Professional Documents
Culture Documents
Mur 078
Mur 078
ABSTRACT
doi:10.1093/jopart/mur078
doi:10.1093/jopart/mur078
Advance Access publication on February 29, 2012
ª
© The
The Author
Author 2012.
2012. Published
Published by
by Oxford
Oxford University
University Press
Press on
on behalf
behalfof
ofthe
theJournal
JournalofofPublic
PublicAdministration
AdministrationResearch
Research
and Theory,
and Theory, Inc.
Inc. All
All rights
rights reserved.
reserved. For
For permissions,
permissions, please
please e-mail:
e-mail:journals.permissions@oup.com
journals.permissions@oup.com
598
2 Journal of Public Administration Research and
and Theory
Theory
The effectiveness of communication may be questioned when the recipients of those mes-
sages do not trust the senders. Citizen trust in government plummeted to 17% in 1994 and
again in 2008, the lowest since these statistics were first kept in 1958. The average of those
who trusted government most of the time was 35% with a high of 77% in 1964 (Pew Research
Center for the People and the Press 2010). While dealing with public distrust, government
communicators also face more limitations on their ability to communicate due to laws and
regulations than their counterparts in the private sector (Liu, Horsley, and Levenshus 2010).
Given the importance and challenges, it is crucial for public administration scholars to
more rigorously study government communication and its impact on media coverage and,
Public cynicism has persisted even though research finds that most government messages
are communicated in an open and honest manner (Graber 2003). Third, proactive interac-
tion with reporters can help build strong relationships between communicators and report-
ers, which increases the chance of fair coverage (Lattimore et al. 2009). Reporters often
rely on government communicators for information, and an interdependent relationship
may help produce favorable reporting (Elsbach 1994; Gellert et al. 1994). Therefore, we
propose:
H1 External communication activities are associated with more favorable media
coverage.
(Lee 2008; Liu and Horsley 2007). This devaluation in government not only reduces non-
communicators’ desires to support communication activities but also places communicators
in relatively low organizational positions that cannot attract colleague support. A recent na-
tional survey found that only 36% of government communicators have management titles
(National Association of Government Communicators 2008). Lovell (2001) found that
police departments that encourage communication with the media from all levels of the
organization tend to have better media image. Therefore, we propose:
H3 Cross-department support for external communication is positively associated with
favorable media coverage.
some controversial ‘‘liberal programs’’ favorable coverage despite being highly political.
In comparison, politics not only hinders the effectiveness of business communicators’ jobs
but also is likely to be viewed as less prominent or healthy for business organizations. Thus,
we pose the following research question given that the extant literature is not able to support
hypothesis generation:
RQ1 To what extent do politics affect government and business external
communication?
Another constraint faced by government communicators is the amount of information
METHODS
This study is based on data collected from a mail and online survey of US government and
business communicators between April and June 2008.
Survey Sample
The survey sample was 2,525 US government communicators and business communicators.
Government communicators were defined as current government employees or contractors at
the local, state, or federal level, or for quasi-governmental agencies such as public utilities
partnerships, whose primary responsibilities are communicating internally or externally to
various publics regarding agency/department/office policies, decisions, or actions, and/or
guiding communication strategy. Business communicators were defined as current employ-
ees or contractors of companies and corporations whose primary responsibilities
604
8 Journal of Public Administration Research and
and Theory
Theory
Survey Administration
Prior to disseminating the survey, we pretested the questionnaire with 6 government com-
municators and 10 business public relations practitioners. For the dissemination, we fol-
lowed a modified version of Dillman’s (2000) tailored design method: (a) we e-mailed
1 To further ensure that the use of a latent factor did not bias our hypothesis testing based on ordinary least squares
(OLS) regression later on (i.e., table 6), we ran separate logistic regression models with each of the dimensions as the
dependent variable. The results were highly consistent across the dimensions, and they are consistent with what is
reported in table 6.
al. Overcoming Negative Media Coverage
Liu et al. 605
9
a brief prenotice letter, (b) we e-mailed a detailed cover letter with a link to the online
survey, (c) we e-mailed a thank-you/reminder letter, (d) we mailed via US Postal Service
a thank-you/reminder letter with a replacement questionnaire to nonresponders, and (e) we
called and/or e-mailed a random sample of nonresponders.
The survey response rate was 41% with 1,043 communicators responding to the sur-
vey. Fifty-five respondents identified themselves as currently working in ‘‘Other’’ types of
organizations such as nonprofits, thus the usable response rate was 39%. The response rate
for the government communicators (40%) and business public relations practitioners (37%)
was roughly equivalent. Approximately two-thirds of the respondents completed the survey
Respondents’ Demographics
The respondents’ demographics are reported in table 1 by employer type (business or gov-
ernment). For our purpose here, we dropped respondents who identified themselves as
working for government but either as a consultant or in a quasi-governmental agency, thus
reducing the sample size for government responses to 571. Similarly, for business respond-
ents we dropped those who worked for a PR or advertising agency or as a consultant, re-
ducing the sample size for business response to 307. The overall relevant sample size is 881.
Among the government respondents, 55% identified themselves as part of the top
management team, 33% were male,3 the average number of employer-paid training semi-
nars attended last year was 2.14, and the average number of subordinates was 3.42, all of
which were not statistically significant from the business respondents. However, govern-
ment respondents on average had lower salaries than did business respondents (p , .001).
On average, government respondents reported salaries in the range of $60,000–$80,000
2 The t-tests further show that online respondents and traditional respondents did not differ on the variables including
multimedia communication, influence of politics, and information need. Online respondents reported statistically
higher scores on engaging public information activities than traditional respondents, but they reported statistically
lower scores on media interaction, communication planning and research, and cross-department support for
communication. In general, even when there was a statistical difference between the two types of respondents, the
substantive difference was very small. In using regression models to test the hypotheses, we included a dummy variable
with online completion being coded as 1. Consistent with the t-test results, this dummy variable was statistically
significant in the regression models and was negatively associated with positive news coverage. However, adding or
excluding the dummy variable neither change the significance and direction of all other variables nor have a significant
impact on the F values and R squares of the models. In other words, our final results reported in table 6 would remain
essentially the same if the dummy variable were included. Given that this dummy variable is not essential to our
hypotheses and that the difference between online and traditional surveys is an important topic in and of itself, we
concluded that a separate methodological article should be produced to specifically address the issue.
3 The gender breakdown of our respondents is consistent with previous surveys of government communicators and
public relations professionals (e.g., National Association of Government Communicator 2008; Toth and Aldoory
2000).
606
10 Journal
JournalofofPublic
PublicAdministration
AdministrationResearch
Researchand
andTheory
Theory
Table 1
Characteristics of the Average Respondents (Government versus Business)
Government (n 5 571) Business (n 5 307)
Salary level 4.33*** (1.28) 5.33 (1.21)
Tenure in the profession (years) 10.26* (8.37) 12.08 (9.74)
Number of subordinates 3.42 (5.77) 9.16 (37.57)
Number of communication associations 2.14 (2.11) 2.01 (1.25)
participated
No. of training seminars attended last year 2.09 (1.97) 2.20 (2.07)
paid for by the employer
4 Missing values are not a problem as fewer than 3% of observations had missing values. As an empirical check, all
the statistical tests were rerun with multiple imputation procedures (SAS MI and MIANALYZE), and the results were
essentially the same as what is reported here.
OvercomingNegative
Liuetetal.al. Overcoming
Liu NegativeMedia
MediaCoverage
Coverage 607
11
Table 2
Factor Analysis Results for Communication Activities
Factor 1 Factor 2 Factor 3 Factor 4
Write news releases and advisories 76* 14 7 13
Hold news conferences 41* 25 18 40
Conduct media interviews 79* 27 24 11
Respond to media inquiries 89* 4 5 5
Pitch stories to the media 72* 7 14 22
Track media clips 55* 2 16 1
Create brochures 0 90* 11 1
Table 3
Descriptive Statistics and t-Tests between Government and Business Respondents
State and
Scale Federal Local
Range Midpoint Government Business Government Government
Positive news coverage 3–15 9 10.51*** (2.34) 11.40 (2.08) 10.32 (2.54) 10.42 (2.27)
(N 5 867)
Media interaction 6–30 18 20.24** (5.68) 21.29 (5.24) 17.09*** (7.34) 21.15 (4.73)
(N 5 869)
Public information 4–20 12 12.26** (4.04) 10.84 (4.72) 10.23*** (3.70) 12.86 (3.94)
communication planning and research (p , .001), but engage less in public information
(p , .01) and conduct fewer multimedia communication activities (p , .01). Among the
constraint variables, government organizations have less adequate communication budgets
(p , .01), are more influenced by politics (p , .001), and face more extensive information
need from outside the organization (p , .01). However, there is no difference on cross-
department support for communication. The reason might be that public sector manage-
ment has increasingly understood the value of communication and public relations. Overall,
government organizations face a tougher communication environment (more politics and
needs) but with weaker budgetary support.
Table 3 also demonstrates that there is no statistically significant difference between
federal and state/local governments regarding communication planning and research, in-
fluence of politics, communication budget adequacy, and cross-department support for
communication. However, federal agencies face less need for information from outside
the organization (p , .001), and they conduct fewer activities in media interaction
(p , .001), public information (p , .001), and multimedia communication (p , .001).
The comparison makes sense as state and local governments more frequently interact with
citizens and clients than do federal agencies. Budget and politics are constraints regardless
of the level of government.
OvercomingNegative
Liuetetal.al. Overcoming
Liu NegativeMedia
MediaCoverage
Coverage 609
13
Table 4
Response Distribution on Perceived Favorable Media Coverage
Percentage
Value Government (n 5 568) Business (n 5 299)
3 0.2% 0%
4 0% 0%
5 0.7% 0.7%
6 4.4% 1.0%
7 4.6% 2.0%
8 5.5% 4.3%
Model Testing
A correlation matrix is displayed in table 5. Most variables are weakly or moderately cor-
related except a relatively strong correlation between media interaction and information
need (r 5 .65, p , .001), suggesting that multicollinearity may not be a problem
(later confirmed with regression diagnostics such as low scores of variance inflation factor).
Among the four types of communication activities, only media interaction is significantly
correlated with positive news coverage (r 5 .19, p , .001). The correlation coefficients
among the four types of activities are either not statistically significant or not strong,
indicating they are indeed distinct types.
To assess the impact of communication activities and constraint variables on media
coverage, three models are tested with OLS regression: One with the whole sample, one
with the government subsample, and the other with the business subsample.5 This is done to
see whether government and business organizations have different dynamics. As table 6
shows, all three models are statistically significant with acceptable F values (13.22, 8.42,
and 4.52, respectively, all significant at the .001 level). The R-square values are not very
large (.12, .12, and .11, respectively) because the survey did not include other variables that
might be relevant given that our focus is on the major explanatory variables discussed by
Liu and Horsley (2007) and Liu and Levenshus (2010).
Overall, the hypotheses are largely supported. Consistent with Hypothesis 1, the
results show that some of the communication activities do lead to positive media coverage:
media interaction for government (Model 2, b 5 .29, p , .001) and public information for
business (Model 3, b 5 .23, p , .001). But multimedia communication is statistically
5 The normality assumption was slightly violated for the residuals, as shown by the Kolmogorov-Smirnov test and the
Shapiro-Wilk test, which are unfortunately sensitive to sample sizes. The normality plot of the residuals showed very
slight violation. Whereas OLS regression is fairly robust for slight violation of the normality assumption (Hair et al.
1998; Tabachnick and Fidell 2007), we tried nonparametric tests (logistic regression) to cross-validate the results, and
the results were similar (same significant variables and same directions of the coefficients). The logistic results can be
obtained from the authors. We also tried regression with robust standard errors and the results were similar too.
Therefore, we report the regular OLS results here.
610
Table 5
Correlation Matrix
Variables V1 V2 V3 V4 V5 V6 V7 V8 V9
V1 positive news 1.00
coverage
V2 media interaction 0.19*** 1.00
V3 public information 20.00 0.04 1.00
V4 communication 0.04 0.37*** 0.27*** 1.00
planning and
research
V5 multimedia 0.03 0.27*** 0.30*** 0.27*** 1.00
advertising
V6 influence of politics 20.11*** 0.06 0.13*** 0.17*** 0.14*** 1.00
Journal of Public Administration Research and Theory
Table 6
Regression Results Predicting Positive News Coverage
Model 1 Model 2 Model 3
Whole Sample, Government, Business,
b (SE) b (SE) b (SE)
Media interaction 0.17 (0.02)*** 0.29 (0.02)*** 20.03 (0.03)
Public information 0.06 (0.02)* 20.01 (0.02) 0.23 (0.03)***
Communication planning 20.12 (0.02)** 20.12 (0.03)** 20.13 (0.04)*
and research
Multimedia communication 20.00 (0.03) 20.01 (0.04) 0.01 (0.05)
significant in none of the models, and communication planning and research negatively
impacts media coverage for both government (Model 2, b 5 2.12, p , .01) and business
(Model 3, b 5 2.13, p , .05). Among the constraint variables, budget inadequacy matters
for both sectors. As predicted in Hypothesis 2, budget adequacy is positively significant for
both government (Model 2, b 5 .19, p , .001) and business (Model 3, b 5 .13, p , .05).
Hypothesis 3 is partially supported: cross-department support for communication is pos-
itively related to affirmative media coverage for government (Model 2, b 5 .10, p , .05)
but is not significant for business.
Of the variables for which we did not specify formal hypotheses, influence of politics
is not statistically significant in the government model but significant in the business model
(Model 3, b 5 2.20, p , .001). Similarly, information demand is not statistically signif-
icant in the government model, but significant in the business model (Model 3, b 5 .23,
p , .01). Finally, government is less likely to have positive media coverage than business
organizations (Model 1, b 5 2.19, p , .001).
empirical research. Consequently, this study’s findings offer some important and intrigu-
ing insights and point to some future research directions. In this next section, we highlight
the primary findings and offer tentative interpretations, based on our results and previous
research. Future research, such as interviews and focus groups, is needed to further val-
idate our interpretations.
Communication Activities
The results offer an empirically validated conceptual framework of government commu-
news coverage for both government and business samples. One explanation is that when
scarce time and resources are spent in planning and research, activities that will directly
influence the tone of news coverage, such as media interaction and public information,
are undercut. This is plausible because organizations on average do not have plentiful budget
and staff for the communication function, but the correlation matrix does not lend much
support as planning and research is positively correlated with other communication activities.
A more likely explanation is the negative association suggests a causal relationship
that is reversed: When organizations face more negative news coverage, they are more
likely to focus on planning and research. This leaves an important question for future stud-
Constraint Factors
Although it is common perception that government communication faces more constraints
than business communication, our results suggest that only some constraints influence per-
ceived government media coverage. Our t-test results confirm that government has higher
degrees of influence of politics, external information need, and budget inadequacy than
business organizations, but the multivariate test results show that only budget inadequacy
and lack of cross-department support are associated with reporting more negative media
coverage for government. Influence of politics and information need do not affect the re-
ported tone of media coverage for government, although they do influence the reported
media coverage for business in interesting ways.
614
18 Journal
JournalofofPublic
PublicAdministration
AdministrationResearch
Researchand
andTheory
Theory
A clear observation from the study’s results is you reap what you invest in: commu-
nication budget adequacy is positively and significantly associated with reported positive
news coverage for both government and business organizations. This is not surprising given
that resource adequacy is an important determinant for any performance model (Pfeffer and
Salancik 1978; Rainey and Steinbauer 1999). The finding, however, is particularly alarm-
ing for government as government tends to have lower levels of communication budget
adequacy than business organizations (see the t-test results). This is in line with the finding
by Liu and Levenshus (2010) from interview data that the lack of resources is the biggest
challenge faced by government communicators. This concern becomes more severe in
need and reported media coverage for government. Instead, information need is associated
with reported positive media coverage for business organizations. It is likely that more
information requests from the media and the public may not necessarily be good for
government agencies that are more conservative in external communication and want
to protect their information territory. For business organizations, however, more informa-
tion requests tend to be a positive opportunity as their proactive managers strive for more
brand recognition and news coverage. It is worth noting that this study measures only the
level of information need, but future studies may also investigate the type of information
requested and the type of stakeholders who make the request.
These findings have great practical implications. Public managers need to recognize
that the nature of government work—more political influence and external information
demand—does not necessarily lead to negative news coverage. Although management
strategies may not be able to fully address the problem of negative news coverage given
factors outside managers’ control, more proactive media interaction does help improve
media coverage. If government agencies want their communication staffs to help generate
positive news coverage, they need to value the role of communication, find a way to in-
crease agency-wide support for communication, and allocate more for communication
budgets. This is tough to do in our current economic situation, but proving the performance
FUNDING
This work was supported by a grant from the Plank Center for Leadership in Public Rela-
tions at the University of Alabama.
APPENDIX
Measurement and Internal Consistency for Index Variables
Communication Activities: How often do you and/or a member of your team engage in the
following communication activities? 1 5 never engage in activity; 5 5 very frequently
engage in activity (i.e., daily)
Media Interaction (standardized alpha 5 .87)
Write news releases and advisories
Hold news conferences
Conduct media interviews
Respond to media inquiries
Pitch stories to the media
Track media clips
Public Information (standardized alpha 5 .86)
Create brochures
Create fliers
Create guides
Create print advertising
Communication Planning and Research (standardized alpha 5 .65)
Develop strategic plans
OvercomingNegative
Liuetetal.al. Overcoming
Liu NegativeMedia
MediaCoverage
Coverage 617
21
Gender
Male 5 1 Female 5 0
Tenure in the Profession
How many years, if any, have you been employed in government (or corporate)
communication?
Number of Training Seminars Attended Last Year Paid for by the Employer
In the past year, how many training seminars, if any, did you attend that were paid for
by your employer?
REFERENCES
Appleby, Paul. 1973. Big democracy. New York, NY: Knopf.
Barnard, Chester. 1938. The functions of the executive. Cambridge, MA: Harvard Univ. Press.
Berman, Evan M. 1997. Dealing with cynical citizens. Public Administration Review 57:105–12.
Blumer, Tom. 2010. Good news: Newspaper circ shrinks just 8.7%. http://newsbusters.org/blogs/
tom-blumer/2010/04/26/good-news-newspaper-circ-shrinks-just-8-7-fair-balanced-wsj-only-top-25-/
(accessed January 5, 2012).
Bohte, John, and Kenneth Meier. 2000. Goal displacement: Assessing the motivation for organizational
cheating. Public Administration Review 60:173–82.
Bok, Derek. 2001. The trouble with government. Cambridge, MA: Harvard Univ. Press.
Bollen, Kenneth A., and Richard Lennox. 1991. Conventional wisdom on measurement: A structural
equation perspective. Psychological Bulletin 110:305–14.
Bozeman, Barry. 1987. All organizations are public. San Francisco, CA: Jossey-Bass.
Cappella, Joseph N. 2002. Cynicism and social trust in the new media environment. Journal of Com-
munication 52:229–41.
Cappella, Joseph N., and Kathleen Jamieson. 1997. Spiral of cynicism: The press and the public good. New
York, NY: Oxford Univ. Press.
Christensen, Tom, and Peer Laegreid. 2005. Trust in government: The relative importance of service
satisfaction, political factors, and demography. Public Performance and Management Review
28:487–511.
Coglianese, Cary, and Margaret Howard. 1998. Getting the message out: Regulatory policy and the press.
The Harvard International Journal of Press/Politics 3:39–55.
Cohen, Jacob. 1988. Statistical power analysis for the behavioral sciences. Hillsdale, NJ: Lawrence
Erlbaum Associates.
Comfort, Louise. 2007. Crisis management in hindsight: Cognition, communication, coordination, and
control. Public Administration Review 67:189–97.
Coombs, W. Timothy. 2008. Parallel process model and government preparedness messages: Beyond duct
tape and plastic sheeting. In Crisis communication and the public health, ed. Matthew W. Seeger,
Timothy Sellnow, and Robert R. Ulmer, 221–34. Creskill, NJ: Hampton Press.
Dillman, Don A. 2000. Mail and Internet surveys: The tailored design method. New York, NY: John Wiley
& Sons.
Downs, Anthony. 1972. Up and down with ecology: The issue-attention cycle. National Affairs 28:38–50.
OvercomingNegative
Liuetetal.al. Overcoming
Liu NegativeMedia
MediaCoverage
Coverage 619
23
Elsbach, Kimberly. 1994. Managing organizational legitimacy in the California cattle industry.
Administrative Science Quarterly 39:57–88.
Fairbanks, Jenille, Kenneth D. Plowman, and Brad L. Rawlins. 2007. Transparency in government
communication. Journal of Public Affairs 7:23–37.
Fallows, James. 1996. Breaking the news: How the media undermine American democracy. New York,
NY: Vintage.
Fitch, Brad. 2004. Media relations handbook for agencies, associations, nonprofits, and Congress.
Alexandria, VA: TheCapitol.Net.
Fornell, Claes, and Fred Bookstein. 1982. A comparative analysis of two structural equation models:
LISREL and PLS applied to market data. In A second generation of multivariate analysis,
ed. Claes Fornell, 289–324. New York, NY: Praeger.
Jones, Jeffrey M. 2004. Americans’ trust in the mass media. Gallup. http://www.gallup.com/poll/11428/
americans-trust-mass-media.aspx/ (accessed January 5, 2012).
Jones, Nicholas. 2002. 24-Hours media. Journal of Public Affairs 3:27–31.
Kalantari, Bethrooz. 2008. Media and the bureaucracy in the United States. In Government public
relations: A reader, ed. Mordecai Lee, 107–14. Boca Raton, FL: Taylor & Francis.
Kirlin, John. 2001. Big questions for a significant public administration. Public Administration Review
61:140–3.
Kozolanka, Kirsten. 2006. The sponsorship scandal as communication: The rise of politicized and strategic
communications in the federal government. Canadian Journal of Communication 31:343–66.
Lattimore, Dan, Otis Baskin, Suzette T. Heiman, and Elizabeth Toth. 2009. Public relations: The pro-
fession and practice, 3rd ed. Boston, MA: McGraw Hill.
Public Relations Society of America (PRSA). 2010. Global research study finds that public relations is
consistently taught as a strategic management function. Public Relations Society of America. http://
media.prsa.org/article_display.cfm?article_id51653/ (accessed January 5, 2012).
Putnam, Robert. 1993. Making democracy work. Princeton, NJ: Princeton Univ. Press.
Rainey, Hal. 2003. Understanding and managing public organizations, 3rd ed. San Francisco, CA: Jossey
Bass.
Rainey, Hal, and Paula Steinbauer. 1999. Galloping elephants: Developing elements of a theory of ef-
fective government organizations. Journal of Public Administration Research and Theory 9:1–32.
Rawlins, Brad L. 2007. Trust and PR practice. Institute for Public Relations. http://www.instituteforpr.
org/essential_knowledge/detail/trust_and_pr_practice/ (accessed January 5, 2012).
Reich, Zvi. 2009. Ethical weaknesses in emergency communication between national authorities and the