is one of the more sane debates in our political system. Neither side takes positions that are completely indefensible. There’s no Michelle Bachman arguing she should be able to slip $2 million in cash in the trunk of a somebody’s car. It’s clear that money in politics is an issue worth watching, but if you value freedom very highly, it’s also valid to be concerned that a solution could open the door for costly free speech suppression.

It seems like the future compromise over campaign finance — whether it’s 2, 5, or 25 years from now — will essentially involve an exchange of more freedom for less anonymity. Feel free to donate whatever you want with no strings attached, but everybody’s going to know. If you want to influence public policy, you’ll have to be a public figure. That’s probably our best chance of curtailing the influence of money while also satisfying those who worry we’re becoming a fascist tyranny.
Dowling, C., & Wichowsky, A. (2013). Does It Matter Who’s Behind the Curtain? Anonymity in Political Advertising and the Effects of Campaign Finance Disclosure American Politics Research DOI: 10.1177/1532673X13480828

How Your Social Status Influences the Way You’re Judged

One of the palpable weaknesses in the American justice system is the tendency for it to produce different outcomes for people from different social classes. Part of this is a result of discrepancies in the quality of legal representation people can afford, but part of it is also due to inconsistencies in the way morally questionable activities are judged.

Research on judgments of misdeeds often focuses on “moral licensing,” or the idea that certain circumstances can make us more amenable to bad behavior. One study that’s received a lot of attention found that people who bought organic food rather than a control food were less likely to volunteer to help a needy stranger. The implication is that doing one good deed made people feel it was less wrong to pass up an opportunity to do another good deed. Other studies have found that this kind of licensing can be influenced by the the actions of people in your group or people with whom you share a random characteristic such as a birthday. One study even found that people grant themselves a license to do bad things by exaggerating the evil of things they previously decided not to do.

However, thus far most research on moral licensing deals with people making judgements about their own actions. We still don’t know much about when or why people are willing to grant third parties a license to do something questionable. Given that you have less knowledge about other people, it seems likely that different factors would be involved.

The University of Wisconsin’s Evan Polman sought to fill this gap by specifically examining whether the social status of a third party influenced the degree to which they were granted a moral license. In two studies Polman and his team told participants about somebody who engaged in questionable activity, but they manipulated status through the person’s name (Billy-Bob (low) vs. Winston Rivington (high) vs. James (control)) or job (janitor vs. CEO). The researchers found that both high- and low-status actors were judged significantly less harshly than actors in the control condition. It would seem that both high- and low-status people tend to be given more leeway to do bad things than people in the middle.

Of course it seems unlikely that high- and low-status people would be granted a moral license for the same reason. That led Polman to investigate a more interesting question, which is how exactly status influences moral licensing. Polman hypothesized that high-status people were given “credentials” while low status people were given “credits.”

While credentials and credits similarly elicit less negative responses to misbehavior, they differ with respect to whether they alter observers’ perception of the negativity of the behavior itself (a form of perceptual change) or merely affect observers’ perception of the extent to which engaging in negative behavior is understandable or justified (a form of attitudinal change). Credentials bias perceptions of norm-violating behavior, leading people to perceive dubious behavior as less dubious; almost as if the behavior was not even a transgression (Effron & Monin, 201o)…

Credits, however, do not change observers’ perception of the behavior but offer counterbalancing capital so that a wrongdoer can transgress as long as their transgressions (so-called moral debits; Miller & Effron, 2010) do not exceed their credits (Nisan, 1991; Zhong, Liljenquist, & Cain, 2009). Thus, credits influence the extent to which observers sanction misbehavior by altering their attitudes toward the wrongdoing, such as viewing the wrongdoing as justified and tolerable.

In other words, when a high-status person does something questionable (e.g. stealing), you judge it less harshly because you believe the action is objectively less wrong (e.g. it was actually a clever manipulation of tax laws.) But when a low-status person does something questionable, you judge it less harshly because you understand why the person did it and you’re sympathetic to their situation (e.g. they needed money for food.)

Polman reasoned that when an action was ambiguous, high status people would be judged less harshly because it would be easier to reinterpret their action as something justifiable. He decided to test this hypothesis by directly manipulating whether the morally questionable action was ambiguous or not. Participants read about a janitor or a CEO who hired only white candidates, but some participants were told the person admitted being reluctant to hire African-Americans (unambiguous condition), and some participants were told the hiring was legitimately based on merit (ambiguous condition). Sure enough, when the action was ambiguous, participants rated the action less harshly when it was done by the CEO rather than the janitor. On the other hand, when the action was unambiguous, and thus it was impossible to view the action as permissible, participants judged the action more harshly when it was done by the CEO and less harshly when it was done by the janitor.

Polman also measured the dispositional sympathy of participants. As predicted, participants who were more sympathetic tended to judge those who were low-status less harshly. However, dispositional sympathy had no effect on the judgments of high-status people. To Polman, this was evidence that low status people were in fact being judged less harshly because their struggles were earning “sympathy credits” rather than “credentials.”

The findings suggest that high- and low-status people are judged differently, but not in a way that is universally advantageous to either of them. When the indecency of an action is ambiguous, high-status people will be judged less harshly because their actions are more likely to be interpreted in a positive light. However, when an action is unambiguous, low-status people will be judged less harshly because they are more likely to elicit sympathy.

As Polman points out, the research has implications for how you ought to defend yourself. If you’re clearly guilty, and thus your transgression is unambiguous, the optimal strategy may be to lower your status in the hope of receiving sympathy. For example, by begging the victim for forgiveness. Alternatively, if there’s ambiguity in your actions, it’s worthwhile to try and raise your status in order to make it more likely your actions will be viewed in a more positive light. Similarly, if you’re a high-status person but the ambiguity of your actions is open to debate, the best defense may be to obfuscate the circumstances of your crime. Because you’re high status, when your behavior is ambiguous a judge is more likely to “credential” your behavior by deciding that your actions weren’t all that bad.

Is there evidence of any of this in the real world? You can certainly string together enough high-status crimes for a solid bout of confirmation bias. For example, when high-status people do something unambiguous, such as commit murder, there doesn’t seem to be any special leniency. On the other hand, when their actions are more ambiguous, such as those involving regulatory improprieties in the banking sector, high-status people do seem to often go unpunished. Of course this doesn’t really prove anything — some actual data is needed. Either way, Polman’s research is a good reminder that social status matters, but not always in the way you might think.
Polman, E., Pettit, N., & Wiesenfeld, B. (2013). Effects of wrongdoer status on moral licensing Journal of Experimental Social Psychology, 49 (4), 614-623 DOI: 10.1016/j.jesp.2013.03.012

We’re All Just Looking For a Patsy

If often seems as though policy-making has devolved into nothing more than a contest where the goal is to blame as many people as possible (but not yourself) for the country’s problems. Fossil fuel companies blame environmental regulations for economic stagnation and high energy prices. Neocons blame civil libertarians for national security weaknesses. And of course, Westboro Baptist blames homosexuality for everything. Much of the blame appears to be farfetched (the exception is homosexuality, which since the days of Newton has been scientifically proven to cause hurricanes), and though the political motivations behind the blame game seem easy to understand, a new study led by Zachary Rothschild of the University of Kansas helps break down exactly when and why people latch on to a scapegoat.

Rothschild and his team were interested in examining how the potential culpability of one’s own group influenced moral outrage and blame for a third-party. They began their experiment by giving participants a survey that led participants to categorize themselves as middle class rather than working class or upper class. Participants then read an article about the struggles of working-class Americans, but in the in-group condition the article blamed the middle class for the struggles, in the out-group condition the article blamed the upper class, and in the unknown condition the article stated that economists don’t know the cause of working-class struggles. Participants then read another article about the status of illegal immigrants. In the viable scapegoat condition the article described the rising fortunes of illegal immigrants, while in the non-viable scapegoat condition the articles describe how illegal immigrants were also struggling to find work.

As expected, when illegal immigrants were viable scapegoats, participants were more likely to blame them for the struggles of the working-class when the cause of those struggles was unknown or attributed to their own group, the middle-class. Participants in the in-group condition also reported more moral outrage at illegal immigrants and a stronger desire for retributive action than participants in the out-group or unknown conditions. When illegal immigrants were not portrayed as a viable scapegoat (i.e. they were shown to also be struggling), participant perceptions about the cause of working class struggles had no effect on how they viewed illegal immigrants. In general, when a third-party was presented as a viable scapegoat, people were more likely to blame the third-party for a negative outcome when their own group was at risk of being viewed as responsible for that outcome.

The findings are intuitive, but nonetheless important. When it comes to the specific issue of immigration, the study highlights how simple political motivations can lead to bad policy. Though most economists agree that increasing immigration will boost the economy, as long as the economy is weak people from both political parties may be motivated to scapegoat immigrants in order to deflect blame and alleviate collective guilt about struggling families. This makes immigration reform more likely to occur once the economy is strong, but in the mean time families miss out on the economic benefits of immigration when they need them most.

More broadly, the politics of scapegoating can prevent problems from actually getting solved.

The irony of the present findings is that moral outrage in the present study arguably ensures the maintenance of the status quo of disadvantaged group’s suffering, while providing advantaged group members with an air of self-righteousness. That is, outraged advantaged group members may punish a supposed third-party culprit in the name of restoring justice for a disadvantaged group while simultaneously continuing to perpetrate harm against the disadvantaged without repair. Stated differently, there is the appearance of justice being served only to ensure that injustice is preserved.

You could alleviate your guilt and protect your group’s moral standing by working to help the disadvantaged, but it’s often easier to accomplish those things by blaming a third-party. In the end the disadvantaged group never gets help. So if you happen to see somebody throwing a lot of blame around, particularly toward groups that don’t seem to have the power ascribed to them, you should probably view that person with extreme skepticism.
Rothschild, Z., Landau, M., Molina, L., Branscombe, N., & Sullivan, D. (2013). Displacing Blame over the Ingroup’s Harming of a Disadvantaged Group can Fuel Moral Outrage at a Third-Party Scapegoat Journal of Experimental Social Psychology DOI: 10.1016/j.jesp.2013.05.005

Teacher Expectations Have a Stronger Impact On Low-Income Students

In their 1968 book Pygmalion in the Classroom, Robert Rosenthal and Lenore Jacobson presented their groundbreaking research that showed teacher expectations are self-fulfilling prophecies. If two students start the school year at the same achievement level, the student the teacher is told is a high achiever will make more gains than the student the teacher believes is a low achiever.

Over the years researchers have confirmed that teacher expectations have small to moderate effects on student achievement, but more recent work suggests that the effects may not be the same for all students. For example, a new study by Temple’s Nicole Sorhagen provides strong evidence that the effects of teacher expectations vary by family income. Using longitudinal data from 10 cities, Sorhagen found that the effects of teacher expectations, whether they’re positive or negative, are stronger for students from low-income families.

This study investigated one aspect of the complex cognitive and behavioral processes underlying student–teacher relationships and found that early inaccurate teacher expectations were a lasting contributor to later academic performance. Furthermore, the findings suggest that self-fulfilling prophecies in the classroom vary across academic subjects and family income. Under- and overes- timation of early math and language abilities, but not reading abilities, seemed to have a more meaningful effect on students from lower income families. The fact that self-fulfilling prophecies in first-grade classrooms exerted an especially lasting impact on the achievement of disadvantaged students raises the possibility that teachers’ underestimation of poor children’s academic abilities may be one factor that contributes to the persistent and worrisome gap in achievement between children from different socioeconomic backgrounds. On the other hand, teachers’ overes- timation of abilities seemed to disproportionally help low-income students, suggesting that knowledge of self-fulfilling prophecies in the classroom could be relevant to policies aimed at ameliorating the achievement gap between low- and high-income students, especially considering the persistence of the achievement gap in America.

One neat thing about the study is that it hints at a mechanism for the power of school culture in low-income communities. A positive “no-excuses” environment is one way to encourage teachers to have high expectations for all students, and the study suggests these expectations will have a larger impact on low-income students. Something similar should also occur with regard to peer effects. If a student ends up in a classroom with higher achievers, the teacher is more likely to expect that they too are a higher achiever, and the impact of those higher expectations will be stronger for low-income students.

The importance of teacher expectations is also a reminder of why it’s important to evaluate teachers using multiple measures. While there’s often a big fuss made about the things teachers do that aren’t captured by test scores, the impact of teacher expectations is something that can only be captured by test scores. Even if you observed a teacher every single day of the year you wouldn’t pick up on whether they’re being sufficiently optimistic and equitable in their expectations for students. Such behaviors are too subtle to capture — what you need is a broad all-encompassing assessment that will incorporate the impact of teacher expectations even if it can’t specifically identify it.
Sorhagen, N. (2013). Early teacher expectations disproportionately affect poor children’s high school performance. Journal of Educational Psychology, 105 (2), 465-477 DOI: 10.1037/a0031754