Public Policy and the Mass Media: An Information ... · PDF filePublic Policy and the Mass Media: An Information Processing Approach ... Influence Theory: The media tell the politicians

Public Policy and the Mass Media: An Information Processing Approach*

Bryan D. Jones Michelle Wolfe

Center for American Politics and Public Policy University of Washington, Seattle

Paper Prepared for the European Consortium for Political Research Workshop on Public Policy and the Mass Media: Influences and Interactions

Helsinki, Finland

May 2007 *This paper is based in part on work supported by the US National Science Foundation under Grant No. SES-0111443-001. We appreciate comments from Lance Bennett and Shanto Iyengar.

1

Public Policy and the Mass Media: An Information Processing

Approach

Bryan D. Jones Michelle Wolfe

University of Washington, Seattle

At the risk of being accused of gross oversimplification, we suggest that the role of the media in the public policy process can be characterized by four distinct theories. All of the theories fundamentally focus on the interactions between ‘the media’ collectively characterized, and ‘the politicians’, again collectively characterized. We can summarize them as follows: Influence Theory: The media tell the politicians what to think. Agenda Setting Theory: The media tell the politicians what to think about. Indexing Theory: The politicians tell the media what to write about.

In this paper, we propose a fourth characterization: Detection Theory: Politicians and the media struggle to identify, characterize, and prioritize complex multiple information streams. [OK, the last one sounds more complex, but it’s because we advocate moving away from an influence-based understanding of the media, which implies direct causal relationships, to a complex systems framework focusing on information flows, which implies nonlinear, interactive relationships.] In what follows, we first detail the first three theories, indicating the research that supports (and fails to support) each. Then we present the first research that systematically compares the media agenda-setting theory with the indexing hypothesis, showing the overall superiority of the idea of indexing.

Next we show how a misplaced emphasis on influence and direct causal effects has led to a mischaracterization of the problem. Rather than focusing on who influences whom, we suggest examining the role of information processing in the complex interaction between policymakers and the press. Finally we elaborate what we term detection theory, which is based on the theory of information processing in the public policy process detailed in Jones and Baumgartner’s The Politics of Attention.

Does the Mass Media Directly Influence Public Policy? Most of the work here depicts the media as private businesses producing profit, increasingly conglomerate and non-competitive. These corporate interests pursue government policies that benefit the bottom line. While this may indirectly affect the flow of information, it is of less interest here than other theories of media influence because it falls more in the domain of interest group

2

politics and can be understood through those models of policymaking. There seems to be little systematic literature on the question on the media using its communication power to influence more global aspects of public policy directly, although this would not rule out agenda-setting effects and other indirect influence paths. However, what anecdotal evidence is presented doesn’t seem to cut in one direction. On the one hand, surely consolidation has led to the concentration of power in the hands of fewer individuals, and the drive for profits has led to less news content and more fluff. On the other hand, new media outlets, including the Internet, and the propensity of news organizations even in oligopolies to compete, may offset this tendency. As far as the ‘pandering to mass tastes’ complaint, we suspect that has been part of the media mix as long as there has been media.

Does the Mass Media Set the Public Agenda? Does the mass media have the power to influence the policymaking

agenda? Media models of agenda-setting have focused mostly on the relationship between public opinion and media coverage rather than on the policymaking process itself (McCombs, Shaw, and Weaver 1997), but clearly there is an underlying assumption in these studies that public opinion influences public policy. Cohen (1963) elegantly asserted the importance of the media in agenda setting when he stated that while the media do not tell voters what to think, it does tell the public what issues to think about. As a consequence, the media may be able to establish boundaries or parameters of debate on policy issues. In accord with McCombs and Shaw’s 1972 findings, Weaver and Elliott (1985) find that newspapers do not mirror political reality; rather, they filter political activities to emphasize some issues and to minimize others (88).

Moreover, the media plays an instrumental role in raising the salience of issues in the public (Iyengar and Kinder 1987). Salience levels of issues in the public arena may determine whether or not issues expand or contract on the government agenda. Issue salience may determine voters’ turnout and choice preferences (Becker 1977). The media educate the public with factual information about public affairs. But, perhaps more significantly, the media educate the public about how much importance it should place on one issue versus another (McCombs and Shaw 1977, 5). The media might not change the publics’ attitudes, but the cognitive effects of mass communication steer attention, awareness and information (6).

Citizens are not alike in their receptivity to the media or in the degree to which the media messages affect their decision making processes. The greater a person’s political awareness, the more likely he or she is able to receive a message. However, the greater a person’s awareness, the more likely he or she is able to resist that message (Zaller 1992). Weaver (1977) also finds that the degree to which a person will orient himself or herself to media messages is in large part influenced by the amount of political information he or she obtains from interpersonal communications. The more information a person obtains from interpersonal communication, the less that person relies on news for gathering political information. Interestingly, McCombs (1977) finds that while newspapers

3

may set the public agenda in political campaign environments, the television media do not – the latter is set by the public agenda.

More recent scholarship puts Cohen and McCombs and Shaw’s “medium consequences” view into some doubt. For instance, Entman (1989) argues that not only do political messages transmitted by the media direct audiences’ attention and awareness, but that these messages also affect what people think (347; italics added) – not just what they think about. In the same way, Funkhouser and Shaw (1990) warn that because the media do not mirror political reality, audiences are exposed to synthetic experiences, which are used in the construction of social reality – a synthetic reality (57). The authors contend that this process changes the publics’ perception of political issues and thus impacts the public agenda (64).

Indexing Who or what sets the news media’s agenda? Is it a question of who

controls the media, or rather, what controls the media? In contradistinction to media agenda-setting models, Lance Bennett’s (1990) indexing hypothesis asserts that official debate sets the parameters of media debate (Althaus et al 1996, 408). As a consequence, official debate also sets the parameters of much public debate.

Bennett (1990) finds that mass media news professionals overwhelmingly use government elites as official sources and that they tend to index news to the range of official debate (106). The major driver of indexing is elite conflict. When disagreements between factions of political elites emerge, media issue coverage increases. The decline of issue coverage does not follow the resolution of a problem, or as an issue disappears from the formal agenda, but rather when official elites stop discussing it. Coverage declines in the absence of internal institutional opposition.

Government officials are preferred as sources because they have power (Gans 1979), they lend legitimacy to news stories (Sigal 1973) and because the economy of information encourages journalists to establish relationships with elite sources (Bennett 1990, 110). These relationships are symbiotic in that all parties benefit. Government officials also “go public” through the media in order to garner support for policies, to explain political actions, and to capitalize on “free” publicity (Kernell 1993 and Sigal 1973).

Althaus et al (1996 and 2003) test the indexing hypothesis and find, contrary to Bennett’s findings, that domestic elites are not the only or majority of official sources in news stories. They conclude that journalists sometimes abandon indexing for balanced coverage by including foreign sources as opposition voices. Similarly, Regina Lawrence (1996) tests the indexing hypothesis on news coverage of two “dramatic events” – incidences of police brutality in Los Angeles. She finds many non-official viewpoints in news coverage and attributes them to a topic – police brutality – that officials would rather keep out of public discourse.

4

Zaller and Chiu (1996) confirm the indexing hypothesis, but cite subtle variations in journalists’ propensity to index. Journalists’ propensity to index varies depending on ideological context, stage of conflict, and prior success of existing policies. Finally, Mermin (1999) visits the indexing hypothesis and finds that news coverage of foreign affairs follows the spectrum of official debate. Furthermore, Mermin finds that the media marginalize non-official voices when there appears to be official consensus on a policy topic.

Indexing studies find that government elites exert a great deal of control in the timing and focus of debate and policy alternatives the media represent in news stories and that elites moreover are able to marginalize opposition viewpoints. In most of the aforementioned studies, researchers test the indexing hypothesis on coverage of foreign policy crises. Additional tests should include an array of domestic policy issues to determine whether news coverage is largely a portrayal of power struggles within elite institutions (which is, after all, newsworthy) rather than critical discussion of societal conditions (Bennett 1996 and Iyengar 1992).

Comparing Media Agenda-Setting with Indexing Tests of the indexing hypothesis are unable to identify causal mechanisms

to explain who is leading whom. The problem with existing studies is that they never compare systematically the agenda-setting power of the media with the agenda-setting power of government officials. We have devised a scheme that allows for this explicit comparison.

The basic model is diagrammed below. We estimate the extent to which media coverage in one time period influences policymaking activities in a subsequent period; this we term the “agenda-setting path”. Simultaneously we estimate the extent to which policymaking activities in the first period influence media coverage in the subsequent period; we term this the “indexing path”.

Figure 1: A Path Model of Media Effects

Media Coverage t-1 Media Coverage t

Agenda-setting path Indexing path

Policymaking t-1 Policymaking t

As a first cut at the question of indexing versus agenda-setting, we calculated entropy, a measure of the supply of information, and the Herfindahl Index, a measure of issue concentration, across the nineteen policy content

5

categories of the Policy Agendas Project (http://www.policyagendas.org/; see Jones and Baumgartner 2005 for a full description of the coding system). These are alternate measures of the same concept. Entropy assesses dispersion across sources for issue-messages under the assumption that dispersion of messages across sources indicates more information. The Herfindahl measures concentration of attention on topics; the more concentration, the less will be the supply of information. Conflict among elites follows the contours of available policymaking venues (Baumgartner and Jones 1993). Policymaking venues are arenas for resolving conflict, and congressional committees play a key role in this regard in the US congress. Issue specialists tend to control issues within key congressional committees, and these committees tend to reach equilibrium on how an issue is defined. Policy solutions follow these issue definitions; so long as issue definitions remain stable, policymaking remains stable. When, however, new aspects of issues are stressed, new committees generally get involved as legislators try to change control of the issue. This involves conflicts over which committees have jurisdiction over the issue (Baumgartner, Jones, and McLeod 2000). Elite conflict emerges within the committee structure of congress. The indexing hypothesis can be extended to inter-institutional factional conflict in a straightforward manner. Where jurisdictional struggle occurs over issue definitions, increased media coverage should result. In the case of congress, we calculate the entropy across the committee structure, which basically tells us whether the information in hearings on a particular topic is ‘bottled up’ in a single committee, or whether it broadly shared in a competition among committees. The latter is indicative of ‘hot’ issues, and the supply of information can be expected to increase as competition among sources of information (that is, the committees) increase. (See Jones, Baumgartner, and de la Mare 2005 for a full discussion). We are interested here in how the spread of attention in congress and the New York Times correspond. More importantly, we’re comparing the extent to which the ‘hot’ issues in congress are covered by the media (the indexing link) with the extent to which the ‘hot’ issues are influenced by prior media coverage. For the New York Times, we calculate the measures directly across the topics, since we’re assessing only one source here.

Before we proceed to the study of agenda-setting and indexing, we examine the trace of entropy for the New York Times and congress across the time period studied (1946-2003) in Figure 2. Peak entropy for the all stories occurred in 1973; not inconsequently, the highest proportion of stories devoted to public policy matters occurred the following year. Since then editors at the Times have imposed a steady and marked decline of policy-relevant material on the paper’s coverage, all the while producing an increasingly less diverse news product. In the 1970s, 40 to 50 percent of coverage was devoted to politics and policy; by the 1990s, this had dropped to around half of the previous level. (It is

6

worth noting that the overall size of the Times, the number of articles, also declined over this time; so there was much more news coverage in the 1970s than there is today.) The New York Times has steadily but surely moved from a strong and diverse outlet for discussion of policy and politics to a less diverse format dominated by style, arts, leisure, and sports. The move to ‘soft news’ has been well-documented in media studies (Patterson 2000); our contribution is to note the clear and unmistakable decline in diversity and in policy content that has become so marked since the 1970s.

Figure 2: The Decline of Policy Relevant Information in the New York Times

R2 = 0.77

R2 = 0.57

1

1.05

1.1

1.15

1.2

1.25

1.3

1946

1948

1950

1952

1954

1956

1958

1960

1962

1964

1966

1968

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

Year

Entr

opy

0

0.1

0.2

0.3

0.4

0.5

0.6

Prop

ortio

n

NYT En tropy

Proportion of NYT policy artticles

Source: Calculated from data made available by the Policy Agendas Project (http://www.policyagendas.org/ ).

Is the decline in press coverage simply a media phenomenon or does it affect government? Figure 3 presents our entropy measure calculated for each house of Congress for major topic categories and committees. As can be seen in the figure, the time path of information supply provided to Congress through its committee system traces a roughly similar pattern as New York Times coverage. We also present the total number of hearings held, showing this number as a proportion of its peak value, 2246 in 1979. Clearly, the capacity to hold hearings has varied substantially over time. Just as clearly, this is fully within the control of Congress and its members, and they have both expanded and restricted the energy and resources they devote to the hearings process. As hearings expanded during the 1960s and 1970s, so entropy and information grew; as hearings have declined since 1980 or so, so has the supply of information.

7

Figure 3: Congressional Entropy and the Number of Hearings

R2 = 0.63

R2 = 0.90

0.85

0.9

0.95

1

1.05

1.1

1.15

1.2

1.25

1946

1948

1950

1952

1954

1956

1958

1960

1962

1964

1966

1968

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

Year

Entr

opy

0.15

0.25

0.35

0.45

0.55

0.65

0.75

0.85

0.95

1.05

1.15

Prop

ortio

n

Entropy

Hearings (proportion of highest value)

Source: Calculated from data made available by the Policy Agendas Project (http://www.policyagendas.org/ ).

Now we turn to the study of the relationship between these two series—whether indexing or media agenda-setting dominate the complex relationship between press and policymakers. Figure 4 presents the results diagrammatically (presenting the appropriate standardized regression coefficients for the path coefficient estimates), and Table 1 presents these results in more standard tabular form, for Entropy. We combined entropy calculations for House and Senate; running these measures separately on each legislative chamber did not alter the results. Results were similar for the Herfindahl Index, as we expected.

Figure 4: Estimates for the Overall Model

-0.14New York Times t-1 New York Times t

0.5 0.73*

0.05

Congress t-1 Congress t 0.91*

Source: Calculated from an analysis similar to Table 1, using standard regression coefficients for the paths. Asterisks indicate statistical significance at 0.05.

8

Table 1: Information Supply in Congress and the Press: OLS Regression Analysis for Figure 4

Models: Dependent

Variable: NYT Policy

Entropy

Dependent Variable:

Congressional Entropy

Constant .954***(.143)

-.002(.150)

NYT Policy Entropy, Lag 1 -.145(.143)

.124(.155)

House-Senate Entropy, Lag 1

.324***(.061)

.897*** (.066)

R2 .451 .870 Adj. R2 .423 .863 DW 1.92 2.54 N 41 41

The findings strongly support the indexing hypothesis. The path from media to legislature is not significant, while the path from legislature to media is. Moreover, the inertial path, from media coverage at one time to coverage at a second time, is insignificant, but the path from legislature to media is. The ‘hot’ issues in congress—those characteristic of high entropy (and hence low Herfindahl scores) capture the attention of the media, and appear as stories subsequent to the congressional controversy. These ‘hot’ issues are pretty much similar from one time period to another, as indicated by the inertial path, but enough ‘slippage’ occurs that new issues are able to break through on occasion. These findings tell us only that concentration patterns across major issues follow an indexing pattern. 1 They say nothing about issue-specific effects. Wolfe (2006) analyzed each of the nineteen major topic areas of the Policy Agendas Project in a similar manner, using the Herfindahl Index to assess issue concentration. The measure is similar in many respects to Entropy, used above. Economists use the Herfindahl Index as a measure of market concentration, and US Congressional scholars have used it to examine changes in jurisdictional monopolies in congressional committees (Baumgartner, Jones and McLeod 2000, Hardin 1998, 2002). The index measures unit concentration or diffusion

1 One problem with this estimate is the long lag time (one year) involved. We really don’t have any other option given the low number of hearings and stories across topics. In any case, the findings are strong given this long lag. The proper interpretation is that when there is a preoccupation with an issue, it tends to sustain itself across time, in the media and in congress. This does not mean there are not shorter-term effects; it is just that we cannot pinpoint them with our design here. Given this type of issue dynamics, the media indexes issue controversy in congress, and does not seem to set the agenda for congress. There may also be reason to expect longer, cumulative type lags. See below.

9

within categories by summing the squared proportion of items in each category. Scores approaching zero indicate near even unit diffusion among several units while large scores indicate near domination of a category by few units.

Wolfe estimated the model depicted in Figure 1 above on each of the nineteen topic areas. Results of the separate regression analyses indicated support for the indexing hypothesis in 10 of the 19 policy topics. Interestingly, of the nine policy topics that do not have statistically significant lagged hearing coefficients, not one is explained better by the lagged media variable. There was no support in this comparative analysis on any issue of the media agenda-setting model.

There is most likely not one simple answer behind the question of who leads whom. It may be the case that from issue to issue there is more or less interplay among the media and governing elites – and the public. Different issues may have different agenda setting dynamics and will thus have different agenda setting results. Soroka (2002) has shown in Canada that policy makers lead on issues, such as debt and deficit, which public citizens do not experience directly, but the media do have an influence on more salient issues. Nevertheless, in a direct comparison between media agenda-setting and indexing, the indexing hypothesis performed far better.

Detection Theory In detection theory, actors—politicians, policy entrepreneurs, interest groups, and the media—are enmeshed in a complex set of interactions focusing on the recognition and interpretation of multiple complex and interacting information signals. While the actors involved all have preferences and goals, the result of this system is the processing of information. A key aspect of an information processing approach to the public policy process is the role of attention allocation (Jones and Baumgartner 2005). Attention is a scarce good, can be allocated only in ‘chunks’ rather than continuously, and is a necessary condition for policy change. As a scarce asset, actors in politics struggle to control it. The genesis of our approach may be found in the work of Baumgartner and Jones (1993) on policy punctuations. In their examination of the role of the media in the policy process, they wrote that “Media attention sometimes precedes and sometimes follows changes in attention by government agencies….each can affect the other, reinforcing the pattern of positive feedback and punctuated equilibrium (125)” [See also Baumgartner, Jones, and Leech 1996]. Situations alone, dire as we may view them, do not automatically generate public policy responses. Information about situations must come from somewhere, and that information must be interpreted in terms that are relevant to government action. Even more importantly, there are many situations, and many that could be improved by appropriate governmental action, but all cannot be addressed at once. They must be prioritized. Indeed, an information processing

10

approach to the study of policy processes focuses on how a political system collects, assembles, interprets, and prioritizes signals from the environment Jones and Baumgartner 2005: 7).

The media is intimately involved in this process, but it is a mistake to think of its role in direct, causal terms. Causation implies influence; influence implies intentionality. Clearly political leaders intend to influence the media, but it is less clear that the media intends to influence politics, once we get beyond the issue of corporate self-interest. It is even less clear that any one set of actors—be they from the media or the political elites, or from business or from ‘the public’—in any sense ‘controls’ policymaking.

Element 1: Information Matters A major problem with existing models of mass media effects on public policy is that they are rooted in an influence framework. The question of ‘who has power’ is certainly not irrelevant, but it has led to overly simplistic models of the interaction between government officials and the media. We now turn to developing a model emphasizing information-processing rather than influence. Why does the indexing hypothesis fare so well in our tests described above? It’s because what policymakers do is news. The media covers the news, but it may be surprising how faithfully media concentration on issues follows congressional concentration on issues. There is a missing variable in our system, however, and it potentially affects the indexing path more than the agenda-setting path. Information about policy-relevant matters is not confined to media-policymaker interactions. Information is at least partially exogenous to this system, and it is therefore possible that both the media and congress are responding to this partially exogenous information stream. Indeed, Behr and Iyengar (1985) find that real world conditions and events affect both the selection of news stories for television broadcast and the public’s perception of the importance of issues. Behr and Iyengar do not, however, examine the relative effects of real world events and media on policymaking. Jones and Baumgartner (2005: Chapter 8) examine policy responses to objective conditions in three policy areas: economics, crime, and social welfare. In the first two, the facts clearly played a role (we have more to say about crime control policy below). On the other hand, social welfare policy was most responsive to political factors; objective conditions did not play a role.

Scholars have produced relatively few studies incorporating objective conditions, public policy activity, and media coverage in a single framework, but one stands out. In a study of US response to climate change using a vector autoregression framework, Liu, Lindquist and Vedlitz (2006) report that both media and congressional attention responded to climate changes (measured by CO2 concentrations in the atmosphere). Controlling for objective conditions (CO2 concentrations), there were no direct effects from the media to congress nor from congress to the media. Scientific publications on the matter influenced

11

congressional attention but not the media. The diagram below summarizes the basic findings.

In at least this policy arena, the evidence supports an information-processing model in which the objective conditions of climate change affects media coverage of the situation, and independently influences congressional attention. Accumulating scientific evidence also independently influences policymaking attention.

Figure 5: Objective Conditions and Media and Congressional Attention

Media

CO2 Concentration

Congress

Scientific Studies Source: Drawn from results presented in Liu, Lindquist, and Vedlitz (2006)

What if a major event happens, such as an urban riot? Is it not to be expected that both the media and policymakers respond, at least by allocating attention? Is it not also to be expected that the media would subsequently devote attention to policymakers? Even absent a dramatic event, circumstances can change such that the media devotes attention to them, as do policymakers. This certainly seems to be true of climate change.

Element 2: The Media Can Help to Set a Tone for Subsequent Policy Action For whatever reasons, the media can become preoccupied with an arena

that may be ripe for policy action. By repeated muckraking and highlighting particular aspects of the information stream, the media may help to set the tone for subsequent policy action. Changes in tone surrounding an issue can presage future policy change, as Baumgartner and Jones (1993) showed in the cases of nuclear power, pesticide regulation, and smoking policy.

This ‘tonal’ component may be picking up changes in the broader political ‘mood’ (Kingdon 1985), or it may have a causal effect on that mood. It may not penetrate the attention of the typical member of the mass public, but it may help to mobilize parts of the attentive public.

It is a mistake to assume that the mass media has a direct policymaking effect via this tonal component; that just rarely happens. But it can facilitate policy entrepreneurship on an issue in certain circumstances.

In some cases, media coverage can be associated with heightened partisan divisions. Baumgartner and Jones (1993) argued that US urban policy over time transformed from a new issue demanding fresh solutions in the 1940s

12

and 1950s to a standard partisan one by the 1970s. Sapotichne (2006) reproduced this basic of the work on US urban policy using updated Policy Agendas Project data, but his models were simpler. Media and congress devoted attention to cities simultaneously. A burst of attention occurred in the 1960s for both congressional hearings and New York Times coverage. But Sapotichne added an important new finding concerning the connection between media coverage and partisanship. He showed that during periods of low attention (1946-1962), the connection between media and congress were similar for Democratic presidents and Republican presidents, but after the burst media and congressional activity deviated. Media attention was much more closely associated with congressional policy attention for Democratic presidents than for Republican ones. Peter John (2006), in a study of budgets for urban affairs in the United Kingdom between 1966 and 2003, finds that a combination of prior media coverage in the London Times on inner city matters and urban riots combined to affect the central government’s allocations for urban matters. This finding is in keeping with those of Baumgartner and Jones (1993). Prior media attention can sensitize the political system to new events in a manner that can result in major policy changes. John finds that media coverage of urban riots is not implicated causally; only more generalized inner city coverage is implicated. It is likely that policymakers and the media are reacting simultaneously to the events, but that prior media attention to the more general topic of inner city matters has set the tone for subsequent policy action. In the path approach to the study of media agenda-setting versus indexing, we used a one-year lag, which seems long (but is unavoidable because of data limits). On the other hand, if media coverage contributes to a tonal environment for policymaking, this actually can be too short. Figure 6 is a graph of US congressional hearings and New York Times stories on economic issues, again from the Policy Agendas Project. Wolfe’s study showed no relationship between media concentration and hearings concentration on economics using the standard one-year lag. There are three periods of increases in policymaking interest in economics, peaking in 1982, 1995, and 2003. In the first period, it seems as if the media leads congressional interest, but in a tonal fashion, with large lags. In the second period, hearings again lead, but with a one-year lag, but in the final period, the media leads, again with about a one-year lag. It is possible that in each of these three periods a different causal relationship characterized the relationship between the media and policymakers: a media tonal effect, an indexing effect, and a media agenda-setting effect. This, however, is not the whole story here. Introducing objective conditions (the unemployment rate), and public opinion (the proportion of the public citing economics as the most important problem facing the nation in Gallup polls), Jones and Baumgartner (2005: Chapter 8) developed a model in which economic conditions led to public opinion which in turn led to policymaking interest. Media was not significant. This does not rule out the observation that causal complexities with temporal sensitivities can characterize the relationship between

13

policymaking and the press, but it does indicate that we ignore the objective conditions that can underlie the process at our peril.

Figure 6: Congressional Hearings and New York Times Coverage, Macroeconomic Issues

0.00%

1.00%

2.00%

3.00%

4.00%

5.00%

6.00%

7.00%

8.00%

1946

1948

1950

1952

1954

1956

1958

1960

1962

1964

1966

1968

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

Con

gres

sion

al H

earin

gs

0.00%

0.50%

1.00%

1.50%

2.00%

2.50%

3.00%

3.50%

4.00%

4.50%

5.00%

New

Yor

k Ti

mes

Sto

ries

Hearings (3-year moving average)

NYT (3-year moving average).

Element 3: Media Attention and Policymaking Activities Can Become Intertwined

in Positive Feedback Systems In some cases, policymakers are quite happy to work behind the curtains

within policy subsystems, with experts and interested parties collaborating to set the course of public policy. But in others politicians may crave attention, seeking to advertise their political careers or their favored policy solutions—especially when they find themselves under electoral threat or when they want to change policy directions.

Politicians with policy solutions need to raise the salience of the problem their solution is supposed to solve. Wolfe’s study of indexing suggests that in many situations, they are successful in this. Mostly this heightened attention levels of or declines after a period of interest, because other problems have bombarded the agenda, simple issue fatigue, or because solutions offered are not convincing. In some cases, however, a ‘Noah effect” (Mandelbrot 1997) kicks in. Success begets success and the political system devotes disproportionate attention to the issue (Jones and Baumgartner 2005). An “arms race” of policy solutions can occur.

In these cases, key variables—objective conditions (the exogenous information stream) media coverage, public opinion, and policymaking attention--come into phase together, and this can result in major punctuations in policy. An

14

exceptionally clear example of this occurred in the late 1960s and early 1970s in the United States in crime policy, as illustrated in Figures 7 and 8. Figure 7 shows the rapid rise in the violent crime rate (as assessed by the Federal Bureau of Investigation), and the associated rise in media coverage (as assessed by the Policy Agendas Project’s sample of New York Times stories). Note that the rise in Times coverage preceded by a year or so the rise in the crime rate (no causal relationship claimed!)2 Note that when the crime rate declined, the Times coverage did not. By 2000 the violent crime rate was about the same as it had been in 1968, just prior to the rapid rise, but the New York Times’ coverage of crime was actually higher in 2000 than it had been at the peak of the violent crime wave in the early 1980s.

Figure 8 shows how the various potentially causal variables (‘streams’ in Kingdon’s terminology) came into phase in the late 1960s. Between 1968 and 1970, the violent crime rate, media coverage, the percentage of the public who thought that crime was the most important problem facing the nation, and congressional hearings all rose rapidly. Figure 6 shows the rise of policy responses to the crime ratchet. The number of laws passed to address the problem increased; by the late 1970s around 5% of all lawmaking activity in a typical year involved crime and justice (the earlier rise in the mid-1960s involved mostly juvenile justice issues). Budgetary commitments followed lawmaking; in 1970 alone the federal budget for crime and justice tripled.

One interesting facet of the late 1960s crime ratchet is that when the crime rate declined, nothing else associated with crime policy (except public concern) followed suit. Media attention, congressional hearings, lawmaking, and budgetary commitments either continued along at the higher level or increased. In the case of budgets, the increase leveled off in the 1970s and early eighties, but then increased spectacularly—in the absence of any rise in the violent crime rate. At this time a kind of policy hysteria broke out in Washington over the allegations of a ‘crack cocaine’ epidemic. A ‘bidding war’ for which party was tougher on criminals yielded major legislation setting new incarceration standards. Yet neither violent nor property crimes rose during the period (but the homicide rate did increase somewhat, but mostly after the laws were passed). The result was draconian sentencing guidelines for drug sales and use that filled federal prisons with drug sellers and generated a virtual crime policy-prison industrial complex. States obliged by engaging in their own policy hysteria. The result was a classic self-sustaining policy subsystem with enormous spending commitments from national, state, and local governments.

This illustrates an important point. Information streams, we claimed earlier, were ‘partially exogenous’. Certainly the big policy buildup in crime control in the late 1960s was associated with an increase in crime, but the secondary buildup in the 1980s was not. A final increase in the Clinton years was associated with a distinct change in crime policy—it was directed more at

2 This could be due to an earlier rise in crime in the New York City region, or sensitivity on the part of editors at the paper to early warning signs of the rise, or for other reasons.

15

getting police on the street than locking up criminals. In the case of the crime policy hysteria of the 1980s, there little evidence that increases in crime drove the policy punctuation; at a minimum the policy response was vastly disproportional to the indicators of the problem. Media coverage fed into the frenzy (note the increased coverage in the period); it was a participant but not a causal influence (Wolfe’s study shows no relationship between hearings and New York Times coverage, or between coverage and hearings, suggesting a simultaneous increase.)

Concluding Comments In this paper we have offered some thoughts about the role of the media in the policymaking process. First we have examined in some detail the primary prevailing models of press-state relations, particularly focusing on a direct comparative test of US congressional policymaking attention and New York Times coverage of major issues. In the simple models we test (linear, with one-year lags), the indexing hypothesis is clearly superior. Indexing, however, is not the end of the story. Our notions of the role of the media in the policymaking process changes when we incorporate elements of an information-processing approach to press-state dynamics. It is likely that for many issues, both the press and policymakers are responding to changes in indexes of objective conditions. Responses of either or both may be disproportionate—a long period of under-reaction with bursts of over-reaction (Jones and Baumgartner 2005). This means that the relationship between press and policymakers is characterized by non-linearities and positive feedback effects, such as tonal, context-setting effects and policy outbursts. Most information ‘signals’ are not processed in straightforward, proportionate processes. Neither is the politics-media interaction generated exogenously, with no reference to the facts. While this implies that our notions of causation, and hence influence, must become more sensitive to context, it by no means implies we need to give up on the development of theory. It does mean that these theories will be increasingly related to the ‘complex systems’ perspective common in the natural sciences today.

16

Figure 7: The Crime Policy Ratchet

2%

3%

3%

4%

4%

5%

5%

6%

6%

7%

7%

1946

1948

1950

1952

1954

1956

1958

1960

1962

1964

1966

1968

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

Year

Pct o

f Tot

al A

gend

a--N

YT

4

5

6

7

8

9

10

11

Hom

icid

e R

ate

Homicide Rate

NYT (3-yr Moving Average)

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%

10%

1946

1948

1950

1952

1954

1956

1958

1960

1962

1964

1966

1968

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

Year

Perc

enta

ge o

f Age

nda-

-Hea

rings

and

NYT

0%

5%

10%

15%

20%

25%

30%

Perc

enta

ge o

f Age

nda-

-MIP

HearingsNYTMIP3 per. Mov. Avg. (NYT)3 per. Mov. Avg. (Hearings)3 per. Mov. Avg. (MIP)

NYT leads; Congress and the public follow

17

Figure 8: Crime and Justice, a) Laws Passed; b) Real Budget Authority, 1954-2000.

0.0%

1.0%

2.0%

3.0%

4.0%

5.0%

6.0%

7.0%

8.0%

9.0%

1946

1948

1950

1952

1954

1956

1958

1960

1962

1964

1966

1968

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

Perc

enta

ge o

f Pub

lic L

aws

on C

rime

and

Just

ice

(3-y

ear m

ovin

g av

erag

e)

0

5,000

10,000

15,000

20,000

25,000

30,000

1954 1956 1958 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 19821984 1986 1988 1990 1992 1994 1996 1998 2000

Rea

l Bud

get A

utho

rity,

Crim

e an

d Ju

stic

e, $

Mil

-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

Prop

ortio

n C

hang

e

US crime budget trippled in 1970

proportion change

real budget

18

References

Althaus, Scott L. 2003. When News Norms Collide, Follow the Lead: New Evidence for

Press Independence. Political Communication 20: 381-414. Althaus, Scott L., Jill Edy, Robert Entman, and Patricia Phalen. 1996. Revising the

Indexing Hypothesis: Officials, Media, and the Libya Crisis. Political Communication 13: 407-421.

Baumgartner, Frank R. and Bryan D. Jones. 1993. Agendas and Instability in American

Politics. Chicago: The University of Chicago Press. Baumgartner, Frank R., Bryan D. Jones and Michael McLeod. 2000. The Evolution of

Legislative Jurisdictions. Journal of Politics 62: 321-49. Baumgartner, Frank R., Bryan D. Jones and Michael McLeod . 1998. “Lessons from the

Trenches: Ensuring Quality, Reliability, and Usability in the Creation of a New Data Source” The Political Methodologist 8(2) Spring 1998: 1-10.

Baumgartner, Frank R. and Beth L. Leech. 1998. Basic Interests: The Importance of

Interest Groups in Politics and in Political Science. Princeton, N.J.: Princeton University Press.

Baumgartner, Frank R., Bryan D. Jones and Beth L. Leech. 1997. Media Attention and

Congressional Agendas. In Shanto Iyengar and Richard Reeves, ed., Do the Media Govern? Thousand Oaks: SAGE Publications.

Becker, Lee B. 1977. The Impact of Issue Salience. In Donald L. Shaw and Maxwell E.

McCombs, eds., The Emergence of American Political Issues: The Agenda setting Function of the Press. St. Paul: West Publishing Co.

Behr, Roy and Shanto Iyengar. 1985. Television News, Real-World Cues, and

Changes in the Public Agenda. Public Opinion Quarterly 49: 38-57. Bennett, W. Lance. 1990. Toward a Theory of Press-State Relations in the United

States. Journal of Communication 40: 103-125. Bennett, W. Lance. 2003. A Semi-Independent Press: Government Control and

Journalistic Autonomy in the Political Construction of News. Political Communication 20: 359-362.

Cobb, Roger W. and Charles D. Elder. 1972. Participation in American Politics: The

Dynamics of Agenda-Building. Baltimore: The Johns Hopkins University Press. Dearing, James M. and Everett M. Rogers. 1996. Agenda setting. Thousand Oaks:

SAGE Publications. Downs, Anthony. 1972. Up and down with ecology - the "issue-attention cycle." The

Public Interest 28: 38-50.

19

Entman, Robert M. 1989. How the Media Affect What People Think: An Information Processing Approach. The Journal of Politics 51: 347-370.

Funkhouser, G. Ray and Eugene F. Shaw. 1990. How Synthetic Experience Shapes

Reality. In Doris A. Graber, ed., Media Power in Politics. 4th ed. Washington, D.C.: CQ Press.

Gans, Herbert J. 1979. Deciding What's News. New York: Pantheon Books. Hardin, John. 1998. Advocacy versus Certainty: The Dynamics of Committee

Jurisdiction Concentration. Journal of Politics 60: 374-97. Hardin, John W. 2002. “Multiple Topics, Multiple Targets, Multiple Goals, and Multiple

Decision-Makers: Congressional Consideration of Comprehensive Health Care Reform.” In Policy Dynamics, eds. Frank R. Baumgartner and Bryan D. Jones. Chicago: University of Chicago Press.

Iyengar, Shanto. 1991. Is Anyone Responsible: How Television Frames Political Issues.

Chicago: The University of Chicago Press. Jacobs, Lawrence R. and Robert Y. Shapiro. 1996. Toward the Integrated Study of

Political Communications, Public Opinion, and the Policy-Making Process. PS: Political Science and Politics 29: 10-13

. John, Peter. 2006. Explaining Policy Change: the impact of the media, public opinion

and political violence on urban budgets in England. Journal of European Public Policy 13: 1053-68.

Jones, Bryan D., and Frank R. Baumgartner. 2005. The Politics of Attention: How Government Prioritizes Problems. Chicago: University of Chicago Press.

Jones, Bryan D., Frank R. Baumgartner, and Erin de la Mare. 2005. The Supply of Information and the Size of Government in the United States. Seattle, WA: Center for American Politics and Public Policy, University of Washington.

Kedrowski, Karen M. 1996. How Members of Congress Use the Media to Influence

Public Policy. In Doris A. Graber, ed., Media Power in Politics. 4th ed. Washington, D.C.: CQ Press.

Kernell, Sam. 1997. The Theory and Practice of Going Public. In Shanto Iyengar and

Richard Reeves, ed., Do the Media Govern? Thousand Oaks: SAGE Publications.

Kingdon, John W. 1984. Agendas, Alternatives, and Public Policies. Boston: Little,

Brown. Lawrence, Regina G. 1996. Accidents, Icons, and Indexing: The Dynamics of News

Coverage of Police Use of Force. Political Communication 13: 437-454. Liu, Xinsheng, Eric Lindquist, and Arnold Vedlitz. 2006. Explaining US Media and

Congressional Attention to Climate Change, 1969-2005. Paper presented at the Southern Political Science Association Meetings, Atlanta Georgia, January 5-7.

20

Lippmann, Walter. 1922. Public Opinion. New York: Simon & Schuster Inc. Livingston, Steven and W. Lance Bennett. 2003. Gatekeeping, Indexing, and Live-Event

News: Is Technology Altering the Construction of News? Political Communication 20: 363-380.

McCombs, Maxwell E. Newspapers versus Television: Mass Communication Effects

Across Time. In Donald L. Shaw and Maxwell E. McCombs, eds., The Emergence of American Political Issues: The Agenda- setting Function of the Press. St. Paul: West Publishing Co

McCombs, Maxwell E. and Donald L. Shaw. 1972. The Agenda-setting Function of the

Press. The Public Opinion Quarterly 36: 176-187. McCombs, Maxwell E. and Donald L. Shaw. 1977. Agenda setting and the Political

Process. In Donald L. Shaw and Maxwell E. McCombs, eds., The Emergence of American Political Issues: The Agenda-Setting Function of the Press. St. Paul: West Publishing Co.

McCombs, Maxwell and Jian-Hua Zhu. 1995. Capacity, Diversity, and Volatility of the

Public Agenda: Trends from 1954 to 1994. The Public Opinion Quarterly 59: 495-525.

McCombs, Maxwell E., Donald L. Shaw, and David Weaver, eds. 1997. Communication

and Democracy. Mahwah, NJ: Lawrence Erlbaum Associates. McQuail, Denis. 1979. The Influence and Effects of Mass Media. In Doris A. Graber, ed.,

Media Power in Politics. 4th ed. Washington, D.C.: CQ Press.

Mandelbrot, Benoit. 1997. Fractals and Scaling in Finance. New York: Springer.

Mermin, Jonathan. 1999. Debating War and Peace: Media Coverage of U.S.

Intervention. Princeton, N.J.: Princeton University Press. Nelson, Barbara J. 1984. The Agenda setting Function of the Media: Child Abuse. In

Doris A. Graber, ed., Media Power in Politics. 4th ed. Washington, D.C.: CQ Press.

Patterson, Thomas. 2000. Doing well and Doing Good: How soft news and critical

journalism are shrinking the news audience and weakening democracy. Joan Shorenstein Center on the Press, Politics and Public Policy.

Rogers, Everett M. and James W. Dearing. 1988. Agenda setting Research: Where Has

It Been, Where Is It Going? In Doris A. Graber, ed., Media Power in Politics. 4th ed. Washington, D.C.: CQ Press

Sapotichne, Joshua. 2006. Media Attention and Congressional Hearings Activity.

Seattle, WA: Center for American Politics and Public Policy, University of Washington

21

Shaw, Donald L. 1977. The Press Agenda in a Community Setting. In Donald L. Shaw

and Maxwell E. McCombs, eds., The Emergence of American Political Issues: The Agenda Setting Function of the Press. St. Paul: West Publishing Co.

Sigal, Leon V. 1973. Reporters and Officials. Lexington, Mass.: D.C. Heath and

Company. Soroka, Stuart N. 2002. Issue Attributes and Agenda setting by Media, the Public, and

Policymakers in Canada. International Journal of Public Opinion Research 14: 264-285.

Talbert, Jeffery C. and Matthew Potoski. 2002. “The Changing Public Agenda Over the

Postwar Period.” In Policy Dynamics, Frank R. Baumgartner and Bryan D. Jones (eds). Chicago, IL: The University of Chicago Press.

Weaver, David and Swanzy Nimley Elliott. 1985. Who Sets the Agenda for the Media? A

Study of Local Agenda building. Journalism Quarterly 62: 87-94. Weaver, David H. 1977. Political Issues and Voter Need for Orientation. In Donald L.

Shaw and Maxwell E. McCombs, eds., The Emergence of American Political Issues: The Agenda setting Function of the Press. St. Paul: West Publishing Co.

Wolfe, Michelle. 2006. Comparing Agenda Setting with Indexing. Seattle, WA: Center

for American Politics and Public Policy, University of Washington. Zaller, John and Dennis Chiu. 1996. Government's Little Helper: U.S. Press Coverage of

Foreign Policy Crises, 1945-1991. Political Communication 13: 385-405 Zaller, John R. 1992. The Nature and Origins of Mass Opinion. New York: Cambridge

University Press.