On the Relationship between Media/Political Leaders and the Public

What follows is a paper that I plan to incorporate in a forthcoming book in some form, probably with a different title and possibly with a slightly different set of connectors to the overall book theme. I post it here primarily to expand upon an argument that frequently comes up in political and media discussions: how should we understand the correct relationship between public sentiments and the presentation of issues by politicians and by media?

First, some quick background: functionalist theory (sometimes referred to as structural functionalism) is the dominant tradition in sociology (at least in the US). Functionalism is consistent with classical and pluralist democratic theory which holds that the public's desires and views are expressed by the political leadership and by mass media. The nature of media's news coverage and that of their entertainment offerings and the kind of laws and public policies made by public officials are ascribed to public interest and public sentiment. Thus, for example, many people would explain the growth of reality-TV shows as due to public interest. Another example would be to account for the increasingly punitive criminal justice system as due to an increasingly conservative and punitive-minded public. The failure so far of impeachment proceedings being already underway (or finished!) is likewise seen by many people as a reflection of the apathy of the average American.

As I argue below, however, I believe that the functionalist model and the democratic/pluralist models are flawed for both theoretical reasons and on empirical grounds. This has momentous consequences in the real world of political action. It is, of course, not surprising that theoretical questions have real world implications. One cannot move forward to change the world if one doesn't understand correctly how things work. And theory is indispensible in this process because a wrong-headed theory will lead you astray inevitably.

The focus of the piece that follows is on social constructionist theory which is the dominant force within social problems theory in sociology. Social problems theory seeks to account for why and how certain issues become "social problems:" issues that garner a tremendous amount of public attention, usually measured by substantial shifts in polls, by public officials holding hearings, making pronouncements, passing resolutions or laws, and/or by media doing a lot of stories about that issue. Social constructionism is a particular thread within social problems theory.

The version I'm posting here lacks a bibliography - to be added later.

Theory of Problems, Problems of Theory:
A Critical Appraisal of the Dominant Models in Social Problems Theory

by

Dennis D. Loo

Social constructionism has yet to satisfactorily resolve one central problem: how to account for why one issue wins out over other issues. Social constructionists have typically solved this in a post hoc fashion: the winning issue triumphed because its claim must have been more compelling (Lowney and Best 1995). How do we know that it was more compelling? The fact that it won proves that it was more compelling. This is, of course, a circular argument. It does not tell us what distinguished the winning issue from the also-rans. In an attempt to overcome this post hoc problem, Lowney and Best (1995) cite three variables as decisive: an issue’s resonance with cultural themes, the mobilization of greater resources by the winning claims-making group, and contingent events (such as the drug-induced death of Len Bias which contributed to the mid-1980s Drug War).

Unfortunately, upon closer examination, Lowney and Best’s answer to the post-hoc problem still does not resolve the problem. Briefly put, their resolution does not fully answer the following objections: 1) There are numerous possible issues that could resonate with cultural themes. Why did this issue and not another get selected? 2) Did the successful interest group actually mobilize more resources than other groups? Frequently, the state and/or media decide on their own to pay more attention to an issue. Interest group mobilization does not necessarily enter into the picture at all. 3) While contingent events can play a role in the genesis of a social problem, whether those events are publicized heavily or not is a decision that rests ultimately with media and/or the state. If mainstream media elect not to call a great deal of attention to an event, then that event will almost invariably pass relatively unnoticed. The post hoc problem in social problems accounts remains unresolved because a key element is being overlooked both by Lowney and Best, and more generally, by the predominant models in social problems theory, structural strain and interest group: the key role media play in selecting, initiating and framing social issues.

The structural strain and interest group models do commonly devote substantial attention to media activity, and in particular, media’s tendency to exaggerate the incidence and prevalence of a social problem. These models, however, treat media as epiphenomenal to the process by which a social problem emerges. That is, the media are held to commonly react to other claims-makers rather than acting as claims-makers themselves. Best (2001:15), for example, describes media’s role as that of conduit for interest groups’ claims: “[s]uccessful activists attract support from others. The mass media – including both the popular and entertainment media … relay activists’ claims to the general public.”(italics added.). Similarly, Surette (1998) describes the relationship between media and claims-makers this way:

"The media serve the role of world knowledge conduit and playing field for the competition between claims-makers. Claims-makers compete for media attention and media favor claims that are dramatic, sponsored by powerful groups, and related to established cultural themes. (p. 10-11. Italics added.)"

Let’s examine each of these three factors in turn. Dramatic value is, of course, important. It is, however, only one among many factors that media weigh in deciding whether to feature a story. Indeed, drama does not by itself, in fact, constitute the most important factor (media studies has addressed this extensively, see, for example: Croteau and Hoynes 2001, Fishman and Cavendar 1998, Cose 1993, Kurtz 1994, Gans 1979, Loo 1996). Regarding powerful group sponsorship: media themselves collectively constitute a powerful group, yet the predominant models exclude media from consideration as a powerful group. Many of the stories that media feature are not sponsored by powerful groups outside of the media themselves.

For example, the 1980s so-called freeway shootings spree, and the 1980s Charles and Carol Stuart case in Boston (in which the pregnant Carol Stuart was allegedly shot and killed by a black male), did not become major, national, news stories because powerful advocacy groups called for media to cover them. Media themselves decided to run these stories. The alleged trend of freeway shootings was a media invention. Finally, for a story to register with an audience it must certainly resonate with one or more cultural themes. But resonance with cultural themes fails to answer the question of why one issue gets attention and another is ignored since there are heterogeneous and contradictory strains within every culture (Beckett 1994) and any number of possible issues could resonate with a given cultural theme. Resonance, in other words, is a necessary, but insufficient, condition. Taken together – dramatic value, mobilization of resources by powerful (non-media) groups, and resonance with cultural themes – these factors do not fully account for why one issue wins out over others.

Media as Mirror

The structural strain model posits that media adopt an issue because they are mirroring the general public’s sentiments. Hence, even when media attention to an issue clearly launched a social problem, the structural strain model does not view this as evidence of media initiating a social problem. Rather, it is interpreted as evidence of media and/or the state acting just like the rest of the polity (e.g., Best 1999, 1990, Surette 1998, Scheingold 1991). Similarly, the interest group model focuses exclusively upon the interest group’s ability to persuade the media (and the state) to recognize its issue. In the interest group model media are viewed as playing a passive role, never initiating attention to a social problem on their own.

The predominant models in social problems treat media as a kind of black box: what comes out of the box is predetermined before it enters the box. In other words, actors other than in the media fundamentally decide what media will feature. Media merely select the best “suitors” according to the predominant models. The fact that media always select, often invent, and frequently frame social problems makes the media, however, a claims-maker of the first stature. Once media’s key role is recognized, social problems accounts’ inability to explain why one issue wins out over another can approach being resolved because the cast of protagonists will no longer be missing one of its leading players.

Explorations into the question of what factors might have entered into media’s decisions to ratify a social problem, and their specific framing of that social problem, are by nature circumstantial arguments, and tendential, rather than determinist. Nonetheless, accounting for why a social problem emerged when it did cannot avoid coming to grips with the question of media’s (and/or the state’s) specific roles. Without such an analysis, determining why a particular issue became a social problem when it did will remain an unanswered question.

The Structural Strain Model

In the structural strain model an alleged generalized public anxiety (Best 1990; Scheingold 1995; Luttwak 1995) finds expression through a surrogate issue. Structural strain lays stress on a collective psychological displacement phenomenon. For instance, crime emerging as the number one social problem would be seen as the public manifesting anxieties about some other issue - such as economic travails or social/political disturbances.

Best (1990), for example, explains the 1980’s moral panic about missing children as a reaction to several different events: widespread public anxiety over the traumas of Watergate, the Vietnam War, Nixon’s resignation, the oil embargo and energy crisis, AIDS threatening to become a global epidemic, America appearing to be in economic decline, and the threat of nuclear war. According to Best, these problems seemed both deeply threatening and impossible for the average person to do anything about. These diverse anxieties were then funneled into the supposedly more manageable issue of protecting the children against unknown individual child-abductors.

Best (1999) also employs a structural strain model. In his book Random Violence, Best tracks what he calls “the general sense that contemporary society is plagued by random violence.” (p. 5) This “general sense” does not come from media activity. Rather, “[a] review of recent press coverage provides evidence of growing concern about random violence.” (p. ) In other words, the press coverage manifests growing concern, but is not a source, or the source of concern. Media coverage does not reflect in any way a preoccupation by media itself with random violence. Instead, media coverage merely reflects a general, concern about random violence by the society as a whole. “There has been a general concern with random violence, and it is that general concern that is the subject of this book.” (p. 7) An absence of agency marks his sentence. Where did this “general” concern come from?

In the following sentence Best summarizes the impact of the theme of random violence. “[T]he notion of randomness distorts what we know about criminal violence. It exaggerates the degree to which violence is patternless, pointless, and increasing. It is imagery calculated to promote fear rather than understanding.” (pp. 21-2. Italics added.) This is an excellent statement about the consequences on public sentiment of certain kinds of media coverage of crime. But anyone who reads Best’s sentence could reasonably ask: calculated by whom? Who has done this calculating? In the structural strain model no one has done this calculating. Instead, “why is this melodramatic imagery so popular? The answer, of course, is that melodrama is powerful…. it challenges our most basic assumptions about social order….If violence is increasing, then everyone is in growing danger.” (p. 22)

This preceding passage bears some scrutiny. When our most basic assumptions about social order are challenged (e.g., the terrorist attacks of September 11, 2001) this constitutes a very disturbing event. Disturbing events are not popular events, outside of the confines of fiction and movies where social order is temporarily ruptured, but importantly, restored in the end. We may watch in horror and fascination as the events are replayed again and again on television, but that does not make the event something that we desire to see happen. Best may be saying that melodramatic imagery is popular among media and among other claims-makers and not the general public. Indeed, if that is his point, media and other claims-makers certainly gravitate towards melodrama and alarmist messages, at least in part because it does deliver the message that everyone is in growing danger, and thus makes their claims seem more pressing. But if Best means the general public, then it should be said that the public does not wish for news messages that disrupt their day-to-day assumptions about the stability of the social order. Explaining the prevalence of melodrama on media and among other claims-makers should not, in short, be attributed wholly or even primarily to public interest.

Scheingold (1991) also employs a structural strain model when he asks why the American criminal justice system is simultaneously the most punitive and the least effective among all the advanced industrialized countries. He concludes that our punitive criminal justice system reflects American individualism. That is, the individual street criminal gets his/her just desserts by being held personally accountable for his or her actions. “The crux is not... that the state somehow imposes retributive scenarios on American popular culture but rather that retribution expresses a collective yearning” (Scheingold 1991: 175).

Plausible as Best’s and Scheingold’s arguments may appear, there are a number of problems with their explanations and with the structural strain model in general. In essence, structural strain lacks a causal mechanism. The absence of a causal mechanism can be discussed with regard to four specific areas. They are the model’s 1) lack of specific predictive power, 2) inability to account for fluctuations in media coverage, 3) assumption that the public naturally selects a more “manageable” scapegoat, and 4) conflating of values (or sentiments) with policy-making. (Most social problems accounts incorporate the interest group model in an attempt to provide a complete causal mechanism. But as I discuss in the second section of this chapter, this does not resolve the problem either.)

1. Lack of Specific Predictive Power

Any number of issues could serve equally well as surrogates for the posited primary public worry. Why this particular issue? The structural strain model, because it asserts a displacement phenomenon, and since a surrogate could assume several different guises, cannot by its nature and by itself indicate the reason why this particular surrogate issue, as opposed to another, was selected. Structural strain in this regard reflects post-hoc reasoning.

2. Inability to Account for Fluctuations in Media Coverage

If public sentiment is merely being mirrored by media attention, as the structural strain model argues, then fluctuations in media coverage of an issue need to be explained. Best (1989; 1990) and Best and Horiuchi (1985) cite data that show a flood of media attention in certain years, and a paucity or absence of said reports in others, on dangers to children and the alleged Halloween candy tampering problem. Inasmuch as the time periods examined all fall within the general time frame of Best and Best and Horiuchi's hypothesized structural strain, how can one then specifically explain why there were more media reports sometimes and not in others?

As Fritz and Altheide (1987) point out, the structural strain explanation should predict a fairly invariant level of media and state attention to a social problem since the source of that attention is an anxious “public:”

"Urban legends are a reaction to pervasive societal stress and strain, and are not regarded as a social construction in and of themselves. Moreover, because the origin of the strain that produces urban legends is the broader social order, then information about such legends is not regarded as a crucial element of the process, merely a catalyst. At most, the information just tells people what they already know to be true. Thus, according to this view, it is not claims-making via the mass media that is important; indeed, the social strain thesis implies that new information, including counterclaimsmaking, would not affect the level of fear. .. [O]ur data suggest otherwise...."(486-487).

The data which Fritz and Athleide refer to is striking. Subjects were polled in a pre-test about their assessment of the danger of stranger child abduction. After the pretest they viewed a videotape that revealed the complexity of the missing children problem, and which called into question the widely propagated figure of 1.5 million missing children annually in the U.S. After viewing this videotape, there was a marked drop in the percentages now saying that they thought stranger child abduction was a serious problem. 61% said it was a serious problem before testing, only 33.3% said it was a serious problem after viewing the counterclaim videotape. While 74% of the cohort showed a concern of over 21 in the seriousness scale before testing, only 28% scored over 21 on the same scale after testing (Fritz and Athleide 1987: 483). In contrast to the structural strain model, Fritz and Altheide (1987: 486) conclude that mass media “contribute the most to the claims-making process.”

3. The Selection of a More “Manageable” Scapegoat

Within the structural strain model is a presumption that scapegoats are chosen by an anxious public as a means of selecting an easier and more manageable enemy. What makes these particular surrogates more manageable? In what sense does the focus on stranger criminals, who make up the minority of criminal offenders, constitute the construction of an easier and more manageable enemy? What makes such worries a more manageable fear than domestic violence hazards, for example? When a woman, for example, faces a nine-fold greater danger of physical injury from someone in her own household compared to someone in the streets, how do enhanced worries about stranger-on-stranger street crimes make her life more manageable? How is it, in other words, that displaced and, particularly, added-on, anxieties perform a psychological function for an anxious public?

The public scapegoats youth gang bangers, for example, because they are allegedly easier for the various publics to cope with psychologically than directly dealing with, say, job loss and job insecurity. There is a considerable body of research on the effects of heightened crime fears (e.g., Brooks 1974; McDermott and Blackstone 1995; Clemente and Kleiman 1977; Wilson 1968; Yin 1980). These studies have found that anxieties are increased by such reporting over and above what the various publics already feel.

In other words, rather than creating an enhanced sense of personal and/or collective mastery over forces that have come to be seen as outside of individual and/or collective control, crime scares generate greater anxiety and intensified feelings of insecurity. While it is not necessarily a structural strain argument that the surrogate issue relieves any psychic strain, the ready assumption that a deviant class as scapegoat makes for a more manageable enemy skips over a closer examination of why more systemic sources for the problem are usually not targeted as scapegoats.

What these particular scapegoats have in common - vicious, amoral, stranger criminals - is that focusing attention on them demobilizes the public, leaving most possible action against them in social control agents’ hands. Is settling on deviants as enemies inevitable and spontaneous? Such a selection process is better understood as contingent, subject to the presence or absence of contending interpretations from different identifiable protagonists, rather than inevitable.

The structural strain model tends to see downward scapegoating as automatic. But the process by which a scapegoat is selected is subject to social processes layered by power differentials in which elites possess unequal powers to direct blame away from themselves. The structural strain model leaves out of its analysis these differentials of power. It depoliticizes the question of the selection of a scapegoat, and naturalizes the process, rather than identifying it as contingent.

4. Conflating Cultural Outlooks with Policy-Making

Another of the weaknesses in the structural strain model is that while the model has heuristic value for identifying the receptivity of various publics to symbolic political movements, it tends to conflate cultural outlooks with policy-making. That is, the structural strain model identifies a cultural outlook (resentments born of economic and/or social/political anxieties) and assumes a translation of that value stance into a matching social policy (more punitive social policies).

Luttwak, for example, argues that the:

"… insecure majority ... vents its anger and resentment by punishing, restricting, and prohibiting everything it can. The most blatant symptom is the insatiable demand for tougher criminal laws, longer prison sentences, mandatory life sentences for repeat offenders, more and prompter executions, and harsher forms of detention (including, of late, chain gangs). Politicians ... have heard the people, and the result is a mass of new federal and state legislation that will greatly add to the staggering number of Americans already behind bars. This, however, is merely part of a much broader urge to prohibit and punish...all expressions of the same deep resentment, turned against convenient targets." (1995: 15-16)

Even if the assumption is made that a cultural outlook is necessarily translated into public policy, however, a general value or sentiment has more than one possible matching social policy.

There are a number of possible policies that could be matched up to a given value or sentiment. Thus, even if the structural strain model is correct in identifying a generalized disquiet among the populace, this still does not answer the question: why this particular social problem rather than another?

The structural strain model has difficulty answering this question with precision because it conflates initiative with receptivity. That is, it assumes that public receptivity is equivalent to public initiative in the emergence of social problems (Beckett 1994). But just because the public shows interest in an issue does not mean that the public originated attention to that issue. Thus, the prevalence of media news reports about a putative surge in random violence that Best (1999) notes does not mean that the public has demanded this kind of news coverage, and that media attention to this “problem” is attributable to public interest. Given the heterogeneity, diffuseness, and contradictoriness of public sentiment, any number of possible issues could equally well garner public attention. The structural strain argument misses a crucial distinction: the difference between public prerogative and public sovereignty (Beckett 1994). Choosing one item from among a set of alternatives is not the same thing as deciding what the alternatives will be in the first place.

"[C]apitalism would work to present consumer sovereignty as the equivalent of freedom, in the common view and the common parlance. ('If you don’t like TV, turn it off.’ ‘If you don’t like cars, don’t drive them.’ ‘If you don’t like it here, go back to Russia.’ ‘If you don’t like Crest, buy Gleem.’ ‘If you don’t like Republican, vote Democratic.’)" (Gitlin 1978).

Similarly, Sassoon (1996) compares the public’s options in determining policy to a person deciding what ethnic cuisine to prepare based on what raw materials can be found at the local grocery. If that grocery is mainstream, one is more likely to end up finding the necessary ingredients for an Italian meal even if the intent was to prepare Vietnamese cuisine. The public may, for example, choose street crime stories over other offered alternatives, but it does not have a say as to what those other alternatives are. Who decides what those alternatives will be?

The Interest Group Model

Goode and Ben-Yehuda (1994) argue that the interest group model answers the question of who decides the alternative. Does it? There are certainly times when an interest group initiated attention to an issue, for example, Mothers Against Drunk Drivers launched the anti-drinking and driving campaign. At other times, however, interest groups were either not pivotal in the launching of a social problem or were not even in existence (e.g., the Drug War declared by Ronald Reagan in 1986; the demonization of Iraq and Saddham Hussein, currently and prior to the war on Iraq in 1990; “road rage” in the 1990s; the home health care issue in the early 1980s (Cook, Lomax, Tyler, et al 1983)). Furthermore, even when an interest group was present and active, this still does not explain why this particular interest group succeeded and several others simultaneously pressing their claims did not.

The answer to this problem has traditionally been reached by circular logic: this claim became celebrated, therefore the claims-makers’ typification strategy must have been more effective than another claims-maker's. Lowney and Best (1995) have attempted to remedy this problem of retroactive reasoning by following the early stages of an issue – namely, the stalking of women -- through to its eventual emergence as a social problem. They conclude that constructions of “social problems become successful when typified in ways consistent with the larger culture’s concerns, and when they mobilize significant organizational support.” (Lowney and Best 1995: 52). In addition, they also point to contingent events as relevant. For example, the drug-induced death of Len Bias played an unmistakable role in the onset of the drug war.

What are those “larger culture’s concerns”? Lowney and Best cite Hilgranter and Bost (1988:64) who cite claims which “relate to deep mythic themes,” Gamson and Modigliani (1989:5-6) who cite “cultural resonance” and Best (1991:50) “cultural resources.” The cultural resonance argument really simply reiterates the structural strain paradigm: an issue becomes celebrated because it strikes a responsive chord in the public consciousness. A diverse array of potential social problems, however, resonate with deep cultural themes.

Lowney and Best (1995) propose to resolve this problem by incorporating the mobilization of organizational resources – the winning group succeeds by mobilizing more organizational support. This resolution, however, does not answer the question of whether interest groups become more visible because they mobilize more support, or because the media and/or the state independently decide to seek out interest group opinions more (e.g., by holding Congressional hearings or running a series of news articles).

A persuasive case can be made for Lowney and Best’s third factor, contingent events, as a “tie-breaker.” In some cases contingent events have clearly played an important role. Nevertheless, events still do not speak for themselves. They must always be interpreted. A presidential assassination, a 7.1 earthquake, a world war, the 911 attacks, the Asian Tsunami of 2004 and a few other similar catastrophes can be guaranteed to be headline material, but most other events cannot. The mugging panic that Hall, Critcher, Jefferson et al (1978) chronicle, or the panic about crimes against the elderly (Fishman 1978), were both “problems” in which the material events paled next to the furor created about them.

Legitimation as an active rather than passive process

Spector and Kitsuse (1973) argue that for an issue to become a social problem, not only must an interest group press an issue, but the group and its issue must be legitimated by media and/or the state. Spector and Kitsuse’s discussion of what leads media and/or the state to legitimate a social problem all concern attributes of the interest group, and none regarding the media and the state. They focus on three points:

1. The power of the interest group. This includes the group’s economic resources, the size of its membership, how large its constituency is, and how organized the group is.

2. The nature of the interest group’s claim. Spector and Kitsuse observe that the more specific the group’s claim, the better. More diffuse or general claims are more difficult to satisfy, and less likely to be legitimated by media and/or the state.

3. The interest group’s strategy and mechanisms for pressing claims. Spector and Kitsuse (and later, Ibarra and Kitsuse 1993) emphasize the rhetorical strategy pursued by claims-makers - how dramatic they make their typifications - as well as the effectiveness of their lobbying efforts and the kinds of allies they can muster.

If one further examines Spector and Kitsuse’s argument, in what they call a natural history of social problems, it becomes apparent that every step in the process applies with even greater force to the media and the state than to interest groups. Spector and Kitsuse identify four stages to that history. In stage one an interest group expresses grievances and makes claims upon government or other official or influential institutions. In stage two the claims-makers’ claims are acknowledged and the group’s standing to make said demands are legitimated by media and/or government. In stage three the interest group claims that official responses are problematic, and in stage four the interest group may seek ways outside of institutional channels to pursue their claims further.

With regard to stage one, a governmental agency or official, or a major media outlet, are ideally positioned to make claims and express grievances - upon themselves. Ronald Reagan as president, for instance, can declare a war on drugs through making a widely publicized speech. No middle-level group need necessarily have gotten involved in the process in its initial stages. No middle-level group did get involved in the launching of the War on Drugs. Alternatively, one or more major media outlets may decide to make an issue of missing children, or freeway shootings, or allegedly tampered Halloween candy, or an epidemic of juvenile homicides.

By this process, stage two in which the claimants’ issue and standing to make said claim are legitimated is already fulfilled by definition. The agencies who must place the imprimatur of legitimacy on a given issue have themselves initiated the issue. In stage three, media and/or state actors will argue that business as usual is somehow inadequate or intolerable regarding issue X, and in stage four, if blocked by recalcitrant sections of the state, major media and/or state actors can appeal to the broader public to by-pass the bureaucratic logjam.

Best argues that inattention to media’s specific role in the social problems process is valid because the media rarely initiate a social problem: “While press coverage may be especially visible and influential, claims-making rarely begins with the media; the press usually covers other, primary claims-makers, and its coverage is a secondary-claim.” Best (1990: 109) Speaking specifically of the missing children scare of the 1980s, Best states:

"In the competition among claims-makers, those able to present claims in compelling rhetoric have an advantage, and the new child savers found it easy to define their cause in terms of threatened, vulnerable innocents. These emotion-laden images encouraged the mass media to incorporate threats to children into news coverage and popular culture; the media’s treatment further emphasized the issue’s dramatic elements. In turn, the portraits drawn by primary claims-makers and relayed through the media’s secondary claims evoked reactions from the public." (Best 1990: 174-5)

Note the relative importance Best gives herein to the primary claims-makers (child savers), secondary claims-makers (media and the state), and the public. According to Best, initiative lay in primary claims-makers’ hands, whose “emotion laden images encouraged the mass media” to carry stories about threatened children. The media did not, in other words, exercise initiative or choice in deciding whether or not to run a spate of stories about threatened children. The media acted as a conduit between the primary claims-makers and the broader public, merely “relaying” the portraits drawn by the primary claims-makers, albeit the media do “emphasize” an issue’s dramatic elements.

Dramatic value certainly enters into media decisions, but to say that media carry a story because it is dramatic is a truism - it tells us nothing new. Best appears to assume that there are degrees of drama, and the most dramatically presented issues will receive media attention. The actual process is far more complicated. Drama is in fact a necessary, but insufficient, condition for a story to be chosen. Properly framed, in fact, any good conflict on either an interpersonal level or larger group level is dramatic. There are many and diverse dramatic stories available to media on any given day, or any given hour, from which media must choose. The question that ought to be asked is: why was this dramatic story chosen over that (equally or even more) dramatic story?

To begin with, decisions about whether to carry a story in the first place are always subject to ideological considerations. One example of how dramatic value does not mainly determine whether a story will be featured or not is cited by Cose (1993): the so-called Central Park Jogger case. In the same week that headlines across the country covered the so-called “wilding” rape and beating of a white female investment banker jogger in Central Park, New York, in the 1980s, an even more dramatic story was buried in the inside section of the New York Times. In that incident, a woman who had gone to the top of her high-rise office building to watch the sunset was set upon by two men, raped, and then thrown off the top of the building. Incredibly, she caught a wire in mid-air on the way down, and was screaming for help, dangling naked, and eventually rescued.

Certainly in terms of dramatic elements, the second story has more than the first. The difference, however, was that the woman in the second story was black and the incident occurred in the Bronx. In other words, media decisions about “drama” are shot through with differential treatment based upon race, gender, class, and often age. These are hardly neutral categories. They are in fact highly political. An interest group that presents a truly dramatic and compelling claim to the media about the plight of, say, lower class black females, is not likely to prevail over less dramatic claims made by a group making claims about the plight of upper middle class females.

Structural strain and interest group model advocates do explicitly note that some victims are more likely to receive media attention than others. This recognition is not, however, taken any further. It exists awkwardly within models that do not treat media as a key claims-maker.

Even more important than who started a claim is the fact that media and the state always exercise the power of deciding how to frame a story or issue. Since the issue frame that dominates the public arena fundamentally dictates how a problem is addressed, the issue frame (along with whether or not an issue makes it to social problem status in the first place) matters more than any other factor in the social problems game. The structural strain model precludes a thorough investigation of the process by which an issue is framed since it presumes that any claims arise out of the whole of the society, and therefore holding any party specifically responsible is a useless exercise.

The interest group model treats media and the state as secondary actors. That is, the interest group model asserts that media and the state simply adopt the problem frame presented them by the successful claims-makers. Since most social problems actually originate from the media or the state, this point is often moot. As for those issues that do not, while the issue frame is sometimes adopted from that pressed by the initiating group, it does not happen as any kind of rule. Gitlin (1980), for example, chronicles the dilemmas that the anti-war movement in the 1960s faced when the media finally covered the movement in any great detail since the problem frame used by the media clashed in important ways with the problem frame the movement itself advocated.

Structural strain and the interest group model are both based on pluralist democratic theory. That is, they assume that media and the state do not strike out on their own, initiating social problems themselves, bringing the public along with them, because public policy is assumed to always reflect public sentiment. Put another way, democracy dictates that government policies, and media attention, manifest popular desires. When public policies take a conservative (or liberal) turn, commentators nearly automatically assume that this can be attributed to a shift in public mood, with policy makers and media organizations sniffing the winds and adjusting their approaches accordingly. This belief that the public ultimately determines public policy and what media features in news and entertainment is so widely and strongly held that it escapes scrutiny.

Who launches a specific social problem at a particular point in time is an empirical question as much as a theoretical question. Theoretical questions are important for a number of reasons. Among these is the fact that a specific theory can tend to obscure the specific evolution of an issue. For example, if one applies a structural strain or Spector and Kitsuse model, issues that actually originate from media and/or the state might be mistakenly and automatically understood as originating from interest groups (or from structural strain).

Even when an interest group is active and prominent, whether or not an issue becomes a social problem is still always subject to media and/or state decisions. An issue becomes a social problem only after the media and/or state actors decide that it will. As Parenti (1970: 501) points out, “[o]ne of the most important aspects of power is not to prevail in a struggle but to determine the agenda of struggle - to determine whether certain questions ever reach the competition stage” (Cited by Gaventa, 1980 at 10). Or, as advanced by Cohen (1963), who was speaking specifically of the media, the media does not so much tell us what to think, as tell us what to think about.

By their nature interest group claim-makers’ claims are not yet before the various publics as a pressing social problem (awaiting its politicization within the public political arena and/or its publicization by major media before becoming a social problem). Media and/or the state are, therefore, not under heavy social pressures from the various publics to cover or deal with an issue. The general polity does not know about, or is not particularly aroused about the issue yet. Media and/or the state, or elites, have the power, in other words, of vetoing an issue simply through their silence. It may be an exceptionally important issue for the public(s), but becoming politically aroused over it invariably requires its being focused in the public arena in ways that only media and/or the state have the ability to provide.

The fact that an issue must be legitimated by media and/or the state is something which has been recognized by all scholars operating in the social problems tradition. But the full implications of this have not always been drawn. Legitimation by media and/or the state has generally been viewed as a passive, as opposed to an active, activity by media and/or the state. But why should media and/or the state’ role be seen exclusively as passive, and not variable - including at times, highly active? There is no reason in principle why not, other than the assumptions inherent in classical and pluralist democratic theory that media and the state invariably respond and react to public sentiment, rather than ever initiating any social problems.

The power exercised by media and/or the state is not, of course, unilateral and undivided. The exercise of influence and initiative in the making of a social problem is an interactive one. But it is a weighted interaction in which, because of their privileged positioning and social role, elites possess distinct advantages, not the least of which is that of initiative (Lippmann 1922; Beckett 1994). This is a rather different view than the rather naive view that concludes that policy shifts or marked changes in media coverage necessarily reflect preceding and corresponding changes in public sentiment.

The approach which focuses on interest group claims-maker's strategy puts the initiative in the hands of the interest groups. So the question of how the media decides whether it will cover an issue or not remains under-explored in the interest group model as much as it does in the structural strain model. Further, the question of how and why the media elect to pursue a particular approach to the issue merits greater attention than either the structural strain or the interest group models give it. Putting the stress on the interest group’s rhetorical skills, therefore, obviates the crucial role that media and/or the state play in the social problem definitional process. In the elite constructionist model, it is precisely the media and/or the state that are the focus of attention. The elite constructionist model has attracted relatively few adherents in social problems studies in recent decades. It deserves much more attention.