​The last decade has seen a huge upsurge in researcher interest in the consequences of pre-k education. That’s due, in part, to the steady increase over the last fifty years in the number of children enrolled in pre-k. In the last twenty years, that increase has been driven by children enrolled in public programs. 

The increase in publicly funded programs naturally enough sparks interest among policymakers as to whether these programs work.

That’s not an easy research question. It’s hard to track children in future years as their families move, it’s more difficult to construct reliable assessments for younger children, and there’s less agreement about what constitutes successful outcomes in pre-k than in higher grades.

Perhaps most troubling, it’s not obvious what the counterfactual is; in other words, you’d like to compare the outcome for a child if he goes to pre-k compared to the outcome for the same child if he doesn’t go to pre-k. That’s obviously impossible, so you’ll compare the child to another child who doesn’t go to pre-k….but what does that child do? Fifty years ago it was a good bet that “didn’t go to pre-k” meant the child was at home with his mother. Today, he might go to a home daycare, or be cared for by a relative. To what we should compare pre-k just isn’t so obvious.

In addition, the answers to these questions are clouded by political factors. Ones assessment of the efficacy of public pre-k programs has the potential to be influenced by one’s preconceptions of the efficacy of government programs more generally, and by preconceptions as to whether such programs ought to be within the role of government, whatever their efficacy.

Some researchers have countered the latter set of concerns by noting that effective pre-k programs can more than pay for themselves; children who attend effective pre-k programs will be less likely to drop out of school, more likely to end up in high paying jobs (and so pay more taxes), are less likely to be incarcerated, to need public assistance, and so on. Some researchers claim that pre-k programs return ten dollars or more for each dollar spent.

The pressing questions are: (1) are these claims accurate; (2) what are the characteristics of a “high quality” pre-k program; (3) can governments create and sustain pre-k programs with these features at scale?

Two groups of researchers recognized the need to bring together existing research and to provide policymakers with some answers that are, insofar as is possible, objective and devoid of political ax-grinding. This academic world is small enough that it was inevitable that each should learn about the other, and they chose to join forces. The result is a report, The Current State of Scientific Knowledge on Pre-Kindergarten Effects.

The heart of the report is a consensus statement. I offer four key conclusions from that statement here, verbatim, in red, with brief comments of my own after each conclusion. 

Studies of different groups of preschoolers often find greater improvement in learning at the end of the pre-k year for economically disadvantaged children and dual language learners than for more advantaged and English-proficient children.

That’s the counterfactual at work. Rich and poor kids would have different experiences if they were not in pre-k, with poor kids having fewer opportunities for an enriching environment than the wealthy kids.

Pre-k programs are not all equally effective. Several effectiveness factors may be at work in the most successful programs. One such factor supporting early learning is a well implemented, evidence-based curriculum. Coaching for teachers, as well as efforts to promote orderly but active classrooms, may also be helpful.

This is not a surprise…the curriculum matters, and providing training and direction to teachers helps. In the details of the report, curricular comparisons are pretty rough-cut: whole-child vs. skills-based (i.e., math, literacy or both). Whole-child curricula have not been successful in developing literacy, math, or socio-emotional skills…but it also sounds like a bit of a basket category.

Children’s early learning trajectories depend on the quality of their learning experiences not only before and during their pre-k year, but also following the pre-k year. Classroom experiences early in elementary school can serve as charging stations for sustaining and amplifying pre-k learning gains. One good bet for powering up later learning is elementary school classrooms that provide individualization and differentiation in instructional content and strategies.

This is one of the most important points. It’s saying that the oft-cited Perry and Abcederian preschool results are atypical. Absent continued intervention, you should expect fadeout of the pre-k benefit. I’ve blogged about relevant studies before, but the main point is intuitive. Academic and social outcomes are a product not just of school experiences, but also of home and other out-of-school experiences. If those out-of-school experiences are not especially enriching, children benefit (to a greater or lesser degree) from substituting pre-k experiences. The out-of-school experiences continue to matter after pre-k.

When you spell it out, the counter-assumption sounds a little strange: children may be behind their peers in important knowledge and skills by age four, but with a good year or two of pre-k they catch up, and can keep pace with their wealthier peers thereafter. This assumption makes sense if you ascribe an outsize importance to the first few years of development, e.g., as Freud did. But that assumption isn’t right. Outside-of-school experiences continue to matter after age six.

Convincing evidence shows that children attending a diverse array of state and school district pre-k programs are more ready for school at the end of their pre-k year than children who do not attend pre-k. Improvements in academic areas such as literacy and numeracy are most common; the smaller number of studies of social-emotional and self-regulatory development generally show more modest improvements in those areas.

There’s more than one way to do this sucessfully, and the public sector can get it right. There are probably three reasons (at least) that evidence is better for literacy and numeracy than for socio-emotional learning. First, we know better how to teach numbers and letters because we’ve been at it longer. Second, there’s less happening at home that might conflict with the learning happening at school when it comes to these skills.  Third, the relative contributions of environment vs. heritable factors (e.g., temperament) is probably larger for literacy and numeracy. 

In addition to the consensus statement, the report includes brief but meaty analyses of questions important to pre-k policy, for example:

• How can scale-up be improved? (Training and monitoring at a level of detail similar to that used during initial design.)
• Is the economic return really 10 dollars or more, per dollar spent? (That figure may apply to small scale programs operating in the 1960’s. Today, with a scaled up program expect more like 2-4 dollars per dollar spent).
• Should pre-k be targeted or universal? (It depends on the circumstances in the state or district—both are effective).

If you have even a passing interest in pre-k, I recommend this report to you. 

​As digital devices have decreased in price, they have become more available to more children. The impact of this availability on children’s social lives have been debated with vigor, often with gloomy foreboding. The concern centers on the possibility that online activities are absorbing so much of children’s time that little is left for other worthy pursuits, e.g., face-to-face conversations.

​But data to inform these opinions have been lacking. We know children who own digital devices (and now, that’s most of them) spend a great deal of time interacting with them—estimates for teens are around 10 hours per day or more. To some, it’s self-evident that must carry a cost, and the cost is assumed to be social. Kids are isolated by digital devices, kids no longer know how to speak face to face, and so on (e.g., here.)

But this was really punditry and speculation. Hard data were lacking, especially hard data to which causality could be ascribed. That is, we might see kids who were socially withdrawn who spent a lot of time in online social pursuits, but such a correlation would be hard to interpret. Were online social interactions replacing face-to-face interactions and causing social isolation, or would this child be withdrawn in any event, and online communication is actually a more social activity than this child would otherwise engage in?

Clearly, a true experiment would help clear the matter up: take 1,000 middle-schoolers, give 500 of them a computer, and see what happens over the course of a school year. Well, darned if someone didn’t do that (Fairlie & Kalil, 2017).

The experimenters administered a survey at the start and the end of the school year. They also had access to administrative data regarding school participation.

It should be noted that children in the control group did have access to computer time at school and elsewhere, and some families purchased computers on their own during the year. The researchers tracked these confounds as best they could. Children given the computers did indeed spend more computer time per week than control kids.

The results:
Friends: The results showed that kids given computers did not report communicating with or hanging out with their friends less…in fact, they reported spending more time with friends.
Social groups: Giving kids a computer had no impact on the probability that they would be part of a sports team, club, or music group.
School participation: There was also no effect of home computers on the number of days absent from school (or tardy), or days suspended.
Competing activities: Self-reported TV time, homework time and leisure reading were unaffected.
Social networking: Children with a home computer were more likely to have a social network page and reported spending more time on social networks. There was also a statistically nonsignificant increase in the probability of reporting cyberbullying, a result that is difficult to interpret because the overall mean was so low (less than 1%).

A few caveats of these conclusions should be borne in mind. First, the study only lasted for one school year. Second, having a smart phone, with the constant access it affords, may yield different results. Third, children were given a computer, but not Internet access. Some kids had it anyway, but the more profound effects may come from online access.

All that said, I am less frightened than some by the threat that digital technologies will eat children’s minds, or making them anti-social zombies. I wrote The Reading Mind before this study was published, but as I put together the data, I suggested that digital activities were not replacing reading, and that’s true for two reasons. First, reading provides a different sort of pleasure than gaming or social networking. If you like reading, that pleasure is only available by reading. Second, digital technology has not reduced reading for most kids because most kids don’t read anyway.

This later point is the most salient to me, and has most influenced how I raise my own kids. It’s not that most digital technologies are so terrible, but most of what my kids can do online is less preferable to me than what they can do offline. I’d rather they make something, take a bike ride, or read a book. But if they horse around on the computer, that’s no worse (or better) than watching Say Yes to the Dress, my ten-year-old's latest television infatuation.

My real concern about digital technology use in teens is hard to quantify. When I was a teen I, like most, probably assigned too much value to the opinions of my peers.  They necessarily stopped influencing me when I got off the school bus, and I was influenced mostly by my parents and two sisters. I don’t relish the thought of children taking their peer groups home with them in their pockets, influencing them 24/7, and diminishing the impact of their families. 

Suppose a teacher asks students to read a text and he wants to be sure that children notice X, Y, and Z about the text. One strategy would be to pose questions before they read, the answers to which are X, Y, and Z. The problem with this strategy is that it increases the chances they will notice and remember X, Y, and Z, and decreases the chances they will notice everything else (e.g., Pressley et al, 1990).

The interpretation of this phenomenon is straightforward: posing questions keys the reader’s attention to certain content, and they pay less attention to everything else. Indeed, when researchers put in boldface type information in the text that was not prequestioned, but would later appear on the test, the disadvantage vanished (Richland et al, 2009). The boldface drew attention to the content that would have otherwise been skimmed.

In some cases, the teacher may view that as a worthwhile tradeoff, but that’s probably infrequent—she’d prefer the boost to X, Y, and Z occur without the cost. 

A new paper (Carpenter & Toftness, 2017) may suggest a strategy to avoid the problem, although the experiment actually tested memory for video content. 

The researchers tested 85 undergraduates, each of whom watched a video lasting about eight minutes. The video was divided into three segments describing the original settlement of Easter Island, religious practices there, and the arrival of outsiders. The researchers generated four questions about each segment. Half of the subjects were asked to guess at the answer to two randomly selected questions about the upcoming segment. The control group simply pushed a button to continue on with the video. At the end of the video, subjects in both groups answered all 12 questions about the video in random order.

​Just as in prior experiments using text rather than video, asking people about specific information before seeing the video made it more likely they would learn that information once they watched the video. (That’s the tallest bar at far left). 

More interesting is that the prequestioned group was also more likely to successfully answer questions for which they had not been precued. 

Why was the cost not observed? Carpenter & Toftness emphasize that a reader controls the pace of reading; the reader can skip over content that she deems less important, and read again content that matters more. The viewer does not control the pace of video; important content might pop up at any time, and once it’s past it cannot be reviewed. So the viewer is more likely to attend closely to the whole thing. 

The researchers note that this attention hypothesis can help explain other instances in the research literature where the prequestioning deficit for other content is not observed. For example, Pressely et al (1990) asked subjects to rate each paragraph for interest. Little & Bjork (2016) showed that non-prequestioned information got a boost if it was mentioned in a prequestion, although not the target to-be-learned information. 

So in the final analysis, can teachers pose prequestions in a way that boosts memory for targeted content but doesn’t incur a cost for everything else? 

​In principle, yes. With the right type of material (boldfaced, or video), you're good, and asking for interest judgments works too. But of course none of these may be practicable as the teacher envisions the lesson plan. 

This work suggests that a teacher could devise another strategy that uses prequestions without cost--a mental task that requires attention to all content, not just the prequestioned would do the trick. True in principle, but the bottom line on prequestions a the moment seems to be “proceed with caution.” 

It's been about ten years since college students on most campuses began to take notes on a laptop in lecture classes. So that's about 10 years of professors fretting about it.

It seems obvious that students are, at the least, themselves distracted during lectures--supporting anecdotal evidence can be collected by popping into the back of any large lecture hall and observing laptop content; you'll see plenty of social media websites open, students responding to messages or email, and a surprising number of people shopping.

​More insidious, some data indicate that using a laptop is distracting to students sitting behind the multi-tasking student.

But most data on the consequences of laptop use are correlational; for example, students who report taking notes on laptops earn lower grades than those who don't, and observational data show that students have non-course related software open during class (e.g., Kraushaar & Novak, 2010). The obvious problem is that laptop use may not be causing low grades; laptop use may be more appealing to those who would earn lower grades anyway. 

Two new studies using different methods suggest that laptop use does, indeed, incur a cost to students.

In one (Carter et al., 2017) students in sections of a required introductory economics course were randomly assigned to use a laptop or to refrain from doing so. In a clever twist, some classrooms were asked to take notes on tablets with the tablet lying flat, hence reducing the likelihood that the screen would distract neighboring students.

The final examination was the same across sections, so scores on the exam served as the outcome measure for the study. Laptop (or tablet) use was associated with lower scores, d = -.18 (controlling for demographic variables, GPA, and ACT scores at admission). It is noteworthy that all of the effect was carried by performance on the multiple choice and short-answer portion of the exam, with no effect on essay grade. The researchers noted that essay grading is not objective and professors might center their grades on student performance. 

We can have greater confidence that laptop use is really causal here, given that class sections were randomly assigned to use them or not. But there are some other caveats of interpretation. It's possible that teachers teach differently when they know their students are taking notes on devices. We should also be cautious in generalizing beyond the institution (West Point) and the class (economics). For example, perhaps economics requires the drawing of many figures and graphs, a process that's clumsy on a device. Perhaps the same experiment conducted in an English literature class would show no effect. Or maybe laptop use offers a small benefit to most students, but there's a minority for whom the forced use of a laptop incurs a big cost.

The second study (Patterson & Patterson, 2017) uses a different, complimentary method that avoids these problems (but has others).

​The researchers noted that, at their institution, different classes required laptops (15%), forbade them (2%), or made them optional (83%). The researchers reasoned that a student who had a laptop-optional class might be nudged into using one if she had a laptop-required class the same day. After all, she would already have the laptop with her, so why not use it? And indeed, their survey showed that students were about 20% more likely to use a laptop under those circumstances.

And students were 48% less likely to use laptop in a laptop-optional class if they had another class the same day that forbade laptop use. 

So researchers compared grades in a class where some students are biased to use laptops and others were biased not to use them--biased, not forced, so if a student believes he really takes better notes on a laptop, he can use it. And both the laptop-user and non-user are in the same class, with the same opportunities to be distracted by other students, and experiencing the same lecture from the professor. 

Researchers analyzed about 5,500 grades in lap-top optional courses from undergraduate and masters students over the course of six semesters. All were enrolled at the same liberal arts college. 

The effect of having a laptop-required course at another time was d = -.32 to -.46.  Having a laptop-prohibited course the same day was associated with a positive effect, d = .14 to .25. (These effects controlled for gender, ethnicity, age, course load, course schedule difficulty, major, and GPA.)

Patterson & Patterson found that the negative effect was larger for weaker students--in fact, there was little or no effect for stronger students. They also found that laptop use (or more precisely, having a laptop-required course the same day) had a larger negative effective in quantitative classes. 
​
So we have two studies--a randomized control trial, and a quasi-experimental design--to add to the correlational studies showing that students are better off taking notes by hand than doing so on a laptop. None is perfect, but different designs have different flaws. This desirable state of affairs is referred to as converging operations--different types of experiments all point to the same conclusion. ​

In closing, it's worth remembering that this study does not concern all classroom laptop use, but rather one function: taking notes in a lecture course. Still, considering that most time in introductory college courses is spent listening to lectures, the impact of this work ought to be consequential.