Do Even Journalists Support Media Restrictions? Presumed Political Media Influences and the Consequences

Abstract

How journalists perceive the influence of different types of media and how this perception affects the journalists’ attitude toward supporting censorship was explored by conducting a survey among German journalists (N = 1,688). As predicted, the stronger journalists perceive the political influence of the media to be on the public, the more they approve of, for example, demands for restricting this influence. At the same time, behavioral intentions regarding censorship depend less on the perception of the media influences directly, and more on attitude toward these restrictions.

Keywords

journalists third-person effect influence of presumed media influence approach censorship

As a rule, journalists react vehemently when it comes to fending off (alleged) threats to press freedom. In Germany, for example, searches of editorial offices for investigations by a public prosecutor or attempted interferences by politicians are heavily criticized by journalists and their associations. Thus, at first glance, German journalists at least agree in rejecting state intervention in the media’s freedom of expression. They probably share this opinion with journalists from other countries, because independence and the right to freedom of expression are central preconditions for the media’s ability to fulfill their obligations in democratic societies.¹ Thus, for instance, Hanitzsch and Mellado have shown that journalists from eighteen countries perceive influence from politicians and that journalists see this influence as a threat to their autonomy.²

How do journalists react when they perceive media influences on others, which are strong and problematic at the same time? Research on the third-person effect³ and on the influence of presumed media influence approach⁴ has shown that people who ascribe a strong and negative media influence on others tend to approve of censorship.⁵ However, the question whether these findings can be applied to journalists remains unanswered. The consequences of presumed media influence, until now, were mainly explored regarding the public,⁶ but also special groups, such as politicians,⁷ scientists,⁸ and physicians.⁹ Remarkably, this is not the case for the creators of media content, not even regarding censorship approval. Yet, for several reasons, it does not appear feasible to simply apply the findings for the general relationship between perceived media influences and demands for censorship to journalists. First, journalists across the globe see it as their task to provide political information and to act as a watchdog of the government.¹⁰ In this respect, journalists’ demands to limit their own influence would, most likely, lead to cognitive dissonance and to normative conflicts. Thus, journalists are likely to have very little interest—even less so than the public—in having restrictions imposed on their own work. Accordingly, their reactions to attempted interference with their autonomy are—as noted before—quite fierce. Second, the perception of media influences differs between journalists and the general public: Although journalists do not perceive strong media influences on the public, they perceive a stronger media impact on themselves (and also on politicians) than the public. In addition, journalists evaluate media influences more positively than the public.¹¹ Thus, it is not unlikely that journalists’ reactions as a consequence of perceived influences differ from the public’s. Third, journalists generally have relatively high educational qualifications and an above-average interest in politics.¹² These variables can also influence the approval of censorship.¹³ Finally, censorship demands from the affected profession itself—which has also good possibilities to distribute its beliefs and to influence public opinion—would have a higher relevance regarding media policy than demands from people in most other professions. Thus, it is necessary to examine if even journalists support media restrictions as a consequence of perceived media influences on others.

Moreover, censorship approval, thus far, has only rarely been examined within the context of political communication.¹⁴ However, media coverage of politics forms a crucial section of democratic societies, and restrictions would have serious consequences.

This paper, therefore, focuses on two questions: How do journalists in Germany assess political media influence on themselves and on others? How does this perception affect the journalists’ support for censorship? The assumed influences of traditional media (such as television and newspapers) and online media are analyzed separately.

Presumed Media Influences

Much research concerning attitudes and intended behavior toward media censorship regarding the third-person effect has been conducted. This approach contains two central assumptions.¹⁵ First, individuals assume that the media and media content influence other people more strongly than themselves (third-person perception). Second, this difference in perception is presumed to have consequences for beliefs, attitudes, and behavioral patterns (third-person behavior).

The assumption regarding third-person perception has been tested in the widest range of contexts and has been confirmed in many studies.¹⁶ Tsfati and Livio,¹⁷ for instance, showed that journalists, too, rate media influence on others higher than on journalists themselves.¹⁸ In the light of these results, the following hypothesis is applied to the present context:

H1: Journalists perceive that the media’s political influence on politicians, other journalists, and the public is stronger than the influence on journalists themselves.

Presumed Media Influences and Support for Censorship

In research on the behavioral component of the third-person effect, support for censorship is the “gold standard.”¹⁹ There are numerous hints that people approve of censorship of problematic media content if the individuals ascribe to this content a stronger influence on third persons than on themselves.²⁰ This assumption can be applied, for example, to pornography,²¹ advertising,²² depictions of violence on television,²³ or violent song lyrics.²⁴ In contrast, other studies failed to prove that third-person perceptions were positively related to censorship demands.²⁵ This is especially true for political media content: third-person perceptions related to the influence of news, political campaigns, and legal procedures did not have any effects (or only slight effects) on support for censorship in several studies.²⁶ The authors explained these findings with relevance ascribed to such content from the point of view of democratic theory.

Another explanation for the inconsistent findings would be that the difference in perception between oneself and others is an inappropriate predictor for behavioral consequences, as it “does not distinguish between those who perceive media content to have high influence on themselves and on others and those who perceive media content to have low influence on themselves and on others.”²⁷ Moreover, if perceived media effects on oneself and on other persons are merged into one variable, it is not possible to determine their separate effects on support for restrictions.²⁸ Consequently, some authors argue that, in all likelihood, it is often just the strongly perceived effect on others that accounts for the observed consequences.²⁹

Thus, many recent studies in this field of research concentrate on perceived media influence on others. The authors follow the “influence of presumed media influence” approach that states that people perceive media content influences on others and react to that perception.³⁰ Wei and Lo³¹ showed, for example, that perceived media influence on others serves as a better predictor regarding support for censorship of political attack ads than does third-person perception. Other studies showed as well that the presumed political media influence on others strengthens support for censorship.³² Hence, in the following, perceived influence on others is used as a predictor of support for censorship.

Censorship demands are mostly accepted when they involve restricting unwanted media content. The perceived political media influences per se are presumably not rated negatively. Therefore, whether the perception of strong political media influences leads to increased support for censorship, especially as far as journalists are concerned, is an open question. However, several factors can bring about a skeptical view (even in journalists) of political media influence: for instance, journalists can presume that political coverage is biased and sides with a specific position. These hostile media perceptions³³ can intensify the perception of media influences on others that are considered strong and negative at the same time.³⁴ Even if journalists, in general, should have a comparatively positive opinion of media influence,³⁵ it seems plausible that they sometimes share such critical opinions. This might also affect journalists’ attitude toward censorship:

H2: The stronger and the more negative journalists perceive political media influence to be on others, the more (a) they will demand a restriction of this influence, (b) relevant they will find stricter media control by the state, (c) they will intend to support stricter media control by the state.

This hypothesis addresses the general assumption on the relationship between the perception of political media influences and censorship demands. In the following, this basic assumption will be differentiated by the perceived impact on different groups and by the perceived impact of different media. Furthermore, the relationship between perceptions and censorship intention will be considered in more detail by considering attitude toward behavior as a mediator variable.³⁶

Presumed Media Influences on Different Groups and Support for Censorship

A widely held attitude is that censorship is accepted to protect defenseless others.³⁷ According to the logics of the third-person effect, the anonymous masses are regarded as particularly susceptible to negative media content and are thought to need protection.³⁸ The majority of the presumed influence studies on demands for censorship therefore examined media influence on “the public.” In most cases, researchers did not make any further differentiations,³⁹ even though it is logical that support for control measures depends not only on the perceived strength of media influence, but also on whom this influence is ascribed.

In a survey of journalists concerning political media influence, testing presumed influences on the public and on politicians and other journalists is suitable.⁴⁰ Politicians often deal with the media for professional reasons. They know media coverage can be useful; that is one reason why politicians are, as a rule, familiar with journalistic working methods and other media-related logics.⁴¹ Many politicians even communicate directly with citizens via online media.⁴² Thus, we assume that politicians possess comparatively high media competence and (what is more important in the present context) are ascribed with that competence. Similar qualities should be applied to journalists, who, due to their profession and their above-average level of education,⁴³ are probably considered particularly media-competent by their colleagues.

If journalists act on the assumption that politicians and other journalists are better at handling media content than the public, both groups should need comparatively less protection against negative media influences. Thus, it is hypothesized that

H3: Perceived strong and negative media influences on the public lead to stronger support for censorship among journalists than perceived strong and negative media influences on politicians and journalists.

Presumed Media Influences of Different Media and Support for Censorship

Furthermore, the general assumption concerning presumed media influences and support for censorship, as expressed in H2, can be differentiated for different media. For instance, it is likely that perceived influences of traditional mass media and perceived influences of the Internet relate differently to censorship support. The Internet has not only changed journalists’ working methods, but also challenged established standards and the professional views and attitudes of journalists.⁴⁴ For example, the Internet offers users the possibility to find almost all kinds of information autonomously, and to disseminate information themselves. Journalists thus cease to be the only gatekeepers. In the face of such changes, journalists have “attempted to retain the cultural authority of the traditional journalism, and refute online news content provided by the Internet.”⁴⁵ The emergence of the Internet thus caused negative reactions among journalists depicting online media as non-credible, unreliable, or unprofessional.⁴⁶

Such skepticism has probably decreased. However, recent studies show that journalists still have some mistrust of the Internet even though they perceive it as useful for their own work.⁴⁷ This skeptical attitude toward online content should bring about a situation in which strong and negatively perceived political influences through online media evoke the impulse to control or restrict these influences. In contrast, more tolerance should be shown to traditional media.

H4: Perceived strong and negative political effects of online media on others have a stronger influence on support for censorship among journalists than perceived strong and negative effects of traditional media.

Presumed Media Influences, Attitudes toward Censorship, and Support for Censorship

Finally, there are theoretical and empirical indications that perceived media effects do not directly influence censorship intentions, but are mediated by the attitude toward censorship. Typically, studies based on the third-person effect or the influence of presumed media influence approach do not examine actual support for censorship (e.g., the signing of relevant petitions), but only the attitude toward such measures. Nevertheless, behavior can hardly be predicted by attitudes only.⁴⁸ According to the theory of reasoned action,⁴⁹ the intention to do something is a significantly better predictor of behavior. According to the same theory, this intention is determined by the attitude toward the behavior and by perceived requirements of the social environment. These assumptions are empirically well-proven⁵⁰ and have been tested—at least in part—in the context of censorship. For example, Zhao and Cai⁵¹ demonstrated that the perception of media influence does not affect the intention to support media control directly, but rather is mediated by the attitude toward censorship.⁵² This result should also be valid for journalists and in the context of political communication.

H5: The intention to support control measures predominantly depends upon the attitude toward control measures, and is only indirectly influenced by perceived media influences.

Method

Sample and Procedure

To test the hypotheses, a standardized online survey among German journalists was conducted in March and April 2012. It took place in cooperation with the Deutsche Journalistinnen- und Journalisten-Union (German Union of Journalists, approximately twenty-two thousand members) and Deutscher Journalisten-Verband (German Federation of Journalists, approximately thirty-eight thousand members), the two main German journalist unions. They represent a large percentage of the community of journalists working in Germany.⁵³

The unions contacted members via e-mail, calling on them to participate in the study. Respondents were informed that participation was voluntary. A total of 2,168 people participated. However, only 1,688 journalists who work for media organizations (including journalists who work as freelancers) were included in the analyses; 480 respondents who work mainly in the public relations sector for businesses and organizations were excluded from the sample. As an incentive for participation, one euro was donated to a charity organization for each completed questionnaire. Precise data on the response or cooperation rate are not available.⁵⁴ There is no information on how many journalists were actually reached through the two associations. On the other hand, it is unknown how many journalists are members of both associations and thus may have received the call for participation twice.

The study falls short of being a representative sample of German journalists.⁵⁵ This limits the scope and validity of the study and should be considered in interpreting the results.

The majority of the participants (57.0%) worked exclusively in classic mass media. In addition, 43.0% worked, at least in part, for online media.⁵⁶ The sample included more men (62.6%) than women (37.4%). The participants were between 17 and 87 years old (M = 49.18; SD = 11.00), and 91.6% named either the Abitur (German university entrance qualification) or a university diploma as their highest education degree; lower levels of education were thus barely represented. About two-thirds of the respondents were freelancers (67.3%), one-third were in a salaried position (32.7%), 46.8% worked primarily for a newspaper or magazine, and roughly 15% worked for radio or television broadcasters. Only 8.9% of the respondents worked predominantly for online services; the remaining journalists worked for other media or news providers (e.g., news agencies). In addition, 43.1% of the respondents indicated they covered politics, 45.7% culture, 31.4% economics, and 34.2% local news (multiple answers were possible). In contrast, fewer respondents covered science (21.2%), media (14.5%), or sports (12.2%).

Measures

Presumed media influences

One item was applied to measure the extent to which the journalists perceive the political influence of traditional media on themselves, on other journalists, on politicians, and on the German public in general: “When thinking of television and the press, how strong do you consider their political influence, namely, the influence on [ . . . ]?” (1 = no influence to 5 = very large influence). In the same way, the respondents were also asked to estimate the extent of the Internet’s political influence on the different groups. In addition, analogous items were used to rate the specific influence of Facebook, Twitter, and Internet news sites.

Evaluation of presumed media influence

The evaluation of the presumed influences also depended upon differentiating between offline and online media. The following question was asked, “How negative or positive do you consider the political influence of [ . . . ], namely, the influence on [ . . . ]?” (1 = very negative to 5 = very positive). Respondents rated the influences traditional and online media exert on themselves (press/television: M = 3.13, SD = 0.74; Internet: M = 3.10, SD = 0.78), on other journalists (press/television: M = 3.03, SD = 0.80; Internet: M = 3.07, SD = 0.79), on politicians (press/television: M = 3. 02, SD = 0.91; Internet: M = 2.99, SD = 0.78), and on the German public (press/television: M = 3.10, SD = 0.86; Internet: M = 2.99, SD = 0.77).

Support for censorship

First, attitude toward restricting political media influence was measured. One item captured the attitude toward restricting the influence of print media and television: “How much do you agree with the following statement: The political influence of television and the press should be restricted?” (1 = absolutely disagree to 5 = absolutely agree; M = 1.60, SD = 1.04). Attitude toward restricting the influence of the Internet was measured with a corresponding item (M = 1.77, SD = 1.09).

In addition, one item determined attitude toward enhanced state control of the Internet: “How important do you consider an initiative in favor of stricter state control of the Internet?” (1 = very unimportant to 5 = very important; M = 2.43, SD = 1.32).

Finally, a possible behavioral intention of journalists was considered concerning the intention to support an initiative in favor of stricter state control of the Internet. The item was formulated as follows: “Would you support such an initiative for stricter control of the Internet by signing a petition?” (1 = by no means to 5 = definitely; M = 2.12, SD = 1.34).

Covariates

The age of the respondents, years of education, sex, and the percentage of journalistic activities for online media out of the total journalistic workload were covariates. Furthermore, journalists’ interest in politics was assessed (1 = not at all to 5 = very strongly; M = 4.42, SD = .71).

Results

H1

The first assumption concerns the differences in the perception of political media influence. Table 1 shows the respective values of the influence ascribed to different media or online offerings combined with different (groups of) people. As expected, third-person perceptions can be discerned throughout. The journalists invariably perceive themselves as less affected than politicians, other journalists, and the public. This is true for classic media such as television or newspapers, the Internet in general, and different online content. Repeated analyses of variance with Bonferroni post hoc tests showed that all differences between the influence ascribed to oneself and to the other groups are significant (p < .001). Thus, H1 was supported. Moreover, it is striking that particularly strong influences of the Internet are ascribed for other journalists and politicians.⁵⁷ The political effect of specific online offerings, such as news sites or Twitter, is also considered particularly strong when it comes to other journalists and to politicians. In contrast, the influence of Facebook is considered as strong on the public as on the two other groups.

Table 1.

Mean Estimates (Standard Deviations) of Presumed Influences of Different Media.

	Presumed influences on:
	Self	Politicians	Other journalists	Public
Intensity of presumed influences
Television and press	3.28 (.98)	4.07 (.89)	3.75 (.92)	4.17 (.77)
Internet	3.00 (1.13)	3.32 (.93)	3.57 (.91)	3.20 (.89)
Facebook	1.78 (1.02)	2.79 (.94)	2.80 (.92)	2.81 (.93)
News websites	3.01 (1.15)	3.72 (.91)	3.76 (.92)	3.12 (.91)
Twitter	1.37 (.77)	2.63 (.98)	2.64 (.96)	2.17 (.83)

Note. All items were measured on a five-level scale; higher numbers indicate stronger perceived impact; n = 1,636-1,684.

H2

It was assumed that journalists who perceive a strong and negatively evaluated media influence would demand restrictions on media influence (H2a), consider stricter state control of media more relevant (H2b), and have a stronger intention to support such media control (H2c). Hierarchical regression analyses were conducted to test these assumptions. H2a was tested regarding press/television and Internet. H2b and H2c involved only the Internet. In the first step, control variables such as age, years of education, sex (dummy coded; 0 = male, 1 = female), political interest, and the type of media the journalists work for (dummy coded; 0 = only for classic media, 1 = [also] for online media) were used as predictors. In the second step, the perceived influence strength of the Internet or of press and television on the public, on other journalists, and on politicians was included in the model, as well as the evaluation of these influences. To test the interaction effect of both variables (perceived influences and evaluation of perceived influence), the product variables of the respective predictors were included in the model.⁵⁸ Tables 2 and 3 show the results of the analyses.

Table 2.

Hierarchical Regressions—Influences on Approval of the Restriction of the Internet Influence, on the Relevance of, and on the Support for Internet Control.

	Restriction of Internet influence (n = 1,596)β	Relevance of Internet control (n = 1,589)β	Support for Internet control (n = 1,595)β
Block 1
Age	.14***	.24***	.25***
Years of education	−.11***	−.07**	−.09***
Sex (0 = male, 1 = female)	.00	.17***	.14***
Interest in politics	−.18***	−.08**	−.06*
Primary areas of journalistic work (0 = traditional media, 1 = online media)	−.07**	−.08**	−.06*
R²	.07***	.10***	.09***
Block 2
Age	.13***	.23***	.25***
Years of education	−.09***	−.04	−.06**
Sex (0 = male, 1 = female)	−.01	.15***	.12***
Interest in politics	−.15***	−.06*	−.05*
Primary areas of journalistic work (0 = traditional media, 1 = online media)	−.04	−.07**	−.06*
Presumed influence of the Internet on public	.09**	.07*	.09**
Presumed influence of the Internet on journalists	−.08**	−.07*	−.06*
Presumed influence of the Internet on politicians	.04	.08**	.07**
Evaluation of presumed influence of the Internet on public	.16***	.13***	.10**
Evaluation of presumed influence of the Internet on journalists	.08*	.02	.04
Evaluation of presumed influence of the Internet on politicians	−.02	.00	−.02
Presumed influence of the Internet on public × evaluation of influence	.01	.00	−.00
Presumed influence of the Internet on journalists × evaluation of influence	−.02	−.04	−.03
Presumed influence of the Internet on politicians × evaluation of influence	.00	−.02	−.04
R² change	.05***	.03***	.03***
Total R²	.12***	.13***	.12***

Note. Coding: education = years of education, evaluation of influence (recoded: 1 = very positive to 5 = very negative); all variables of perception and evaluation of influence are centered.

p < .05. **p < .01. ***p < .001.

Table 3.

Hierarchical Regression—Influences on Approval of the Restriction of Press and Television.

	Restriction of press and TV (n = 1,599)β
Block 1
Age	.05*
Years of education	−.10***
Sex (0 = male, 1 = female)	.02
Interest in politics	−.13***
Primary areas of journalistic work (0 = traditional media, 1 = online media)	.02
R²	.03***
Block 2
Age	.02
Years of education	−.06**
Sex (0 = male, 1 = female)	−.01
Interest in politics	−.10***
Primary areas of journalistic work (0 = traditional media, 1 = online media)	−.01
Presumed influence of press and TV on public	.05*
Presumed influence of press and TV on journalists	−.08**
Presumed influence of press and TV on politicians	−.05
Evaluation of presumed influence of press and TV on public	.21***
Evaluation of presumed influence of press and TV on journalists	.09**
Evaluation of presumed influence of press and TV on politicians	.01
Presumed influence of press and TV on public × evaluation of influence	.08**
Presumed influence of press and TV on journalists × evaluation of influence	−.08*
Presumed influence of press and TV on politicians × evaluation of influence	−.05*
R² change	.11***
Total R²	.14***

Note. Coding: education = years of education, evaluation of influence (recoded: 1 = very positive to 5 = very negative); all variables of perception and evaluation of influence are centered.

p < .05. **p < .01. ***p < .001.

H2 has to be largely rejected. The multiplication terms of strength and evaluation of influence, in most cases, do not affect the dependent variables. However, the perceived strength of influence has significant effects.⁵⁹ They become evident almost continuously. The stronger the journalists consider the political influence of the Internet on the public, the more they ask for the political influence of the Internet to be restricted, the more relevant they find stricter state control of the Internet, and the more they would support a corresponding initiative. In addition, the stronger the political influence of the press and television on the public is perceived by journalists, the stronger their approval of restricting the influence of such media.

The perceived strength of the influence on politicians leads to similar results. Strong presumed influences of the Internet lead to a higher evaluation of the relevance of control measures and a more pronounced intention to support state control of the Internet. Restrictions on political influence for the Internet, the press, or television are not supported, however.

Significant effects were also detected concerning the perceived influence on other journalists. However, the effects are the opposite of what was anticipated. The stronger the presumed political effects of the Internet, the press, and television, the less the journalists tend to support censorship and to regard censorship as relevant.

When the evaluation of media influences is considered, the following pattern emerges: negatively perceived political media influences on the public cause journalists to demand restrictions on the influence of the Internet, the press, and television; to regard state control of the Internet as relevant; and to be willing to support a corresponding initiative. A negative evaluation of the influences on other journalists results only in a demand for restrictions on the political influence of the Internet, the press, and television. The evaluation of the media influences on politicians did not have an effect on the dependent variables.

In many cases, media influences that are perceived as strong thus affect attitudes or behavioral intentions toward censorship—regardless of how the presumed influences are evaluated. Significant interaction effects of influence strength and its evaluation are only found regarding the support for restricting the political influence of the press and television. Simple slope analyses were conducted to interpret these interactions.⁶⁰ Results show that at least the strength of perceived press and television influence on the public leads, in the case of a negative evaluation, to significantly higher demands for restrictions on influence. In contrast, the strength of the presumed influence does not have a significant effect in the case of a positive evaluation. This corresponds to H2.⁶¹

H3

According to the third assumption, perceived strong and negative media influences on the public enhance the approval of or the intention to support censorship more than supposed influences on politicians and other journalists. As already mentioned, perceived strong or negatively evaluated media influences on politicians and other journalists often do not significantly affect the dependent variables. In the case of other journalists, even negative relations are detected. Tables 2 and 3 show that the standardized regression coefficients of the strength and of the evaluation of influence (provided they are significant and have a positive algebraic sign) are mostly lower for politicians and for journalists than for the public. There is one exception: the supposed influence of the Internet on politicians has a stronger effect on the relevance of control measures than the supposed influence on the public. Apart from this, H3 can be confirmed. From the respondents’ point of view, it is mainly the public that has to be protected from strong or negatively evaluated media influences.

H4

H4 suggested that perceived strong and negative political effects of online media on others exert a stronger influence on support for censorship than perceived effects of classic media. First, journalists are slightly more in favor of restricting the Internet (M = 1.77, SD = 1.09) than they are of restricting the press and television, M = 1.60, SD = 1.04, t(1673) = 7.34, p < .001.

Considering the values in Tables 2 and 3, however, H4 cannot be confirmed. The regression coefficients are not remarkably different, whether the reasons for support for Internet censorship or for support for press/television censorship were examined. Considering the overall results, there is no evidence that support for Internet censorship can be better explained by perceived media influence than support for press/television censorship.

In the context of H4, the type of media the journalists work for was examined. As shown in Table 2, journalists who (also) work for online media consider state control of the Internet less important and would support such an initiative less than their colleagues who work for classic media. However, whether the journalists also work for online media is irrelevant when it comes to demands for restricting the influence of the Internet, the press, and television.⁶²

H5

Finally, it was hypothesized that the intention of journalists to support control measures depends mainly on their attitude toward such measures. In addition, this attitude was thought to mediate the influence of perceived media influences on intentions to support Internet control. Two path analyses were conducted to reassess this hypothesis, for the perception of influence on the public (Figure 1) and on politicians (Figure 2).⁶³ The perceived political influence of the Internet on the respective group and the evaluation of this influence served as independent variables. They were assumed to have a direct effect on the evaluation of the relevance of Internet control. This evaluation, in turn, complies with attitude toward Internet control that, according to the theory of reasoned action, affects intention to support Internet control measures. The path models were saturated; thus, assessing the pertinent characteristics regarding the quality of the model was not possible.

Figure 1.

Path model, attitude toward Internet control as a mediator between the perceived influence of the Internet on the public or the evaluation of this influence and the behavioral intention as the endogenous variable (standardized coefficients).

Figure 2.

Path model, attitude toward Internet control as a mediator between the perceived influence of the Internet on politicians or the evaluation of this influence and the behavioral intention as the endogenous variable (standardized coefficients).

The results shown in Figure 1 demonstrate that the perceived strength of the Internet influence on the public (β = .05, p = .06) and the evaluation of this influence (β = .17, p < .001) (tend to) have positive effects on attitude toward state control of the Internet. However, in contrast, only the influence strength, not the evaluation, directly affects the intention to support Internet control measures (β = .03, p < .05). In turn, the latter is highly dependent on the attitude toward control measures (β = .84, p < .001). As expected, an indirect, attitude-mediated effect of the influence strength (β = .04, p = .05) and the influence evaluation (β = .14, p < .001) on the intended action was detected. The stronger and the more negative journalists perceive the influence of the Internet on the public, the more relevant the journalists find the need to put this influence under state control. This leads to an enhanced intention to support a corresponding initiative.

The same applies to the perceived Internet influences on politicians (Figure 2). Once again, the effects of presumed influence strength and of influence evaluation on behavioral intention are mediated by attitude. Unlike the direct paths, the indirect effects concerning the perceived influence strength (β = .07, p < .01) and the influence evaluation (β = .09, p < .001) are significant.

Discussion

The aim of this study was to test which influences journalists ascribe to offline and online media and whether these perceived influences lead to enhanced approval of censorship. First, the results show that journalists perceive the political influence of the press or television and different online media on other people as significantly stronger than on themselves. Considering the results of previous third-person research,⁶⁴ this finding does not come as a surprise. However, the consistency of this third-person perception across all media types and media offerings is remarkable. Moreover, the results reveal differences regarding the examined groups. The journalists presume that press and television mainly exert a political effect on the public and on politicians, while leaving other journalists comparatively unaffected. On the contrary, as far as the political effect of the Internet is concerned, other journalists are considered much more susceptible than the public and politicians. This probably reflects the presumption that journalists use the Internet much more frequently for informing themselves about political proceedings. This interpretation corresponds to results finding that the perceived reach of a medium or media coverage is the central predictor of the medium’s presumed influence.⁶⁵

The comparison of different media shows that journalists perceive the political influence of the press and television as significantly stronger than the influence of the Internet or specific online offerings. Future studies should test which factors—in addition to the perceived reach—are responsible for the different perceptions. For example, Web 2.0 offerings might be considered less informative compared to press or television and might, therefore, be thought to have minor political influence.

Journalists believe the media affect other people. However, what consequences do these perceived influences have? The results can be summarized as follows: the stronger journalists perceive the political influence of the respective media on the public to be, the more strongly the journalists approve of demands to restrict the political influence of the Internet, the press, and television; and the more relevant the journalists find enhanced state control of the Internet, and the sooner they express an intention to support corresponding initiatives. This is remarkable in three respects. First, the results of previous research are also applicable to journalists, who, as a rule, are exceedingly critical of all attempts at restricting their freedom.

Second, approval of censorship items is linked to political media content. This area occupies a special position from the point of view of democratic theory. The results illustrate that even the sensitive area of political communication is not exempt from demands for stricter regulation and restriction.

Finally, media influences do not necessarily have to be evaluated negatively to increase approval of censorship. The results show that the perceived strength alone suffices for this purpose. Apparently, even journalists show a critical mind-set concerning (too) strong media influences. This might be explained by journalists’ general skepticism toward a plenitude of power that is, from their point of view, generally thought to need control. However, it must be pointed out that these findings are not based on a representative sample of German journalists.

One further consideration regarding these results is crucially important. The journalists did not loudly call for prohibitions, controls, and restrictions and are not positively disposed to such measures. The perception of strong media effects does not necessarily lead—in absolute terms—to approval of censorship. This perception only tends to increase this approval. The descriptive results are also important. The data reveal that the approval of control measures and the demands for influence restriction are generally low. Thus, the majority of journalists oppose interference with the freedom of the media. However (and this is the crucial result of this study), perceived strong media influences on others can at least soften this attitude.

The results clearly demonstrate that journalists want to shield, most of all, the public from strong media influences. In contrast, presumed media effects on politicians exert a rarer and weaker influence on the different variables of censorship. This might be due to the higher media competence the journalists ascribe to politicians rather than to the public. Someone capable of dealing with media content does not have to be protected from its influence—not even if these media actually exert an influence on the individual. This is perhaps why journalists do not believe their fellow journalists need protection. In this particular case, even converse effects are revealed: the more strongly the respondents perceive the media influence on other journalists, the less the respondents support censorship. Strong media influences on journalists are not only regarded as not requiring regulation, but also even as desirable. One possible reason is that journalists also orient themselves to media coverage by others.

In general, the respondents show they are slightly more open-minded toward stricter regulation or influence restriction of the Internet than toward censorship regarding the press and television. This is especially true for journalists who work exclusively in classic media. The “new media” characterized by interactivity and user-generated content might be experienced as a bigger threat to their own work by many journalists who fear losing their roles as professional news messengers. Future studies should undertake a finer differentiation between particular media types to permit more detailed statements regarding the types of media that bring about the highest approval of demands for censorship. In addition, future studies could differentiate more strongly between the journalists questioned. For example, because politicians are members of the supervisory boards of Germany’s public broadcasters and can exert influence, journalists working for public networks might be particularly disapproving of censorship. Consequently, it would be interesting to see whether there is a difference between these journalists and their colleagues who work for private networks. However, this question cannot be answered with the present data.

Furthermore, future research should measure actual behavior. This is comparatively difficult in the context of censorship, as most people cannot carry out regulating or restricting measures themselves. Corresponding initiatives one could actively support are extremely rare, too. Thus, it seems necessary to examine not only attitudes toward censorship, but also behavioral intentions—as has been done here. According to the theory of reasoned action, the latter allow decisively better conclusions on actual behavior than do attitudes. However, attention has to be paid to the fact that the behavioral intention (at least according to a result of the present study) depends less directly on the perception of media influences and more on the attitude toward behavior. Thus, the latter forms an important mediator variable.⁶⁶ Therefore, measuring attitudes and behavioral intentions seems necessary to model the influences of presumed media influence appropriately.

Finally, it might be useful to concretize the dependent variable “support for censorship demands” in future studies. In the present study, the questions were deliberately aimed at “restrictions” in a very general sense, and no specific measures or precise content were mentioned. Yet, the range of “censorship measures” and of content that they could be referring to is very broad. Thus, it cannot be ruled out that the respondents had quite different kinds of interference and content in mind when giving their answers. It would, for example, make sense for them to lean toward approving restrictions on other journalists’ work while strictly refusing interference in their own autonomy. This potential “third-person effect” of a different kind will have to be examined in the future.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the German Research Association (Research Unit “Political Communication in the Online World,” Subproject 3, Grant 1381).