Week 5 - Assignment: Understanding Peer-Reviewed, Primary, and Secondary Sources

Educational Researcher, Vol. 44 No. 3, pp. 173 –184 DOI: 10.3102/0013189X15574903 © 2015 AERA. http://er.aera.net

ApRIl 2015 173

M ost members of the American public will never read this article. Even if they are interested in education and charged with making important decisions about

schooling, it is unlikely that they will seek information from a peer-reviewed education journal. Instead, they will probably turn to other sources. Despite their well-publicized demise, the news media are some of these sources. Daily newspapers are a top source of education information in America, second only to family and friends (West, Whitehurst, & Dionne, 2011). Online news ties with newspapers as a source of information about local education (Rosenstiel, Mitchell, Purcell, & Rainie, 2011). Additionally, after years of decline, news use may be rebounding as a result of the increasing popularity and functionality of tab- lets and mobile devices (Pew Research Center, 2013).

Although researchers disagree about issues of mechanism and degree, a large body of literature suggests that the news media can and do influence decision making, perception, and even behavior (Croteau, Hoynes, & Milan, 2012; Gamson, 1992; Gerbner, Gross, Morgan, & Signorielli, 1982; Graber, 2009). In the education sphere, this means that news media coverage can

guide, to varying degrees, policymaking, practice, voting, and selection of schools. For instance, Howell (2008) found that aca- demic research results that contradicted the popular perception that private schools are superior to public schools changed the opinions of large percentages of study participants. In a day and age in which personal choice plays an expanding role in educa- tion, parents, students, and other nonexperts need evidence and information more than ever in order to select the most suitable school or register for the most appropriate level of coursework. As such, one might hope that the news media would lean toward presenting the best available evidence (Henig, 2008). At the very least, this means that the education research mentioned in the news media would have undergone the quality control measure of peer review. Further, one might expect (or at least hope) that academics, who are generally required to possess advanced graduate degrees in their areas of study, would be a major source of expertise. Yet little if any academic research has examined

574903EDRXXX10.3102/0013189X15574903Educational Researcher research-article2015

1Education Week Research Center, Editorial Projects in Education, Bethesda, MD

One Small Droplet: News Media Coverage of peer-Reviewed and University-Based Education Research and Academic Expertise Holly Yettick1

Most members of the American public will never read this article. Instead, they will obtain much of their information about education from the news media. Yet little academic research has examined the type or quality of education research and expertise they will find there. Through the lens of gatekeeping theory, this mixed-methods study aims to address that gap by examining the prevalence of news media citations of evidence that has undergone the quality-control measure of peer review and expertise associated with academics generally required to have expertise in their fields. Results suggest that, unlike science or medical journalists, education writers virtually never cite peer-reviewed research. Nor do they use the American Educational Research Association as a resource. Academic experts are also underrepresented in news media coverage, especially when compared to government officials. Barriers between the news media and academia include structural differences between research on education and the medical or life sciences as well as journalists’ lack of knowledge of the definition and value of peer review and tendency to apply and misapply news values to social science research and expertise.

Keywords: communication; content analysis; in-depth interviewing; media; regression analyses; research utilization

FEATURE ARTIClES

174 EDUCATIONAl RESEARCHER

what kind of education-related studies, data, and experts are actually mentioned by the news media. In a sense, the issue has fallen through the cracks of the mass communication and educa- tion fields. Further, potentially relevant studies are difficult to assemble in that they span a wide variety of fields such as mass communications, education, and political science.

The purpose of this study is to address the gaps in the litera- ture by examining the prevalence of news media mentions of “higher quality education-related evidence.” From here on, this term is operationalized to mean research that has undergone the quality control measure of peer review in that it has appeared in peer-reviewed journals. The term also includes research and expertise associated with academic experts. Such researchers are most likely to publish in peer-reviewed journals and work for employers that require advanced graduate degrees in their areas of research.1 While this definition is far from perfect and open to debate, this focus on university and peer-reviewed research and expertise is also important because most academic research is either in whole or in part supported by public funds in that its authors work at public universities or receive government con- tracts or grants. This adds a layer of urgency to the need to understand the degree to which such research and expertise is disseminated to the public that has helped pay for it and is, indi- rectly at least, responsible for ensuring its continuing support. In order to provide a better understanding of the prevalence of such research and expertise, this study also examines the journalistic decision-making processes that might help explain why educa- tion writers often hesitate to mention research and expertise from peer-reviewed and/or academic sources.

Literature Review

Relatively little academic research addresses news media coverage of education research and expertise, much less academic research and expertise. One of the few exceptions is Jeffrey Henig’s 2008 book, Spin Cycle. Henig’s study includes an analysis of the affili- ations of the experts whom two large newspapers consulted for articles on school choice. Nearly half of the experts mentioned (46.3%) were university affiliated. Of the 12 experts mentioned more than three times, two thirds (8) were affiliated with a uni- versity for at least some point during the sample period (1980– 2004). The study also included interviews with journalists from elite organizations. Henig concluded that these journalists lacked the knowledge to sort the research wheat from the chaff. Further, they were skeptical about education research in general and char- ter school research in particular in part as a result of method- ological disagreements between high-profile researchers. They were also overwhelmed as a result of deep staffing cuts in their newsrooms.

Henig (2008) did not analyze research affiliations or news coverage unrelated to school choice. In a broader study of the education coverage of three elite publications (The New York Times, The Washington Post, and Education Week), Yettick (2009) found that Education Week articles most frequently cited univer- sity research, whereas the two newspapers most frequently cited government research, followed by university research. This study did not include an analysis of expert citations or interviews with journalists. Further, neither study examined online-only outlets

or local publications, which may be predisposed to cover more education news since, in the United States, K–12 schooling is largely funded and controlled at state and local levels.

In their landmark overview of media coverage of social sci- ence research (including education research), Weiss and Singer (1988) found that elite, national news outlets were most likely to mention government studies but that university research was a close second. They also found that social science was the subject of 4 of the top 10 stories generated by meetings of the American Association for the Advancement of Science, meaning that social science attracted a disproportionate amount of media attention. However, the outlets studied completely ignored the research presented in five top peer-reviewed social science journals.

By contrast, a research synthesis of qualitative studies of health and science reporters found that both types of journalists relied heavily on scientific journals, which were their main source for story ideas, in part because they were skeptical of pub- lic relations efforts and trusting of scientists (Amend & Secko, 2011). In line with these results were the findings of a quantita- tive study that found that more than half of front page newspa- per articles about medical research (57%) were based on studies published in peer-reviewed journals (Lai & Lane, 2009).

Overall, research suggests journalists cover social science at major academic conferences (Weiss & Singer, 1988). They favor peer-reviewed studies of science and health (Amend & Secko, 2011; Lai & Lane, 2009). In some instances, they favor university- based education research (Yettick, 2009). However, an overarch- ing issue is that the news media, in general, mention very little peer-reviewed research or expertise of any kind (Henig, 2008). One reason may be that peer-reviewed education research has long been criticized as low status, irrelevant, and trivial, to the point that even practitioners in the field (i.e., teachers) often felt justified in ignoring it (Lagemann, 2000). Yet even in the health field, where the news media have arguably helped transform cer- tain peer-reviewed journals into household names, 0.34% of peer-reviewed studies attracted attention from the news media (Suleski & Ibaraki, 2010). As Suleski and Ibaraki write,

If the output of science articles were the volume of a swimming pool, the total papers that made it to a mainstream audience through news media would fill only a quart, and the non-health/ medicine papers would be just two tablespoons. (p. 120)

The results of this study suggest that in the education field, the medicine dropper might be a more appropriate instrument of measurement.

Conceptual Framework

The conceptual framework for this study is gatekeeping theory. Gatekeeping theory provides a model for how and why decisions are made. In essence, everyone is not equally important when it comes to mass social decisions, such as the foods a family con- sumes or the articles that appear in a newspaper (Lewin, 1997; White, 1950). Instead, decision making flows through one or more channels interspersed with gates. “Gatekeepers” determine whether the gates swing open or remain shut (Lewin, 1997, p. 300). In the case of communication, channels might, for instance,

ApRIl 2015 175

represent the path from a university news release to a newspaper article. Examples of gatekeepers would include the editors who sort through news releases and the reporters who decide whether to write about the sorted releases. An extensive body of mass communication research has examined gatekeeping influences, which include newsroom social control (Breed, 1955), cognitive processes (Stocking & Gross, 1989), and ideology (Gans, 2004).

More recent theoretical explanations have suggested that even in the Internet era, gatekeeping has not so much disappeared as evolved to varying degrees to incorporate more audience input (Bowman, 2008; Bruns, 2009; Cassidy, 2006; T. Haas, 2005; Livingston & Bennett, 2003; Singer, 2006; Storm, 2007). In a nod to the loosening of the editorial reigns by newer and more experimental online media outlets, Bruns (2009) proposed renam- ing gatekeeping as “gatewatching” because online journalists are not so much locking and unlocking the gates as they are watching what flows through a perpetually permeable opening that admits content from a seemingly endless supply of contributors.

Shoemaker and Vos’s Multilevel Model

A hallmark of contemporary notions of gatekeeping theory is the idea that multiple levels of analysis should be taken into account. Drawing upon more than half a century’s worth of gatekeeping research and theory, Shoemaker and Vos (2009) envision five levels that progress from narrow and individual to broad and societal: individual, communication routines, organizational, social institution, and social system/ideological.2 These levels interact and should not be considered in isolation. For instance, individuals (Level 1) put their own stamps on communication routines (Level 2).

The Shoemaker and Vos (2009) model is the conceptual framework for this research study. As such, the research questions examine the extent to which forces at various levels influence the prevalence of academic and/or peer-reviewed education research in the news media. The research questions are as follows:

1. What is the prevalence of peer-reviewed research and academic research and expertise in U.S. print news media coverage of education?

2. How do the journalistic decision-making processes help us understand the prevalence of academic and peer- reviewed research and expertise in news media coverage of education?

Past research suggests that potential influences on news cover- age of education and/or research and expertise might include per- sonal preferences, risk avoidance, and educational backgrounds at Level 1; skepticism as a news value at Level 2; reporters from com- peting outlets at Level 3; economic forces, think tanks, and audi- ences at Level 4; and neoliberalism at Level 5 (Dunwoody, 1980; Henig, 2008; Stack, 2007; Tunstall, 1971; Ungderleider, 2006).

Methodology

This mixed-methods research study sampled two populations: news media coverage and the journalists and bloggers who cre- ated that coverage. The first population consisted of U.S. print

news media coverage that mentioned education research or experts. It was unclear, based on past research, how journalists actually defined education research or expertise. So education research was broadly defined as evidence relevant to decisions about education. This definition excluded purely anecdotal statements. “Experts” were defined as researchers. Also included in the definition were non-research-conducting observers com- menting from the sidelines, in much the same way that sports commentators describe and analyze what is happening on the field in an effort to provide broader context and meaning to a game. Although these expansive definitions almost certainly included people, studies, and data that some people would not consider to be research or experts, the writers interviewed for this study generally agreed with these classifications.

This study operationalized media coverage as print news media articles, commentaries, blog entries, columns, and edito- rials about K–12 education, described from here on out as “items.” All items appeared during the first 6 months of 2010 in daily newspapers, online-only outlets, or the trade publication Education Week. The study excluded other outlet types because the print media have historically produced more education cov- erage, employed more journalists, and set the agenda for radio and broadcast news (Pew Research Center, 2013; Reese & Danielian, 1989; Weaver, Beam, Brownlee, Voakes, & Wilhoit, 2007).

Past research has largely neglected local news coverage of edu- cation even though one might expect local news coverage to be more comprehensive since K–12 education in this country is funded and controlled at the state and local level. So this study sampled content from the 650 daily newspapers, big and small, that are catalogued in the America’s News database.

In addition to newspapers, the study sampled online-only sites that either focused on education or wrote frequently about the topic. As a result of rapid and decentralized growth in the sector, it was difficult to find a comprehensive list of education blogs and other online-only outlets from which to work. For this reason, the sampled outlets were gleaned from a list of 86 education- related online-only outlets monitored regularly by Alexander Russo, who runs the news aggregation and commentary site This Week in Education. Russo chooses sites to follow “that include a variety of content, views that aren’t all predictable, and update regularly” (A. Russo, personal communication, March 20, 2010).

The final sampled text was Education Week. This periodical was included because it is an influential source of information for policymakers, academics, journalists, and others with an interest in education-related policy (Swanson & Barlage, 2006).

I used “constructed week sampling” to select daily newspaper articles and online-only texts. In a constructed week sample, each day of the sampled week is randomly selected from the entire list of those days (e.g., all the Wednesdays) that occurred during a given period. In this case, the time period was the first 6 months of 2010. This means that the Monday of the con- structed week may have occurred in January while the Tuesday occurred in June. The advantage of this method is that it accounts for the fact that news is heavier on certain days by including every single day of the week while also representing the entire range of months that occurred during the sample

176 EDUCATIONAl RESEARCHER

period. One constructed week represented 6 months of newspa- per coverage while 2 constructed weeks represented the same time period for online-only reportage, which is more variable (Hester & Dougall, 2007; Riffe, Aust, & Lacy, 1993).

On the 7 days sampled, a total of 223,400 newspaper items ran. Based on the research of E. Haas (2007), search terms such as “study,” “expert,” and “school” narrowed down these results to 40,000 items, each of which this author skimmed, selecting 1,332 newspaper items for coding because they addressed K–12 education and mentioned research and/or experts. These 1,332 items represented 0.6% of the 223,400 news items that appeared in newspapers on the 7 sampled days. They were drawn from 395 newspapers.

I did not use search terms for online-only outlets because many lacked search functions and because search engines gener- ated too many irrelevant results. Instead, this author skimmed 3,390 items that appeared on the 14 sample dates, selecting 361 items (11%) for coding. Finally, 7 randomly selected weeks rep- resented 6 months of coverage in Education Week (Lacy, Robinson, & Riffe, 1995). Because only 305 items ran in those editions, this author skimmed them all, selecting 159 items (52%) for coding.

The codes assigned to selected items are summarized in Table 1. Pursuant to methodological research on content analysis, an assistant recoded a random sample of 100 items, with Cohen’s Kappa used to gauge interrater reliability (Lacy & Riffe, 1996), discarding codes if κ < .90.

The second population sampled for the study consisted of the writers who produced the news media coverage. The selection of these 33 writers was purposive, with an eye toward gaining a deeper understanding of the results of the initial analysis of the sampled texts. Writers participated in open-ended, protocol- guided interviews that lasted 20 minutes to 2 hours apiece.

Analysis

This study employed logistic regression to analyze the 1,852 coded items. The item was the unit of analysis. The models spec- ified controlled for the fact that item length varied significantly (see Table 1) while also exploring interactions between different codes (e.g., item topic and research type). Pursuant to research suggesting differences between news coverage that mentions research and news coverage that mentions expertise, the study analyzed items that mentioned research separately from items that mentioned experts (Weiss & Singer, 1988). Items that men- tioned both research and expertise were included in both sets of analyses. Since the goal was to examine the individual parameter estimates for each variable, parsimony and model fit took a back seat to parameter estimates. Table 1 summarizes the variables employed in the models.

The analysis specified four pairs of models, with each pair consisting of items that mentioned research and items and men- tioned experts. The first pair used binomial dependent variables that indicated whether or not an item had mentioned university research or experts. The second set of models compared differ- ences between outlet types by employing a multinomial depen- dent variable that indicated whether an item had run in a

newspaper, an online-only outlet, or Education Week. A third set of models explored differences among online-only outlets because they varied widely, from one-woman blogs run by hob- byists (e.g., It’s Not All Flowers and Sausages) to professional news organizations with paid staffs (e.g., Voice of San Diego). For these models, which only examined the subset of items that ran in online-only outlets, the outcome variables indicated whether an item had come from outlets affiliated with one of five types of sponsors: associations, educators, “media model,” think tank, or “other.” Finally, a fourth set of models used a binomial depen- dent variable to examine differences between items that were commentaries (generated by readers or thought leaders) and items that were articles.

Analysis of interview transcripts entailed what LeCompte and Schensul described as “coding from the top down” by “choosing a set of concepts first and then sorting out the data in terms of which of the concepts they fit best” (1999, pp. 66, 46). The analy- sis of interviews was deductive in that transcripts were coded to one or more levels and sublevels of Shoemaker and Vos’s (2009) conceptual framework of gatekeeping influences. Once the tran- scripts were coded, each category was examined as an entity, with an eye toward identifying the influence’s scope, its various forms and manifestations, its benefits, and its drawbacks.

Results

Peer-Reviewed Research

Peer-reviewed academic journals are barely a footnote to news media coverage of education: Of the 227,095 items that ran in the sampled outlets during the sampled days, 1,558 (0.69%) mentioned education research. Of these 1,558, 45 (3%) men- tioned peer-reviewed journal research (see Figure 1). By contrast, 695 (45%) mentioned government research.

Of the three most frequently mentioned journals, two were not even from the education field, but from medicine. Further, “health” was the most common topic of news items that men- tion peer-reviewed journals: More than a third of news items that cited peer-reviewed journals addressed the “health” topic.

The most frequently mentioned education journal was Education Next. Although the publication describes itself as peer- reviewed, it is not assigned an impact factor by Thomson- Reuters. More generally, there was virtually no overlap between the education journals mentioned by the news media and the education journals with the highest impact factors. With one exception (Child Development), none of the educational journals mentioned were assigned impact factors in the top five of their 2010 Thomson-Reuters education-related categories. By con- trast, the medical journals mentioned did have high impact fac- tors (see Table 2).

No American Educational Research Association (AERA) journals were mentioned by the print news media during the sample period. Further, only 3 of the 33 interviewees said they used AERA or its annual conference as a resource. “Why I didn’t use them before—I just didn’t know about them,” said 1 of the 3. A veteran education reporter for a major metro daily newspa- per, she had only just learned of the organization’s existence.

ApRIl 2015 177

Table 1 Number of News Items Assigned to Each Code

Code Frequency Example

Institutional affiliation: Item mentions at least one: Examples from coded items Government study 695 A Tangipahoa Parish School Board report Study of unknown origin 438 Research clearly shows … University study 247 A University of Texas at Austin study Think tank study 230 An evaluation by the American Institutes for Research News media study 185 A 2008 Chicago Sun-Times series, called Schooled in Fear Association study 98 A report released by the National Association of Charter School Authorizers For-profit study 93 Study from the Evergreen Education Group, a consulting group based in

Evergreen, Colorado Foundation study 89 A report issued this week from the Annie E. Casey Foundation University research center study 84 Institute for Research on Education Policy and Practice at Stanford University Peer-reviewed journal study 45 A new study … in Education Next Study by an unaffiliated/independent research 34 Author Dan Pink Union study 21 A United School Employees of Pasco survey Government expert 742 Associate superintendent for curriculum, instruction, and accountability University expert 406 An education professor at Stanford Think tank expert 248 The vice president of national programs and policies for the Washington-

based Thomas B. Fordham Institute Association expert 228 A former president of the New York chapter of the National Association of

School Psychologists For-profit expert 170 An educational consultant who has worked with school districts throughout

Texas Expert with unknown affiliation 137 Experts say News media expert 87 Jay Matthews advises us that, based on recent research University research center expert 87 Associate director of the Center on Reinventing Public Education at the

University of Washington Foundation expert 64 Bush Foundation President … released the group’s report Unaffiliated expert 61 Susan Ohanian Union expert 51 President of the American Federation of Teachers

Outlet types: Item ran in: Examples of outlets An online-only outlet 361 National Journal Education Experts Blog Education Week 159 Education Week A newspaper 1332 The Orlando Sentinel Newspaper items: Newspaper size Examples of newspapers Small (circulation < 100,000) 794 The Dalton Daily Citizen Medium (circulation 100,000–499,999) 462 The Denver Post Large (circulation > 499,999) 76 The New York Times Headline example Item is Education Week or newspaper commentary 126 Charters: Students With Disabilities Need Not Apply? Online-only outlets: Outlet sponsor Outlet examples Association-sponsored outlet 43 BoardBuzz Educator-sponsored outlet 67 Sherman Dorn Media model outlet 147 Inside School Research Other outlet 45 Intercepts Think tank–sponsored outlet 59 The EnterpriseBlog Item topic (each item can be assigned to multiple topics)

Headline examples Politics and governance 455 Race for education dollars—Maryland legislators should heed the state

superintendent’s advice.

(continued)

178 EDUCATIONAl RESEARCHER

Even online-only outlets with university sponsors (mainly faculty bloggers) gave short shrift to peer-reviewed education research. Peer-reviewed journals were mentioned in less than 5% of the university-sponsored outlets’ coded items that mentioned research. This means that faculty member–sponsored outlets mention peer-reviewed journals at rates similar to the sample as a whole.

Why don’t education writers mention more peer-reviewed education research? Interestingly, one obvious obstacle was not a significant barrier: Just 5 of the 33 study participants said they avoided peer-reviewed education journals because of the cost associated with downloading articles from journals that erect pay walls. Rather, the stumbling blocks were more basic: For instance, this author was forced to revise the interview protocol to define peer review because 9 people were uncertain what it meant. This was likely related to the background of the interviewees, which was typical of American journalists: Although all interviewees were college graduates, just 20% had pursued majors in the social sciences. The remainder pursued journalism and/or humanities- related majors. This is likely also why nearly half the interviewees (n = 15) found peer-reviewed journals difficult to understand. In fact, difficulty understanding peer-reviewed journals was one of two top reasons why interviewees avoided citing peer-reviewed education research. A veteran newspaper reporter explained,

I don’t have enough personal expertise to know what I’m looking at. … I sort of went fishing. I don’t know whether this is credible research or not. … I don’t know if it’s groundbreaking or something that says what another study says. I found it to be diminishing returns.

In the wake of industrywide financial pressures that leave little room for any activity with the possibility of providing diminishing returns, few study participants had even tried such a foray into the world of peer review (Pew Research Center, 2011, 2013).3 So it makes sense that time constraints were the other top reason why interviewees did not mention peer-reviewed journals. One educa- tion reporter at a newspaper that had recently undergone layoffs was among the 16 interviewees who lacked time to decipher peer- reviewed research. He said, “I don’t know what would happen if my editor came over and I was just reading a magazine.”

Another obstacle for interviewees was the perception that peer-reviewed education journals did not publicize their articles

1,852 of 227,095 items that ran

1/1/10-6/30/10 were coded

because they men�oned ed res.

and/or experts (.8%)

159 Edweek items coded (52%)

137 items men�on research

8 items men�on p.r. jrnls (6%)

41 items men�on univ. res. (30%)

124 items men�on experts

60 items men�on univ. exp. (48%)

102 items men�on exp. AND res.

361 online-only items coded (11%)

281 items men�on research

12 items men�on p.r. journals (4%)

67 items men�on univ res (24%)

250 items men�on experts

109 items men�on univ. exp. (44%)

170 items men�on exp. AND res.

1,332 newspaper items coded (.6%)

1,140 items men�on research

25 items men�on p.r. journals (2%)

139 items men�on univ. res. (12%)

1,060 men�on experts

237 items men�on univ. exp. (22%)

868 items men�on exp. AND res.

FIGURE 1. Numbers and percentages of items that mentioned education research and expertise Peer-reviewed journals were only coded as “research” because a peer-reviewed journal “expert” category would imply that the person worked for a peer-reviewed journal.

Code Frequency Example

Money 353 South Adams seeking further cuts in budget Assessment & accountability 269 New York State places dozens of NYC schools on replacement list Health 258 What if schools couldn’t serve chocolate milk? Personnel 255 Union’s offer underwhelms Teaching and learning 255 Momentum Building for Hands-On Science Learning Special populations and equity 235 CUSD tackles equality in sports Parents and community 199 New NORD leaders in tough spot—Low funding limits summer programs Legal and discipline 184 Bullying May Be Decreasing, Survey Finds Technology 87 8-hour school day?—Plan would give some students extra 2 hours of

computerized math, reading Miscellaneous Headline example Item was focused on research 572 Second Study Gives Thumbs Up to N.Y.C. Charters Locally focused outlet example Item ran in a locally focused outlet 1348 Voice of San Diego Average number of words per item 643

Table 1 (continued)

ApRIl 2015 179

with news releases (n = 8). Even interviewees who regularly perused peer-reviewed journals found it more efficient to rely on news releases rather than wade through articles. Interviewees selected because they had made the rare decision to mention peer-reviewed studies said they found the research they had mentioned as a result of publicity efforts instigated by the researcher or via search results featured prominently on Google. More generally, Internet search engines were the main way in which interviewees found academic research.

Certainly, when it comes to many of the obstacles described in this section, science and medical reporters face many of the same constraints as education writers. Yet they quote peer- reviewed journals more often. Interviewees believed that jour- nals in those fields more aggressively publicized their articles. But a veteran education reporter for a daily newspaper also noted differences between the education and medical fields: “I think medical research is on its face more interesting. We’ve had great steps forward. The world [has been] changed by medical research. I don’t think we can say the same for education research.” This may, of course, be a self-fulfilling prophecy. In ignoring peer- reviewed education research, journalists may also be limiting its potential to influence the broader world.

Universities

Universities serve a core research function. So one might expect them to be the top source of the education research that reaches

the public via the news media sampled for this study. They are not. They are a distant third, behind government (45%) and “unknown research,” a category that includes all research cita- tions that did not mention a person or institutional affiliation (28%).4 University research and think tank research are men- tioned at nearly the same rates (16% vs. 15%). This is the case even though universities produce 14 to 16 times more research than think tanks (Yettick, 2009). Further, most of the think tanks mentioned (87%) were advocacy-oriented think tanks rather than organizations like RAND that embrace a more aca- demic approach in that they strive to conduct empirical research and do not seek to promote a specific cause.

Audiences and thought leaders who submit commentaries (i.e., op-eds) mention university research most: Commentaries in newspapers and Education Week are twice as likely as articles to mention university studies.

Online-only outlets are nearly twice as likely as newspapers to mention university research. One reason is that newspapers are so thoroughly focused on government that other types of research and expertise get squeezed out. More than 64% of coded newspaper items mentioned at least one government expert and/or study.

Like other types of online-only outlets, educator-sponsored media outlets gravitate toward research and expertise that matches their sponsor type. As such, universities are a particu- larly important source for educator-sponsored outlets: Educator- sponsored outlets are three times more likely to mention

Table 2 Complete List of Peer-Reviewed Journals Mentioned in Study

Sample, by Number of Media Citations, Impact Factor, and Field

Journal Name Total Cites Impact Factor Field

Pediatrics 7 5.4 Pediatrics Education Next 5 NR N/A Archives of Pediatric and Adolescent Medicine 4 4. 3 Pediatrics Developmental Psychology 3 3. 4 Psychology, developmental Educational Leadership 3 0. 2 Education and education research New England Journal of Medicine 3 53 Medicine, general and internal Health Affairs 2 3. 8 Health policy and services Journal of the American Medical Association 2 30 Medicine, general and internal Kappan 2 0. 2 Education and education research American Journal of Health Behavior 1 1. 2 Public, environmental, and occupational health British Journal of Educational Psychology 1 1 Psychology, educational Child Development 1 3. 8 Psychology, educational Current Directions in Psychological Science 1 3. 5 Psychology, multidisciplinary Economics of Education Review 1 1. 1 Education and education research Elementary School Journal 1 1. 1 Education and education research International Journal of Science Education 1 1. 1 Education and education research Journal of Adolescent Health 1 3. 1 Psychology, developmental Journal of Education for Students Placed at Risk 1 NR N/A Journal of Human Resources 1 2. 1 Economics Literacy Research & Instruction 1 NR N/A Music Educators Journal 1 NR N/A Psychological Science and the Public Interest 1 NR N/A

Source. 2010 Thomson-Reuters Social Science Edition. Note. NR = not ranked by Thomson-Reuters; N/A = not applicable. Impact factor and field are provided by Thomson-Reuters.

180 EDUCATIONAl RESEARCHER

university research and experts than are the “media model” out- lets that dominate the online-only sample.

However, when the “educator-sponsored” category is divided into K–12 educators and university educators, it becomes clear that university-sponsored outlets prioritize university research more than outlets run by K–12 educators. About 40% of the coded, research-citing items from university educator–sponsored outlets mention university research as compared to 16% of the items drawn from K–12-sponsored outlets. This gap makes sense when one considers that, unlike in the life sciences, practi- tioners in the education field (i.e., teachers) rely more on practi- cal experience than research (National Research Council, 2002).

In the sample as a whole as well as in all three outlet types, universities were much more frequently called upon to provide educational expertise than research: Among coded items that mentioned educational experts, 28% cited university-affiliated experts. Among coded items that mentioned research, 16% cited university research (see Figure 1).

As a source of educational expertise in the sample as a whole, universities are second only to government. However, online- only outlets and Education Week mention university experts about twice as much as newspapers.

Interview results helped explain the underrepresentation of university research, especially as compared to university experts. As with other sources, news values were the standard against which this type of research and expertise were judged. Localism (proximity) and objectivity were the two news norms most com- monly employed to judge both academic research and experts and other research/expert types. Unlike Henig (2008), this study did not find that interviewees were skeptical of education research as a result of high-profile academic disagreements, such as the debate between Caroline Hoxby and Margaret Raymond, which was covered extensively by some outlets. (See, e.g., Viadero, 2009.) In fact, most were unaware of such issues. One of the few interviewees to even mention the Hoxby debates

dismissed them as “insider baseball” irrelevant both to him and to the readers of his locally focused news outlet. More generally, although a handful of interviewees expressed deep skepticism or scorn related to the field of education research, the predominant view was more along the lines of this response: “That’s like ask- ing me what I think of politics. There is good education research, lousy education research; influential, not influential. … I see lots of different patterns, not one.”5 The differences between my findings and Henig’s are likely due in large part to my decision to focus on a wide range of education writers rather than restrict- ing my sample to affiliates of elite, national outlets. Local writers were simply too distracted by local affairs to spend a great deal of time pondering the nature of education research or following scholarly debates, high profile or not.

Nine interviewees preferred to mention research and experts associated with local institutions. The news value “elitism” was localism’s kryptonite, with reporters sometimes willing to inter- view an out-of-state professor if he or she was an Ivy League pro- fessor. Perhaps as a result, most of the universities most frequently mentioned as sources of education research and expertise are elite, Research I institutions (see Table 3). Unlike elitism, other news values acted in concert with localism: For instance, a local research study that was in line with news values emphasizing clarity was preferable to a local study that was more ambiguous.

Journalists’ emphasis on localism may help explain why uni- versity research is less frequently mentioned than expertise: The human subjects rules that faculty members must follow may lead researchers to disguise the locations of their study sites. But the same researchers are not similarly constrained when it comes to providing expert commentary on, for instance, a proposed pol- icy. By contrast, when the local school district issues a report, journalists can rest assured that it is a study of local schools. Given that nearly every newspaper sampled had a specific local geographic focus, it is unsurprising that government research and expertise dominated newspapers.

Table 3 Universities Most Frequently Mentioned by the News Media as Sources of Education Research and Expertise

Name Items That Mention Research Name Items That Mention Experts

Stanford 19 Harvard 31 Harvard 17 Stanford 31 New York University 14 New York University 30 University of Chicago 10 Columbia University 15 University of California, Santa Barbara 8 University of Illinois at Chicago 15 University of Illinois at Chicago 6 University of Chicago 12 University of Virginia 6 University of Virginia 11 Johns Hopkins 6 University of California, Los Angeles 10 University of California, Los Angeles 5 University of Arkansas 9 University of Arkansas 5 University of Washington 9 University of Washington 5 University of Wisconsin–Madison 9 University of Wisconsin–Madison 5 University of California, Berkeley 8 University of California, Berkeley 5 University of Michigan 8 University of Maryland 5 University of Minnesota 8 Florida State University 5 University of Missouri, Columbia 7

ApRIl 2015 181

The extreme form of localism emphasized at newspapers and hyper-local online-only outlets has expanded in the past half a decade. It results from efforts to increase advertising revenue and stem reader losses by focusing on local news that is not available from other outlets. Such localism has led newspapers to shutter Washington bureaus as they (and competing, hyper-local online- only outlets) devote resources to youth sports scores, police led- gers, and other minutiae of provincial interest (Pew Research Center, 2011). This approach frustrated interviewees, nearly half of whom reported that localism made it difficult to mention any research or expertise whatsoever because the material they reported was too particular to a specific locality to lend itself to a broader perspective.

Objectivity was another commonly cited news value used to select and judge academic research and experts. Study partici- pants rarely read the research to assess bias. Rather, they used proxies to signal objectivity. Institutional affiliation (n = 19) was the most commonly used signal, meaning some institutions or types of institutions were viewed as objective, and some were not. Funding was the most common means of deciding whether an institution was objective (n = 8). Another signal of objectivity was research that disconfirmed initial or past hypotheses or claimed to lack a hypothesis (e.g., “I began the study with no idea of what I would find”).

Experts were more likely to be considered objective if their comments were middle of the road politically and/or they acknowledged the validity of alternative views. Only a small minority of interviewees (n = 5) said they tried to create objectiv- ity by providing contrasting research conclusions and/or view- points, creating a situation in which certain types of views (i.e., moderate) are privileged over others, regardless of what the pre- ponderance of evidence might suggest.

Six interviewees placed academic research and experts in the “objective” category. Six found them biased. Even those who found university research and experts to be objective tended to qualify their statements. “University studies seem to be more neutral,” was a typical comment. Objectivity was viewed on a sliding scale: To many interviewees, every source of education research and expertise was biased, but some sources were more biased than others.

One education writer proposed that even when biased, aca- demic research was biased in a more harmless manner than other research and expert types: “The main source of bias, if I want to call it bias, has to do with the topic of research.” For instance, a charter school study could investigate parental satis- faction, or it could examine racial stratification. The first exam- ple reflects an interest in choice as a driver of liberty interests, while the second example reflects an interest in choice’s impact on equity interests.

Another interviewee was more troubled by her perceptions of academic bias, saying that professors had a tendency to suggest that a research body was inadequate or nonexistent when the cumulative results contradicted their political or ideological beliefs:

I remember a … professor who teaches reading was asked: “What is it we really know about teaching reading?” She said, “Nothing.” We know a lot! Even when something is known, we

don’t know anything if it disagrees with our worldview … or politics.

Localism and objectivity were not the only news values that influenced the gatekeeping surrounding university research and expertise. Seven interviewees also said that academic research and expert comments were difficult to adapt to news values emphasizing simplicity and/or clarity.

In addition to being influenced by news values, university expert and research selection were also affected by interviewees’ limited knowledge of statistics and research methods. The subset of interviewees who desired more training in these subjects did not want more formal education. Instead they preferred a form of on-the-job learning in which they cultivated journalistic sources as personal tutors. Not surprisingly, these “tutors” were often affiliated with universities. An outcome of the resulting “teacher-student” relationship was that faculty members some- times became trusted “super-sources” who shaped coverage in ways that were both highly visible (e.g., via frequent quotations in articles) and more difficult to detect (e.g., by providing input on which stories to cover). This is perhaps one of the reasons 6 interviewees described academic experts as trustworthy. Trust was a risk-management technique for individual interviewees in that relying on trusted sources reduced the chance of making embarrassing and potentially career-damaging errors.

As such, university super-sources could have an outsized influence on coverage of education research and expertise. For example, one young education writer turned to a super-source professor when she found herself unable to quell what she viewed as her editor’s misguided enthusiasm for publishing teachers’ names alongside value-added scores based on student assess- ments. This writer invited her super-source to lunch, where he proceeded to persuade her editor to take a different approach. In a sense, he became an adjunct journalist.

Discussion

The results of this study suggest that peer-reviewed academic edu- cation research and the AERA organization are barely a blip on the radar of American education reporting. While university research and experts are more prominently featured in news coverage, their role is less extensive than one might expect, given that few other types of organizations devote as many resources to research.

It is important to note that the disconnect between academic education research and the news media is due, in large part, to forces external to journalism, which are found at Level 3 of the Shoemaker and Vos (2009) framework. First, purveyors of peer- reviewed education research have been less proactive with pub- licity than those in the science and medical fields, where news coverage of academic and peer-reviewed research is more com- mon. For example, in a 2010 AERA presentation, veteran higher education editor Scott Jaschik described being “bombarded” with “dozens and dozens” of university press office requests and briefings related to (non–social science) research scheduled to be presented at the annual meeting of the American Association for the Advancement of Science (Baker, 2010). By contrast, Jaschik reported receiving just two university emails publicizing studies to be presented at that year’s AERA annual meeting.

182 EDUCATIONAl RESEARCHER

This lack of publicity on the part of universities may help explain the finding that think tanks, which do publicize their research, generate relatively more news per study than do aca- demics (Rich, 2004). An additional reason is that think tanks generally aim to influence policy by producing research that is relevant to issues of public interest, which are often well aligned with topics of news coverage (Rich, 2004). By contrast, academ- ics have a different set of incentives (e.g., earning tenure, publi- cation in peer-reviewed journals) that lead them to prioritize studies that aim to advance the state of knowledge in their fields in ways that are not necessarily relevant to the public, much less newsworthy.

Yet even if universities wanted to publicize the subset of edu- cation research that is relevant to the public, they might not have the money: In 2011, American universities spent more than $65 billion on research and development, the majority of which they received from federal grants and contracts (Britt, 2012). More than half of these funds ($37.2 billion or 57%) went to the life sciences, mainly the medical sciences ($20.4 billion or 31%). By contrast, education research received $1.104 billion, or less than 2% of total expenditures.

Yet adequate education research may actually be more expen- sive to implement even than some medical trials since relevant studies can rarely be conducted in controlled, laboratory envi- ronments. Seldom do they result in the discovery of one simple, uncontested main effect: Rather, unintended consequences and context-related interactions muddy the waters, limiting the cer- tainty with which claims can be made (National Research Council, 2002). This is problematic for journalists, who may find that the qualifications inherent to the study of human sub- jects in nonlaboratory contexts result in a nuanced interpreta- tions that can seem like a simple lack of objectivity on the part of the researcher. Certainly, perceptions of subjectivity were problematic for participants in this study. A further issue is that the more clear-cut results that emerge from randomized, con- trolled, laboratory-based experiments permit medical researchers to make claims with a relatively high degree of certainty com- pared to researchers in education. Such straightforward claims align with news values (mentioned by interviewees for this study) emphasizing clarity, simplicity, and the need to write with firm authority.

Adding to the confusion is, as Henig (2008) notes, educa- tion’s lack of a small set of peer-reviewed, flagship journals such as the Journal of the American Medical Association or the New England Journal of Medicine that clearly signal to journalists that the studies contained within are groundbreaking and method- ologically rigorous.

Journalists’ perceptions that education is relatively subjective may also have a basis in the nature of the field. Values have long played a central role in the discipline because the education of young people is tied to hopes and expectations for the future of society (National Research Council, 2002). As a result, policy- makers, practitioners, parents, and even researchers find it diffi- cult to give research findings too much weight if they conflict with deeply ingrained expectations and beliefs.

Despite these important (discipline-based, Level 3) differ- ences between the fields of education and the life sciences, news values found at Level 2 of the Shoemaker and Vos (2009)

framework also appear to explain key aspects of the disconnect between education writers and academia. In many cases, jour- nalists simply screen out research that is not associated with their direct locality or, in the case of some academic research, in which the location has been disguised. They assess research and experts based on the application and misapplication of journalistic norms of objectivity and bias and simplicity and clarity. In this way, news values interact with the Shoemaker and Vos Level 1 influence of educational background in that interviewees often used news values as “signals of quality” because they lacked the skills to evaluate research and expertise using more appropriate methodological standards.

Conclusion

Just as education research is unlikely in the foreseeable future to be funded at the levels of medical research or to become less context driven, news values are not going anywhere. If university researchers wish to reach the public via the news media, they need to figure out how to demonstrate that their work is relevant to specific localities. They need to embrace simplicity and clarity by writing lucid, concise news releases about their work. Further, journalists have education levels similar to those found in other populations (e.g., school board members, K–12 teachers) that make important, direct decisions about schools. This suggests a more universal need for the accessible communication of research since statistical terms and results that are inscrutable to journal- ists may be just as inscrutable to these other key stakeholders. There are almost certainly lessons to be learned from the strate- gies of peer-reviewed publications such as Education Next, the Journal of the American Medical Association, and the New England Journal of Medicine that have successfully penetrated the public (and news media) consciousness. Future research might also examine whether there are journals that have communicated, in an accessible manner, less decisive findings of education research, which often generates results that are more context dependent and nebulous than those associated with the physical sciences or medicine.

News values emphasizing objectivity and bias are more diffi- cult to overcome in that it is not necessarily desirable for aca- demics to stick to middle-of-the-road comments when their research suggests more extreme statements or limit themselves to studies that lack or contradict hypotheses. Such notions involve, in a sense, the misapplication of journalistic values to social sci- ence research. They represent an opportunity and a need for pro- fessional development.

Given interviewees’ preference for on-the-job learning facili- tated by expert sources, faculty members with an interest in doing so have a golden opportunity to help fill the gaps in the educational backgrounds of the journalism and English majors who report most of the news about our schools. Here, too, jour- nalism schools would appear to have a role by doing more to instruct students on how to incorporate social science research results into the coverage not only of education but of crime, gov- ernment, and other areas of broad public interest. In doing so, journalism educators would help their graduates—the majority of whom will probably work for small, geographically focused outlets—add meaning and perspective to their coverage rather

ApRIl 2015 183

than presenting local news as an unending series of unrelated, idiosyncratic events. In the physical, environmental, and life sci- ences, this movement has already begun in that numerous uni- versities offer science and medical journalism degrees that incorporate research communication. Universities even offer sports writing specialties. Yet this author is aware of no similar programs devoted to education. This is despite the fact that edu- cation has traditionally been a bread-and-butter beat in American newsrooms, with the Education Writers Association (the field’s main professional organization) boasting 1,591 current journal- ist members (Carlson & Roy, 2014).

A final recommendation is that the general interest news media may always have a relatively limited interest in education research and expertise because they serve audiences with interests ranging from education to NFL football. Even if the amount of research mentioned were to increase to double the rate found by this study, it would still constitute only 1.2% of total coverage. By contrast, specialist publications are more amenable to research and experts: More than half of Education Week’s coverage men- tioned education research and/or expertise. Interview results suggest the publication has incorporated increasing amounts of research in the wake of a Spencer Foundation grant that helped fund an education research beat. Far from alienating readers, this type of coverage appears to fulfill audience needs: Of the 20 most frequently viewed articles on Education Week’s website in 2010, 9 went beyond merely mentioning education research to focus on it. This suggests that there may be even more opportu- nities for outreach. This outreach has the potential to influence other journalists since Education Week was study participants’ most frequent news source of information on education research and expertise.

Even specialist news media may have a limited appetite for education research. Yet social media platforms provide an ever- expanding array of opportunities for researchers to directly address members of the public. One of the easiest ways to increase the amount of peer-reviewed education research that is publicly available would be for university-based bloggers to mention it more often.

Given that barely a medicine dropper full of peer-reviewed research currently makes it into the pool of news media educa- tion coverage, it would likely take only a small droplet of effort to exponentially increase the dosage of peer-reviewed education research and academic expertise that is readily and publicly avail- able to practitioners, parents, policymakers, and voters who use that information to make crucial decisions about our schools.

NoTES 1This definition of higher quality education evidence is far from

perfect. Peer review does not guarantee high-quality research. Nor are peer-reviewed journals and academia the only source of high-quality research. However, peer review does provide a measure of quality con- trol. Similarly, so does university affiliation, which, at the very least, tends to require a graduate degree, usually at the doctoral level, in a subject relevant to the research topic. Certainly, other types of organiza- tions such as school districts or non-advocacy-oriented think tanks may employ staff researchers who possess such degrees, but the credential- ing is generally more rigorous in academia. However, E. Haas (2007) finds that advocacy-oriented think tank affiliates who are described as

experts in news media accounts often possess no more than a bach- elor’s degree in a topic unrelated to the research topic while also lacking work experience in the area. A caveat, of course, is that possessing an advanced degree in the research area does not guarantee a high-quality study. However, on a more practical level, for the purposes of this study, it was necessary to use a quality measure that could be discerned from news coverage: Institutional affiliations and names of peer-reviewed journals are commonly mentioned in news coverage that cites research and expertise, whereas detailed explanations of research design are rare. In fact, often news accounts provide so little information even on insti- tutional affiliation that it is difficult if not impossible to track down the specific study or studies being referenced.

2In their book, which is as much a literature review as a theory- building exercise, Shoemaker and Vos (2009) draw upon and describe more than half a century’s worth of empirical studies, most of it within the communication field. This research deeply informed the study described in this article.

3Only 4 interviewees regularly browsed peer-reviewed educa- tion journals, looking for sources and ideas. All 4 were associated with education-focused outlets.

4“Unknown research” was largely a phenomenon of commentaries written by readers and thought leaders. It is mentioned in more than half the commentaries found in Education Week and in newspapers. (Many of the online-only outlets sampled blur the distinction between opinion in news, so they are excluded from analyses that break down news items into articles and commentaries). Unknown research is two times more common in commentaries than in articles.

5Out of the 33 total interviewees, 21 made at least one positive comment about education research, whereas 25 made at least one nega- tive comment.

REFERENCES

Amend, E., & Secko, D. M. (2011). In the face of critique: A meta- synthesis of the experiences of journalists covering health and science. Science Communication, 34(2), 241–282. doi: 10.1177/1075547011409952

Baker, P. (Producer). (2010). Scott Jaschik, part 1 of 2. Retrieved from http://www.youtube.com/watch?v=PRads2ZVqhY

Bowman, L. (2008). Re-examining “gatekeeping.” Journalism Practice, 2(1), 99–112. doi: 10.1080/17512780701768550

Breed, W. (1955). Social control in the newsroom: A functional analy- sis. Social Forces, 33(4), 326–335.

Britt, R. (2012). Universities report highest-ever R&D spending of $65 billion in FY 2011 (InfoBrief). Arlington, VA: National Center for Science and Engineering Statistics, National Science Foundation.

Bruns, A. (2009). Gatewatching: Collaborative online news production. New York, NY: Peter Lang.

Carlson, C., & Roy, M. (2014). Mediated access: Education writ- ers’ perceptions of public information officers’ media control efforts. Retrieved from http://www.spj.org/pdf/sunshineweek/ewa-survey- report-2014.pdf

Cassidy, W. (2006). Gatekeeping similar for online, print journalists. Newspaper Research Journal, 27(2), 6–23.

Croteau, D., Hoynes, W., & Milan, S. (2012). Media society: Industries, images, and audiences. Los Angeles, CA: Sage.

Dunwoody, S. (1980). The science writing inner club: A communica- tion link between science and the lay public. Science, Technology & Human Values, 5(3), 14–22.

Gamson, W. (1992). Talking politics. New York, NY: Cambridge University Press.

Gans, H. (2004). Deciding what’s news: A study of CBS Evening News, NBC Nightly News, Newsweek and Time. 25th Anniversary Edition. Evanston, IL: Northwestern University Press.

184 EDUCATIONAl RESEARCHER

Gerbner, G., Gross, L., Morgan, M., & Signorielli, N. (1982). Political correlates of television viewing. Public Opinion Quarterly, 48(1), 283–300.

Graber, D. (2009). Mass media and American politics. Washington, DC: Congressional Quarterly Press.

Haas, E. (2007). False equivalency: Think tank references on education in the news media. Peabody Journal of Education, 82(1), 63–102.

Haas, T. (2005). From “Public Journalism” to the “Public’s Journalism”? Rhetoric and reality in the discourse on weblogs. Journalism Studies, 6(3), 387–396. doi: 10.1080/14616700500132073

Henig, J. (2008). Spin cycle: How research is used in policy debates: The case of charter schools. New York, NY: Russell Sage Foundation and The Century Foundation.

Hester, J., & Dougall, E. (2007). The efficiency of constructed week sampling for content analysis of online news. Journalism and Mass Communication Quarterly, 84(4), 811–824.

Howell, W. (2008). Education policy, academic research, and public opinion. In F. Hess (Ed.), When research matters: How scholar- ship influences education policy (pp. 135–153). Cambridge, MA: Harvard Education Press.

Lagemann, E. (2000). An elusive science: The troubling history of educa- tion research. Chicago, IL: University of Chicago Press.

Lacy, S., & Riffe, D. (1996). Sampling error and selecting inter-coder reliability samples for nominal content categories. Journalism & Mass Communication Quarterly, 73(4), 963–973.

Lacy, S., Robinson, K., & Riffe, D. (1995). Sample size in content anal- ysis of weekly newspapers. Journalism and Mass Communication Quarterly, 72(2), 336–345.

Lai, W., & Lane, T. (2009). Characteristics of medical research news reported on front pages of newspapers PLos ONE, 4(7). doi: doi:10.1371/journal.pone.0006103

LeCompte, M., & Schensul, J. (1999). Analyzing and interpreting eth- nographic data. Walnut Creek, CA: Altamira Press.

Lewin, K. (1997). Resolving social conflicts; and, Field theory in social sci- ence. Washington, DC: American Psychological Association.

Livingston, S., & Bennett, W. L. (2003). Gatekeeping, index- ing, and live-event news: Is technology altering the construc- tion of news? Political Communication, 20(4), 363–380. doi: 10.1080/10584600390244121

National Research Council. (2002). Scientific research in education. Washington, DC: National Academy Press.

Pew Research Center. (2011). The state of the news media: An annual report on American journalism. Washington, DC: Pew Research Center’s Project for Excellence in Journalism.

Pew Research Center. (2013). The state of the news media 2013: An annual report on American journalism. Washington, DC: Pew Research Center’s Project for Excellence in Journalism.

Reese, S., & Danielian, L. (Eds.). (1989). Intermedia influence and the drug issue: Converging on cocaine. Hillsdale, NJ: Lawrence Erlbaum.

Rich, A. (2004). Think tanks, public policy, and the politics of expertise. New York, NY: Cambridge University Press.

Riffe, D., Aust, C., & Lacy, S. (1993). The effectiveness of random, consecutive day and constructed sampling in newspaper content analysis. Journalism Quarterly, 70(1), 133–139.

Rosenstiel, T., Mitchell, A., Purcell, K., & Rainie, L. (2011). How people learn about their local community. Washington, DC: Pew Research Center’s Project for Excellence in Journalism.

Shoemaker, P., & Vos, T. (2009). Gatekeeping theory. New York, NY: Routledge.

Singer, J. (2006). Stepping back from the gate: Online newspaper editors and the co-production of content in campaign 2004. Journalism and Mass Communication Quarterly, 83(2), 265–280.

Stack, M. (2007). Constructing “common sense” policies for schools: The role of journalists. International Journal of Leadership in Education, 10(3), 247–264. doi: 10.1080/13603120701373649

Stocking, H., & Gross, P. (1989). How do journalists think? A pro- posal for the study of cognitive bias in newsmaking. Bloomington, IN: ERIC Clearinghouse on Reading and Communication Skills: Smith Research Center, Indiana University.

Storm, J. (2007, May). The endurance of gatekeeping in an evolving newsroom: A multimethod study of web-generated user content. Paper presented at the International Communication Association, San Francisco, CA.

Suleski, J., & Ibaraki, M. (2010). Scientists are talking, but mostly to each other: A quantitative analysis of research represented in the mass media. Public Understanding of Science, 19(1), 115–125. doi: 10.1177/0963662508096776

Swanson, C., & Barlage, J. (2006). Influence: A study of factors shaping education policy. Washington, DC: Editorial Projects in Education Research Center.

Thomson Reuters (Ed.). (2010). Web of Science Social Sciences Citation Index. Philadelphia, PA: Thomson Reuters.

Tunstall, J. (1971). Journalists at work: Specialist correspondents: Their news organizations, news sources, and competitor-colleagues. Beverly Hills, CA: Sage.

Ungderleider, C. (2006). Government, neo-liberal media, and educa- tion in Canada. Canadian Journal of Education, 29(1), 70–90.

Viadero, D. (2009, October 14). Scholars spar over research methods used to evaluate charters. Education Week, pp. 6–7.

Weaver, D., Beam, R., Brownlee, B., Voakes, P., & Wilhoit, G. (2007). The American journalist in the 21st century: U.S. news people at the dawn of a new millennium. Mahwah, NJ: Lawrence Erlbaum.

Weiss, C., & Singer, E. (1988). Reporting of social science in the national media. New York, NY: Russell Sage Foundation.

West, D., Whitehurst, G., & Dionne, E. J. (2011). Americans want more coverage of teacher performance and student achievement. Washington, DC: Brookings Institution.

White, D. (1950). The gatekeeper: A case study in the selection of news. Journalism Quarterly, 27, 383–396.

Yettick, H. (2009). The research that reaches the public: Who produces the educational research mentioned in the news media? Boulder, CO: National Education Policy Center.

AUThoR

HOLLY YETTICK, PhD, is the director of the Education Week Research Center at Editorial Projects in Education, 6935 Arlington Road, Bethesda, MD, 20814; [email protected]. Her research focuses on news media coverage and policy related to K–12 education.

Manuscript received July 11, 2013 Revisions received January 2, 2014,

and January 12, 2015. Accepted January 12, 2015