37

Keio Communication Review No. 26, 2004

Factors Influencing Tabloid News Diffusion:
Comparison with Hard News
by Yasuhiro INOUE
Yoshiro KAWAKAMI*
Statement of Purpose
The tabloidization of journalism has blurred a distinct line between hard
news and tabloid news. Television news magazines have been transforming
into "Infotainment" programs and shifting their focus from political and social
issues to scandalous and personal stories such as celebrities' love affairs and
bizarre crimes (Bird, 1998). Even mainstream news organizations are buying
into the trend (Alter, 1995). These two news categories are merging together;
however, attributes of hard news and tabloid news may still matter to the news
audience. Tabloid news may have different attributes to activate particular de-
mographic groups' networks from those of hard news. People may perceive and
react differently to hard news and tabloid news, respectively. Thus, the dis-
semination process of tabloid news may differ from that of hard news. Little,
however, is understood about the diffusion of tabloid news and people's reac-
tion to the news category.
This study attempts to fill these gaps in the literature by examining tabloid
news diffusion and comparing the results of tabloid news diffusion with those of
hard news. The purpose of this study is to investigate whether or not there is a
distinct line between hard news and tabloid news in terms of news diffusion
process by testing the applicability of news diffusion theory to tabloid news.
In this study, two tabloid news events and two hard news events in Japan
are examined by surveying college students. These tabloid stories are the en-
gagement of a popular sumo wrestler and actress, and the "surprise" second
marriage of a popular singer. All these tabloid news stories were sensational
and attracted national attention in Japan. These hard events are the selection of
a socialist Prime Minister and the Great Hanshin earthquake.
Research about how news events diffuse began in the 1940s. However,
these studies focused exclusively on "hard news" such as assassinations and
catastrophes. No literature clearly distinguished between hard news and tabloid
news in terms of diffusion patterns, interpersonal roles, and gender roles. News
* INOUE Yasuhiro is Associate Professor at Hiroshima City University, and KAWAKAMI
Yoshiro is Professor at Seijo University. Correspondence regarding this article should be
directed to INOUE Yasuhiro; email: [email protected]
38
diffusion studies generally found that interpersonal communication plays a role
in diffusing news events. Even though fast and far reaching new media revolu-
tionized our information environment, word of mouth still remains as a means
of a primary news source for important events (for a comprehensive review, see
DeFleur, 1987).
1
This interpersonal role in diffusion process ought to be inves-
tigated with tabloid news events because it is necessary to examine the diffusion
of non-hard news to understand the comprehensive process of news diffusion
(McQuail, 1994). Therefore, examining tabloid news diffusion will provide more
insight and empirical evidence into news diffusion research that has been biased
toward hard news diffusion patterns.
Literature Review
Tabloid News
The definition of tabloid news is controversial (Grabe, 1997). In this study,
tabloid news is defined as news that is (1) celebrities' matter, and (2) primarily
covered by tabloid newspapers and magazines and "infotainment" television
programs, so-called "Waido show" in Japan. This news category is often called
"geino news" (celebrity news) in Japan and "Hollywood gossip" in the United
States. Celebrities' love affairs and engagements/marriages/divorces fall in this
news category. This does not necessarily mean that news stories covered by
"hard" newspapers and television news programs are excluded. While tabloid
news has been criticized by established journalists and critics as lowbrow and
degrading journalism, it has established a niche and has become a bona fide
commodity even in respected daily newspapers and news organizations such as
The New York Times and ABC News (Case, 1992; Alter, 1995). Thus, the bound-
ary between hard news and tabloid news is blurred, and sometimes, a news piece
belongs to the hard and the tabloid news categories simultaneously (Tharp &
Streisand, 1994). For example, news about the U.S. President Clinton's affair
with Monica Lewinsky is a political affair as well as a "soap opera" extramarital
love affair. Or, some assert that Clinton's affair deserves only tabloid coverage,
and vice versa.
From the perspective of news diffusion, there might not be any difference
between soft and hard news, and the only difference may be news value and its
personal salience. People who think a news story is important should be more
likely to tell others about it regardless of the category (hard or tabloid). In
addition, the distinction between tabloid and hard news is oftentimes dependent
upon an individual's interpretation. As the boundary between hard news pro-
grams and "infotainment" news programs is becoming blurred, the distinct line
that divides hard news from tabloid news seems non-existing.
39
Keio Communication Review No. 26, 2004
News Diffusion
The history of systematic research about news diffusion can be traced back
to more than a half century. One of the pioneering studies was Miller's study on
the news story about the U.S. President Roosevelt's death (1945). The research
findings suggest that the word of mouth communication increases the spread of
the news very quickly, though the initial source of information was a mass me-
dium, radio. In the 1960s, more scholars gave attention to news diffusion, and
the number of the studies increased. Greater attention could be attributed to The
United States' President John F. Kennedy's assassination. This tragedy eventu-
ally provided a window of opportunity for news diffusion scholars to examine
the various aspects of the spread of the news (for example, Greenberg, 1964;
Hill & Bonjean, 1964). Hill and Bonjean found that a majority of people (57 %)
learned the news by interpersonal communication even though electronic media
had become ubiquitous at that time. The study confirmed that the significance
of the news event activated the interpersonal channel as the initial news source.
Levy (1969) studied people's knowledge about and their information source
for six assassinations that occurred in the United States in the 1960s: President
Kennedy, Senator Robert F. Kennedy, Dr. Martin Luther King, Jr., Medgar Evers,
Malcolm X and George Lincoln Rockwell. A national probability sample of
1,200 people were asked whether or not they had heard of these events and what
was their first news source. The entire sample had heard about the Kennedy and
King assassinations, although there were significant racial (white and non-white)
differences whether they had known about the other assassinations. For ex-
ample, while 82 percent of non-whites and 58 percent of whites heard about the
Evers assassination, for the Malcolm X, the percentage were 81 percent and 69
percent, respectively. In terms of news source, the majority of respondents
learned about the events through television. There was an inverse relationship
between educational levels and the percentage of respondents whose first source
was television, however. People with higher educational levels were more likely
to learn the news by radio than those of lower educational levels.
Bantz, Petronio, and Rarick (1983) examined the influence of demograph-
ics and political affiliation on post communication behaviors by using the at-
tempted assassination of President Ronald Reagan. As is the case of crisis news,
the majority of respondents (68 %) reported interpersonal communication as the
first news source. The authors cross-tabulated demographic variables such as
gender, race, age, political affiliation and occupation against first information
source and post-communication behaviors such as informing others. On one
hand, there were no differences between demographic characteristics and the
first news source. On the other hand, some demographic variables were related
with post behaviors: younger people were more likely to tell someone else
than older people; men and younger people used television for a follow-up
40
information source more than women and older people, respectively. Emotional
responses on communication behavior after learning of the explosion of the
space shuttle Challenger were analyzed by Riffe and Stovall (1989). This study
found that those who were very upset were more likely and faster to diffuse
the news interpersonally.
News Diffusion Propositions
The findings of these studies produced the following general propositions
about the news diffusion process (Gantz & Tokinoya, 1987: 198):
1. The rate and extent of diffusion are directly related to the importance/
salience of the event; diffusion is more rapid and widespread as impor-
tance/salience increases.
2. The role of mediated and interpersonal channels in the process is func-
tionally related to the importance/salience of the event; only with very
important/salient events will interpersonal channels be highly active in
the initial dissemination or follow-up processes.
3. Source of first awareness is functionally related to the availability and
accessibility of mediated and interpersonal channels; interpersonal dis-
semination will be greatest when the media are unavailable.
4. Source of first awareness is functionally related to normal usage pat-
terns associated with the mediated channels; aside from exceptional
events, most people will become aware of unanticipated events from the
medium they normally rely upon for news.
Research Questions
These propositions were drawn almost exclusively from "hard news" diffu-
sion research and thus should hold within the "hard news" category (Gantz &
Tokinoya, 1987). However, the process of diffusion is expected to vary depen-
dent upon the characteristics of news because individuals' reactions may differ
according to the attributes of a news story, namely, hard or soft. Therefore, it is
important to investigate this phenomenon by incorporating tabloid news for un-
derstanding the comprehensive process of news diffusion.
This study thus examines whether the diffusion process (or pattern) of tab-
loid news is different from hard news diffusion, or what both news categories
have in common in terms of diffusion pattern. It also explores what factors
affect the velocity of news diffusion. For example, daily media use, especially
television viewing, is assumed to increase the chance to hear about a news story.
News attributes such as the degree of surprise may increase or decrease the like-
41
Keio Communication Review No. 26, 2004
lihood that people inform others of the news. For example, it is said that the
more surprising a news story is the more likely people are to pass it along. Re-
search questions are:
R1: What factors affect the speed of news receiving, and are there any dif-
ferences between hard news and tabloid news?
R2: Do the news attributes such as impression, surprise, and reliability of
news affect post communication behaviors, and are there any differ-
ence between hard and tabloid news?
R3: Is news diffusion theory applicable to tabloid news?
Methods
Data were gathered about two "tabloid news" events and two hard news
events in Japan from 1992 to 1998. Within a few days after these news events,
one of the present investigator asked students in his regular social science class
sessions to fill out questionnaires.
Since respondents in this study were all college students, one should be
careful to generalize the results. In addition, the respondents' universities are
located in the Tokyo metropolitan area. The results might be different if re-
spondents from other colleges in rural areas had been used because information
environment and daily routines of urban areas should differ from those of rural
areas. External validity is in question. This study, however, is not intended to
make point estimates of population parameters, but to provide empirical evi-
dence to fill the gaps in the literature. Since respondents for all the four news
events were similar in terms of living and information environment, confound-
ing factors have been reduced. Thus, the differences in dependent variables
could be at least partly attributed to independent variables. In addition, rela-
tively immediate questionnaire administration is thought to have contributed
to the accuracy of respondents' recall. The events and the brief backgrounds
are as follows:
Tabloid News Stories
The Engagement of Star Sumo Wrestler and Super Actress
Sumo wrestling is totally different from pro-wrestling in the U.S. Sumo is
a national sports game and one of the most popular professional sports in Japan.
Sumo wrestlers are highly respected as strong and disciplined in the nation. In
the evening of October 26, 1992, a television news program scooped the en-
gagement of the number one star wrestler, Takanohana, and one of the most
42
popular actresses, Rie Miyazawa. Their popularity at that time was equivalent
to NBA superstar Michael Jordan and Hollywood actress Julia Roberts. This
news was sensational because both are super stars. After the scoop, all the other
television networks' late night news programs followed the news. Not only tab-
loid newspapers but also all the "Big 3" national elite newspapers covered the
story in the front page.
Data were collected three and four days after the news revelation from a
convenience sample of two classes at a private university in the Tokyo Metro-
politan area. Two hundred and twenty-seven undergraduate students filled out a
questionnaire. One hundred and seventy-seven (78 %) were men and fifty (22
%) were women.
"Surprise" Second Marriage
Seiko Matsuda, a scandalous singer, faxed a marriage announcement to the
major media companies including network television stations in the morning on
May 25, 1998. The announcement said that she was going to get married in the
evening of the same day. She is as scandalous and well known as "Madonna" is
in the U.S. This news story was as sensational as and reported as intensively as
the above engagement stories.
Data were collected four days after the announcement from a convenience
sample at a private university in Tokyo, which is different from the above listed
university. Two hundred and eighteen undergraduate students filled out ques-
tionnaire. One hundred and thirty-five (62 %) were men and eight-three (38 %)
were women.
Hard News Stories
Selection of Socialist Prime Minister
The Japan Socialist Party's chairperson, Tomiichi Murayama, was selected
as the Prime Minster at 9:58PM, June B9, 1994. He was the first socialist Prime
Minister in 40 years.
Data were collected from four universities located in the Tokyo metropoli-
tan area between July 1 and 5. Four hundred forty-one undergraduates filled the
questionnaire. Two hundred seventy-five (63%) were men and one hundred
sixty-five (38%) were women.
The Great Hanshin Earthquake
A great earthquake jolted Japan's Hanshin area (Osaka and Kobe) at 5:46AM,
January 1995. It claimed more than 5,000 lives and devastated the area.
43
Keio Communication Review No. 26, 2004
Three days later, two hundred forty undergraduates of a private university
were asked to fill questionnaire. One hundred eighty-one (79%) were men and
fifty-one (21%) were women.
Questionnaire and Variables
Each questionnaire was nearly identical. The questionnaire asked respon-
dents what time (Time: time lapse between the time a news event was first re-
ported and the time a respondent received the news
2
) and how they first heard of
these news stories (News Source), as well as whether they passed on the news to
others (Inform). The News Source item was dichotomized into two categories
(binominal): whether they first heard about the news stories through mass me-
dia or interpersonal communication channels. They were also asked whether
the news stories became the topic of conversation among their friends (Topic).
Media use items asked about their usual media exposure on Likert scales
(newspaper reading and television news program watching = 4-point, television
watching = 5-point). The scales were coded in inverse relation to the frequency
of media exposure for the purposes of an easier interpretation in multiple re-
gression analysis. Numerical values of the scale become bigger as the frequency
of media exposure becomes smaller.
3
Each media use item serves as an inde-
pendent variable for Research Question 1. In addition, impression items as-
sessed respondents opinions about a news story such as the degrees of (1) Good/
Bad (For all the tabloid news stories and "Prime Minister," Did you feel favor-
able when you heard about the news?; For "Earthquake"; Did you feel bad...?),
(2) Surprise (Were you surprised at the news?), and (3) Reliability (Did you
think the news story reliable?). Each impression item was answered on Likert
scales ranging from 1 to 4 (1 = no, 2 = not very, 3 = relatively, yes, and 4 = very
much). All these impression items are used as independent variables to answer
Research Question 2.
Demographic information other than Gender such as Level (freshman, sopho-
more, junior, or senior) and Home (whether they live with parents or not) were
also collected. Because life styles of freshmen and seniors are thought to be
different, Level is used in the analysis of Research Question 1. Variable Home
is also included in the analysis as an independent variable because the informa-
tion environment of students living with their parents is thought to be signifi-
cantly different from that of students living alone.
Results
Multiple regression was used for Research Question 1 to measure the rela-
tive correlations between the frequency of media exposure, demographic fac-
44
tors, and the speed of receiving news (Time). Three demographic variables were
used: Gender (male = 0, female = 1); Level (freshman = 1, sophomore = 2,
junior = 3, and senior = 4); and Home (whether they live with parents or not: no
= 0, yes = 1). Since a smaller unit in Time means faster reception of news, a
negative sign in Gender and Home coefficients means faster news reception for
female respondents and respondents living with parents, respectively. These
variables were entered first in the equation. In order to measure the relative
correlation, media exposure variables were then entered second. Unlike Gender
and Home, a positive sign in media exposure variables indicates faster news
reception. Table 1 shows the results of the multiple regression test.
Table 1: Multiple Regression Results Indicating Media Use on Diffusion Speed
ENGAGEMENT MARRIAGE
R
2
Final R
2
Final
Step R R
2
Change R R
2
Change
DEMOGRAPHICS 1 .095 .009 .009 .274 .075 .075***
Gender -.04 -.26***
Level .05 .02
Home .04 .09
MEDIA EXPOSURE 2 .257 .066 .057*** .379 .144 .069***
Newspaper .09 .01
Television News .17** .18**
Television -.09 -.14**
Step 1: F (3, 217) = .66 Step 1: F (3, 212) = 5.73***
Step 2: F (6, 214) = 2.52** Step 2: F (3, 212) = 5.86***
PRIME MINISTER EARTHQUAKE
R
2
Final R
2
Final
Step R R
2
Change R R
2
Change
DEMOGRAPHICS 1 .159 .025 .025** .289 .084 .084***
Gender -.09* -.20***
Level .02 .01
Home -.13** -.15**
MEDIA EXPOSURE 2 .193 .037 .012** .339 .115 .031**
Newspaper .04 .02
Television News .05 .18***
Television -.07 .01
Step 1: F (3, 431) = 3.73** Step 1: F (3, 235) = 7.16***
Step 2: F (6, 428) = 2.77** Step 2: F (6, 232) = 5.02***
NOTE
Independent variables are coded as follows: Gender (0 = Female, 1 = Male), Level (1 = Freshman, 2 = Sopho-
more, 3, Junior, 4 = Senior), Home (0 = yes, 1 = no), Newspaper and Television News (4-point Likert scale),
and Television (5-point Likert scale).
Significance for multiple regression: * p < .10, ** p < .05, *** p < .01
45
Keio Communication Review No. 26, 2004
In "Engagement," none of the demographic variables were significant at
the first step. The final equation after including media exposure variables ac-
counted for 7 percent of the diffusion speed (R = .26). Among media exposure
variables, television news viewing is the only significant variable ( = .17, p <
.05): the frequency of television news viewing is positively associated with the
speed of news reception.
On the other hand, demographic variables accounted for 8 percent of the
"Marriage" diffusion speed on the first step (p < .01). At this step, Gender was
the only significant predictor. Media exposure variables also made up 7 percent
of the diffusion speed on the second step. The final equation accounted for 14
percent of the variance in the "Marriage" diffusion speed (R = .38). While tele-
vision news viewing ( = .18, p < .05) was a significant positive predictor, tele-
vision viewing ( = -.14, p < .05) was a significant negative predictor among the
media exposure variables. In other words, the more television news one watches
the faster one receives the news; on the other hand, the more television one
watches in general the slower one receives the news. Gender ( = -.26, p < .01)
remained significant: women are faster in receiving the news even after control-
ling for other variables.
In "Prime Minister," demographic variables accounted for 3 percent of the
diffusion speed on the first step (p < .05). None of media exposure variables
contributed to the diffusion speed. Only Home is related with the speed of news
diffusion in the final equation ( = -.13, p < .05). Those who lived with their
parents were faster in receiving the news. Gender is marginally significant ( =
-.09, p < .10). The final equation accounted for 14 percent of the variance in the
"Marriage" diffusion speed (R = .38).
Demographic variables in "Earthquake" accounted for 8 percent of diffu-
sion speed on the first step (p < .01). At this step, Gender and Home were
significant predictors. Media exposure variables also made up 3 percent of the
diffusion speed on the second step. The final equation accounted for 12 percent
of the variance in the "Earthquake" diffusion speed (R = .34). Television news
viewing ( = .18, p < .01) was the only significant predictor among the media
exposure variables. Gender ( = -.20, p < .01) and Home ( = -.15, p < .05)
remained significant. Being a woman and living with one's parents are predic-
tors for faster speed of news diffusion. Faster reception of news for women is
again confirmed in "Earthquake."
To sum up the results in the regression test, television news viewing is an
important factor for receiving news events faster regardless of news category.
In terms of diffusion speed by gender, the findings of regression test generally
suggest that women are generally faster in receiving news events regardless of
news category. Home (living with one's parents) is a positive predictor for faster
news reception only in hard news events. This finding in Home indicates the
existence of a distinct line between hard news and tabloid news as is discussed
46
in the following section.
For Research Question 2 that deals with post communication behaviors, mul-
tiple logistic regression was used. Logistic regression measures the odds ratio of
a dichotomous response. The dependent variables of RQ 2 are binary response
variables, (1) whether or not one informed others of the news story (Inform: 1 =
yes, 0 = no) and (2) whether or not one made the news story a topic of conversa-
tion (Topic: 1 = yes, 0 = no). Thus, this statistical method can measure relative
contributions of independent variables to the odds ratio of the post behavior.
4
The
independent variables are the degree of news impression such as Good/Bad, Sur-
prise, and Reliability as detailed above. In addition to Gender, News Source (0 =
mass media, 1 = interpersonal) is also included as a control variable since the
initial news source may affect the post communication behaviors.
Table 2 shows the results of multiple logistic regression. Generally, Gen-
der is a predictor for the post communication behaviors: women are more likely
to engage in the post communication behaviors regardless of news category
(Topic in "Engagement": = 1.23, p < .05; Inform in "Marriage": = 1.23, p <
.05; Topic in "Marriage": = 1.49, p < .01; Inform in "Prime Minister": = .44,
p < .05). The information channel by which one first received news does not
seem a good predictor for the post behaviors. The only significant relationship
is found at Inform in "Engagement" ( = -.90, p < .05). Respondents who re-
ceived "Engagement" by media channels are more likely to pass along the news
event to others than those by interpersonal channel.
Among impression items, the degree of Surprise appears the only predictor
for post behaviors. The finding is interesting, but not surprising: the more sur-
prised people were the more likely they were to engage in post communication
behaviors (Inform in "Engagement": = .92, p < .01; Topic in "Engagement":
= 1.10, p < .01; Topic in "Marriage": = .76, p < .01; Topic in "Prime Minister":
= .45, p < .01). On the other hand, neither the degree of Good/Bad or Reliabil-
ity in any news event could predict the post behaviors except for a marginally
significant item (Good/Bad, Inform in "Marriage"). Only in "Earthquake," Sur-
prise was not a significant predictor for the post behaviors. The insignificance
could be attributed to the tremendous news value of the news event. The per-
centage of "surprised" respondents in "Earthquake" is the highest in these four
news events, 94 percent. The percentages of respondents who engaged in the
post behaviors in the new event are also among the highest, 90 percent for In-
form and 93% for Topic. This non-variance in response probably led to the
insignificant finding in "Surprise."
As for research question 3 (applicability of news diffusion theory to tab-
loid news), over all findings in tabloid news diffusion seem to be in accord
with hard news diffusion. News diffusion theory posits that diffusion is more
rapid and widespread as the importance of a news event increases (Gantz &
Tokinoya, 1987).
47
Keio Communication Review No. 26, 2004
ENGAGEMENT MARRIAGE
Inform Topic Inform Topic
Independent Vs Coefficient SE Coefficient SE Coefficient SE Coefficient SE
Gender -.11 .49 1.23** .59 1.23** .51 1.49*** .56
News Source -.90** .43 .01 .45 -.21 .49 -.03 .52
Impression Items
Good/Bad .33 .23 .21 .22 .51* .29 -.12 .28
Surprise .92*** .25 1.10*** .25 .28 .27 .76*** .29
Reliability .11 .30 -.31 .28 -.21 .33 -.33 .35
Logit Constant 4.68*** .99 2.84 .85 1.30 1.10 .36 1.10
PRIME MINISTER EARTHQUAKE
Inform Topic Inform Topic
Independent Vs Coefficient SE Coefficient SE Coefficient SE Coefficient SE
Gender .44** .21 .10 .21 1.77* 1.05 .49 .81
News Source -.69* .37 .62* .36 .18 .51 .57 .67
Impression Items
Good/Bad .16 .13 .07 .13 .26 .29 .43 .34
Surprise .19 .13 .45*** .14 .44 .35 .54 .38
Reliability .23 .16 .09 .16 -.06 .25 .31 .29
Logit Constant 1.23** .54 1.16** .55 3.25*** .69 4.72*** .89
NOTE
Independent variables are coded as follows: Gender (1 = Female, 0 = Male), News Source (1 = Interpersonal,
0 = Mass Media), Good/Bad, Surprise, and Reliability (1 = no, 2 = not very, 3 = relatively, yes, and 4 = very
much).
Dependent variables are coded as follows: Inform and Topic (1 = yes, 0 = no).
Significance for Multiple logistic regression: * p < .10, ** p < .05, *** p < .01
The data about the speed of receiving news (Time) were presented by time
interval in Chart 1. The slopes of the lines show the velocity of the diffusion
process for each news event. Since sleeping patterns during the night hours
affect the velocity of the news diffusion process, the hours between 0:00 to
6:00AM were deleted from the calculations in order to construct the standard-
ized velocity of diffusion (Rosengren, 1973; Rosengren, 1987). Then, respon-
dents who learned of the news events after midnight were considered to have
received them at 6:00 in the morning. Graphic plots of the accumulated num-
ber of knowers (Chart 1) show the differences among these news events. Zero
in the charts indicates receiving news at the same hour when the news was
first reported.
Table 2: Multiple Logit Regression Analysis: Factors Affecting Post Behaviors
48
Chart 1: News Diffusion Rates
0
10
20
30
40
50
60
70
80
90
100
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hours

P
e
r
c
e
n
t
Engagement Marriage Prime Minister Earthquake
49
Keio Communication Review No. 26, 2004
Taking a quick glance at the chart, the diffusion speed of "Marriage" is
apparently slower than that of the others. It took 19 hours to reach 90 percent
diffusion rates in the "Marriage" news. The other three news events reached 90
percent diffusion rates within nine hours. Whereas this study did not content
analyze the amount of coverage on these events or measure the degrees of im-
portance of these news events, "Marriage" was the least surprising and news-
worthy story among the others. There are a couple of reasons that explain the
relative "un-newsworthiness" of "Marriage": (1) only "Marriage" was not treated
as a front page story by the three elite newspapers, and (2) "Marriage" was an
unexpected event; however, the relative popularity of the couple (the bride was
a scandalous singer who had passed her prime and the bridegroom was an un-
known dentist) was much lower than that of the "Engagement" couple.
Relative slow diffusion of "Earthquake" in the first few hours could be at-
tributed to the early occurrence, 5:46AM. The diffusion rate, however, acceler-
ated as time went by. The diffusion rate of "Earthquake" caught up with and
eventually passed those of "Engagement" and "Prime Minister." This finding
matches one of the news diffusion propositions; the speed of diffusion is in ac-
cordance with the importance of a news event.
Another proposition of news diffusion theory postulates that interpersonal
communication becomes more active as the importance of a news event increases.
This is also confirmed in the tabloid news diffusion. Even though "Engage-
ment" was reported late at night, more respondents heard the news through in-
terpersonal channels than they did in "Marriage" which was aired early in the
morning. Further, "Engagement" triggered post communication behaviors more
than "Marriage" did. The literature suggests that news events occurring in the
evening are less likely to be diffused through interpersonal channels since many
people are at home. "Engagement" may have overcome the disadvantage in
interpersonal channel diffusion because of its greater news value.
Although it is not clearly stated, the news diffusion propositions by Gantz
& Tokinoya (1987) indicate that daily media use is positively associated with
the faster awareness of news. In all tabloid news events examined here, daily
television news viewing is positively associated with the speed of diffusion. It
is thus summarized that the diffusion of tabloid news events are similar to that
of hard news events except that (1) Home is a positive predictor of diffusion
speed only for hard news, and (2) women are slightly more active in post behav-
iors of tabloid news events than in hard news events.
Discussion
This study demonstrated news diffusion theory is generally applicable to
tabloid news even though the theory has been constructed exclusively with hard
50
news. Female respondents appear to engage more in tabloid news post behav-
iors than in hard news, yet the difference is not decisive. The only factor that
exclusively affects hard news diffusion is Home. Living with one's parents is a
significant positive predictor for the faster diffusion of the hard news events
only. It would be right to infer that parents inform their children more about
hard news stories than about tabloid news because of paternal responsibility.
Assuming this, the news categories, hard or tabloid, still matter to parents and/
or older generations. Put another way, there exists a distinct line between hard
news and tabloid news at least for certain demographic groups. This needs to be
verified by future studies.
The study has provided new evidence that news diffusion theory is gener-
ally applicable to the diffusion of tabloid news while indicating the possibility
of a distinct line between the news categories. Therefore, it augmented the theo-
retical scope in news diffusion study and helped understand the comprehensive
process of news diffusion. This is a major contribution to the literature.
The findings, however, should be further confirmed by empirical evidence
from a variety of news events and demographic groups. First, the hard news
events used in this study are exceptional and tremendous events. For example,
Prime Minister was not only a political news story but also an extraordinary
event considering the selection of a socialist Prime Minister in the second larg-
est capitalist nation in the world. Therefore, it cannot be concluded that women
talk about politics as an everyday topic. Likewise, the tabloid news events are
exclusively Hollywood gossip news stories. Topics of tabloid news range
from celebrities matters to shocking crimes and varieties of human interests. It
is questionable that shocking crime news events such as a bizarre murder trigger
female post behaviors in the same way as Hollywood gossip did.
NOTES
1. In the age of the Internet and other state-of-the-art information technolo-
gies, this statement may not be relevant anymore. Relative importance of
word of mouth may be diminishing. Research should incorporate the role
of the Internet. However, the data of this study were collected before the
proliferation of the Internet. Thus, this proposition should be appropriate
and this study did not consider the Internet a factor.
2. Variable Time was calculated by subtracting the time when respondents
received a news story from the time when the news was event first re-
ported. Thus, a smaller time lapse is interpreted as faster news reception;
a greater time lapse as slower news reception.
3. For newspaper reading, 1 = Always, 2 = Usually, 3 = Occasionally, and 4
= Rarely. For television news watching, 1 = Always, 2 = Usually, 3 =
51
Keio Communication Review No. 26, 2004
Occasionally, and 4 = Rarely. For television watching, 1 = Equal to or
more than 4 hours, 2 = Equal to or more than 3 hours and less than 4 hours,
3 = Equal to or more than 2 hours and less than 3 hours, 4 = Equal to or
more than 1 hour and less 2 hours, and 5 = Less than 1 hour.
4. Coefficient, , represents the change in the log odds of an event happening
for a unit change in each independent variable. Odds ratios can be ob-
tained by taking the inverse of the natural log of logit (p), e

. For example,
= 1.23, and e

= 3.42 which indicates that for each unit increase in an

independent variable, the odds of p change to 342% of its previous value,
an increase of 242%. For further details, see Hutcheson and Sofroniou
(1999).
REFERENCES
ALTER Jonathan (1995). "American Goes Tabloid". Newsweek, December 26:
34-36.
BANTZ Charles R., PETRONO Sandra G.., & RARICK David L. (1983). "News
Diffusion After the Reagan Shooting". Quarterly Journal of Speech, 69: 317-
327.
BIRD S. Elizabeth (1998). "News We Can Use: An Audience Perspective on the
Tabloidization of News in the United States". Javnost/The Public, 5 (3): 33-49.
CASE Tony (1992). "Gossip: Three New York City Columnists Discuss What
Has Become a Bona Fide Commodity to Newspapers". Editor & Publisher,
May 2: 36-38.
DEFLEUR Melvin L. (1987). "The Growth and Decline of Research on the Dif-
fusion of the News, 1945-1985". Communication Research, 14 (1): 109-130.
GANTZ Walter & TOKINOYA Hiroshi. (1987). "Diffusion of News about the
Assassination of Olof Palme: A Trans-Continental, Two-City Comparison of
the Process". European Journal of Communication, 2: 197-210.
GRABE Maria Elizabeth (1997). "Tabloid and Traditional Television News
Magazine Crime Stories: Crime Lessons and Reaffirmation of Social Class
Distinctions". Journalism & Mass Communication Quarterly, 73 (4): 926-946.
GREENBERG Bradley S. (1964). "Diffusion of News of the Kennedy Assassi-
nation". Public Opinion Quarterly, 28: 225-232.
HILL Richard J. & BONJEAN Charles M. (1964). "News Diffusion: A Test of
the Regularity Hypothesis". Journalism Quarterly, 41 (3): 336-342.
52
HUTCHESON Graeme & SOFRONIOU Nick. (1999). The Multivariate Social
Scientist. London: Sage.
LEVY S.G. (1969). "How Population Subgroups Differed in Knowledge of Six
Assassinations". Journalism Quarterly, 45: 685-698.
MCQUAIL Dennis (1994). Mass Communication Theory: An Introduction. Lon-
don: Sage.
MILLER Delbert C. (1945). "A Research Note on Mass Communication: How
Our Community Heard about the Death of President Roosevelt". American
Sociological Review, 10: 691-694.
RIFFE Daniel & STOVALL J.G. (1989). "Diffusion of News of Shuttle Disaster:
What Role for Emotional Response?" Journalism Quarterly, 66 (3): 551-556.
ROSENGREN Karl Erik (1973). "News Diffusion: An Overview". Journalism
Quarterly, 50 (1): 83-91.
ROSENGREN Karl Erik (1987). "The Comparative Study of News Diffusion".
European Journal of Communication, 2: 227-255.
THARP Mike & STREISAND Besty (1994). "Tabloid TV's Blood lust". U.S.
News & World Report, July 25: 46-48.