SITUATIONAL INFLUENCE OF POLITICAL INVOLVEMENT ON INFORMATION SEEKING: A FIELD
Stacey Frank and Steven Chaffee
Paper submitted to the Communication Theory and Methodology Division for the
1996 Annual Convention of the Association for Education
in Journalism and Mass Communication, Anaheim, California
Stacey Frank is Assistant Professor of Mass Communication at the University of
St. Thomas in Saint Paul, Minnesota. Steven Chaffee is Janet M. Peck Professor
of International Communication at Stanford University. This paper is based on
the doctoral dissertation of Stacey Frank at Stanford University.
Phone: (612) 962-5265 Fax: (612) 962-6360
E-mail: [log in to unmask]
2115 Summit Ave. #4371
St. Paul, MN 55105-1096
SITUATIONAL INFLUENCE OF POLITICAL INVOLVEMENT ON INFORMATION SEEKING: A FIELD
Political involvement is conceptualized as a situational variable and evaluated
experimentally. Each subject was included in a 30-minute current events
discussion in one month, and a control condition in another month. This
discussion stimulated subjects to acquire more political news than they did in
the control condition, and to seek information via newspapers. These findings
support a model of political involvement as a condition that varies over time,
not simply a stable individual difference.
Situational Political Involvement / 1
Political involvement, or a general engagement with politics and public
affairs, is often used to explain why people read newspapers, watch television
news, and talk to others about politics. This paper will focus on the
situational component of involvement and its potential to stimulate political
communication behavior. We will explore whether political involvement, which
usually is assumed to be a matter of deep-seated personal characteristics, can
be temporarily stimulated by a minor social interaction. Then, in a
within-subjects field experiment, we will investigate the effect of a
situational manipulation of involvement -- via interpersonal discussion -- on
information seeking in the media and reception of political news.
Dispositional involvement. Early political scholars defined involvement as both
interest in politics in general, and interest in specific election campaigns in
particular (Lazarsfeld, Berelson and Gaudet 1944). This definition was refined
in the University of Michigan's national election studies (Eulau and Schneider
1956; Campbell, Converse, Miller and Stokes 1960). These scholars considered
involvement a psychological trait, a person's deep-seated and enduring sense of
political efficacy and citizen duty, as well as concern about a particular
election campaign and its outcome. Many political involvement studies follow
these definitions; some conceptualize involvement as both general and specific
political interest (Pettey 1988; Faber, Tims and Schmitt 1993), while others
focus on concern with a specific election (Atkin, Galloway and Nayman 1976;
Pederson 1978; Morley 1984).
Income and education are typical of the stable personal characteristics that
predict political involvement as it is tapped with survey questions (Verba and
Nie 1972; Conway 1985). Indeed, in each of the studies cited above, involvement
was considered a predisposition, measured to account for individual differences
in political interest.
Situational involvement.1 Social psychologists have amassed a great deal of
empirical evidence that behavior is less a function of differences in
personality, or stable individual dispositions, and more related to differences
between situations -- the social context in which individuals find themselves
(Ross and Nisbett 1991). Political communication has been found to be influenced
by at least two situational factors: (1) characteristics of specific elections;
and (2) social participation in political discussions.
Rothschild and Ray (1974) used differences between elections in a study of
advertising effects. Involvement was operationalized by type of election:
Presidential (high involvement) and state and local races (lower involvement).
They found effects on voting intention due to repetition of an advertisement in
the low involvement elections, but for high involvement offices advertisement
repetition made little difference.
Milbrath and Goel (1977) cite correlational evidence indicating that
participation in political discussions may stimulate political involvement. In a
panel study, Tan (1980) inferred a causal relationship between discussion and
newspaper use, suggesting that discussion may be a situational factor that
precipitates involvement and information seeking. Social participation theorists
also suggest that interpersonal interaction, where politics is discussed, can
stimulate political interest (Olsen 1972; Erbe 1964).
Political behavior that reflects structural factors such as income and
education, or deep-seated attitudes, is subject to little change. But when
involvement is stimulated by a temporary social situation, political activity
may increase as well (Atkin 1972). Olsen (p. 318) argued that social context may
elicit political behaviors above one's habitual level: "People caught in
traditional patterns [of political participation] must be mobilized through
involvement in new social contexts if they are to become politically active" (p.
318). Erbe (1964) also noted that "interaction with people" stimulates interest
and curiosity with the topic at hand.
Situational involvement in politics activates a need for political or public
affairs information (Chaffee and McLeod 1973), and this need drives a motivated
information search (McCombs 1972). The assumption that individuals can be active
information seekers is consistent with the uses and gratifications research
tradition (Katz, Gurevitch and Haas 1973; Blumler and Katz 1974). Active seeking
moves beyond passive, habitual information intake; it is said to be purposeful
(Atkin 1973), deliberate (Gantz, Fitzmaurice and Fink 1991), and a result of
effort (Donohew and Tipton 1973).
Information-seeking channels. Research indicates that motivated information
seekers turn to newspapers to learn about politics (McLeod and Becker 1981;
Culbertson and Stempel 1986). Pettey (1988) showed that variance in political
knowledge is explained by information seeking in newspapers, but not on
television. Tan (1980) also concludes that when people need to be informed about
politics, they seek information in newspapers, but not on television.
Atkin (1973) examined the relationship between expenditure of effort (to seek
information) and media use. He suggested that newspapers require more effort to
use than television, and are thus more costly. However, as need for information
increases, people turn to more costly sources if they expect a greater
informational benefit. Specific situations can stimulate information searching
in print media. In one study, women who were likely to have an abortion sought
more information from print sources, compared to women who were not; the latter
got their information from a variety of channels (Atkin 1973). Print sources
become more valuable as utility increases. Chew (1994) found that respondents
with more diverse information needs extend their search beyond television, and
turn to newspapers and magazines to satisfy those needs.
To be sure, most people say they habitually rely on television for political
information (Roper 1983), and people do learn about political issues from
watching television news (Chaffee and Yang 1990). Gantz et al. (1991) reported
that most respondents wanting to learn about the 1986 Reagan-Gorbachev summit
turned to television news, for example.
Discussion about politics can be considered a supplemental information source
(Chaffee 1982). Studies on the diffusion of news (Rosengren 1973) indicate that
people seek information from others when the event is sudden and of crisis
proportion, such as the assassination of a political leader or a natural
disaster. Ordinarily, though, people talk about politics less for informational
and learning purposes, and more to enjoy themselves or to persuade others to
adopt similar political views (Rubin, Perse and Barbato 1988; Barbato 1986).
Information-seeking measures. A standard measure of information seeking is
frequency of media news exposure, measured by asking how often the person reads
a daily newspaper or watches television news about government and politics
(McLeod and McDonald 1985; Culbertson and Stempel 1986; Tan 1980; Gantz et al.
1991). A more motivated aspect of information-seeking, specific to news, is
assessed by asking people how much attention they pay to television (or,
separately, newspaper or radio) stories about politics or public affairs. This
measure of information seeking is especially important when examining television
use (Chaffee and Schleuder 1986). For television news, "attention" questions
more strongly predict learning than does frequency of mere exposure, probably
because much exposure to television is unmotivated, often accompanied by other
activities. Asking about "attention" may better capture the active component
inherent in information seeking.
Information seeking may also be assessed by recall of media content, or by
measuring an individual's knowledge of particular media messages. These tests of
reception include newspaper or television campaign items, and political
advertisements. Martinelli and Chaffee (1995) used news recall to detect active
information seeking, for instance. Price and Zaller (1993) label this concept
"news reception." According to these researchers, reception indicates who
actually "gets the news": it requires "attending to, comprehending and retaining
news" (p. 134). While news reception may not be specific to channel, it does map
variance in the general activity of news communication, and is a more reliable
indicator than is self-reported media use (see also Gandy et al. 1987).
Seeking information via interpersonal discussion, as measured by self-reports,
is correlated with other information-seeking behaviors (Eulau et al. 1979; Tan
1980; Robinson and Levy, 1986; Kinsey and Chaffee 1994). Berkowitz and Pritchard
(1989) examined whether people use interpersonal and media channels to seek
information; they found no effect on knowledge for interpersonal channels,
SITUATIONAL INVOLVEMENT AND EXPERIMENTAL EFFECTS
In studies that examine the relationship between political involvement and
information seeking, involvement almost invariably has been a measured
dispositional variable. Its inferred effect is thus confounded with many
correlates, such as education or media use itself. Conceptualizing involvement
as "situational," on the other hand, implies that it is subject to experimental
manipulation, separated from potential confounding factors. The two social
situations that scholars have shown experimentally affect involvement are
general election characteristics (Rothschild and Ray 1974) and group discussion
(Atkin 1972). The manifest importance of different offices is an attribute of
the macroscopic situation, while a discussion about politics is a "microscopic
situation" that is subject to social and experimental change, and is the
situation we use in this study. Involvement has been manipulated in other
communication effects studies (Bybee 1978; Roser and Thompson 1995), but the
manipulations involve personal relevance of a message or issue, not an
engagement with politics outside of self interest.
Research questions. Can involvement in politics be stimulated by a social
situation (in contrast to being measured by enduring individual differences)?
Can this situational stimulation result from a single discussion about politics
without any instruction or occasion for further action? And will this
involvement lead to information seeking and news reception? This set of
questions will be addressed by comparing the media use and related behaviors of
an individual after he or she (1) has participated in a recent discussion about
politics, and (2) has not participated in such a discussion.
This study is a within-subjects field experiment, in which involvement is
manipulated by participation in a discussion about politics in one of two time
periods. Subjects were randomly assigned to either a discussion or control
condition in one month, and then to the other condition in the following month.
In the control condition, subjects were not asked to participate in a discussion
about politics. The study was conducted over a two-month period in a
counterbalanced design, so that each subject eventually was in both the
experimental condition and the control condition and in effect served as his or
her own control. Measures were gathered at three time points, so there are
baseline data (Time 1) and then dependent measures for the experimental
conditions after each time period (Times 3 and 5). The experimental design is
shown in Figure 1.
The data were collected at a mid-sized, residential university in California in
the fall of 1994. The first time period coincided with a California election
campaign, which included races for governor and U.S. Senator and several
controversial referenda. The second time period was post-election, and the news
topics were quite different. Subjects (N = 121) were students in Mass
Communication and Society, an introductory course.
Experimental Procedure. The class met in a large lecture hall three times per
week; students also met in 12 one-hour discussion sections once a week. The
survey questionnaires were administered in three of the lectures during the
quarter; the experimental conditions were run once in each of the 12 discussion
sections. The principal investigator of this study was the head teaching
assistant for the course and administered all surveys; she also led the
discussions that constituted the experimental manipulation. These activities
were not connected, however. Six of the 12 discussion sections were randomly
assigned to Group 1, the other six to Group 2. Subjects in Group 1 participated
in a discussion about politics before the November 8, 1994 California election;
subjects in Group 2 participated in a discussion about politics after the
election. The control condition, i.e. no discussion, was thus assigned to Group
2 in the pre-election period, and to Group 1 in the later period. The procedure
Time-1 (October 10): Survey I. A questionnaire, entitled "Media Survey," was
distributed during the large lecture class. It was explained that, as a class
project, the class was going to track its media behavior as it related to the
ongoing election campaign. Students were told participation was voluntary, and
that it would not affect their grade, but it would be a useful experience to
help them understand the relationship between the mass media and the electorate
as audience (a topic covered in the course). They were told that results of the
survey would be tabulated and presented in a lecture at the end of the quarter.
(This lecture was given). The questionnaire included items measuring
information-seeking behavior: news reception, news exposure, news attention, and
interpersonal communication, as described later here.
Time-2 (Week of October 31): Experiment Phase I. Six sections of 13 - 18
students each comprised Group 1. In each section students discussed politics
during one weekly meeting. Each section leader introduced the head teaching
assistant (i.e. the experimenter), and explained that she would be running that
day's discussion. The experimenter explained that she was a Ph.D. student,
studying the relationship between the mass media and the election process. She
reminded them of the upcoming election and pointed to the board, which listed
six items on the ballot: (1) Pete Wilson v. Kathleen Brown for Governor; (2)
Dianne Feinstein v. Michael Huffington for U.S. Senator; (3) Health care (an
initiative for a state-run, single-payer health care system); (4) Illegal
immigration (an initiative to prohibit public social services for illegal
immigrants); (5) Smoking (an initiative to relax smoking restrictions in
public); and (6) Three strikes (an initiative for mandatory life sentences for
felons on their third conviction).
The experimenter then told the students, "I want us all to participate in a
group discussion about these -- or any other -- topics concerning politics." The
list was provided as a reminder of election-related topics, but students were
encouraged to discuss anything related to the election or politics. She
explained that she was more interested in having a class discussion about
politics than on these items in particular. She told the class she was
tape-recording the discussion so she could later content analyze it, and that
participation was voluntary and would not be graded. Excluding introductory
remarks and unrelated class business, each discussion lasted approximately 30
Time-3 (November 9): Survey II. The second questionnaire was distributed to all
subjects before lecture began, which was the day after the election. Items
similar to those in Survey I remeasured information-seeking behavior and news
Time-4 (Week of Nov. 28): Experiment Phase II. The other six sections, ranging
from 10 to 16 students each, which were randomly assigned to Group 2 and which
had been the control group during Phase I, discussed politics at their weekly
meeting in this post-election period. Subjects in Group 1 now constituted the
control group, and did not have a class discussion about politics. (The control
topic of discussion was simply the regularly scheduled lesson, led by the usual
teaching assistant as had been the case for Group 2 the previous month.) In the
treatment sections, the experimenter repeated the procedure used when Group 1
had been the treatment group. However, she changed the topics listed on the
board to reflect current political news. Topics were: (1) Prayer in schools
(proposed by new Republican leaders in Congress); (2) Cuts in federal welfare
(proposed by new Republican leaders in Congress); (3) Sen. Jesse Helms's attacks
on President Clinton (criticizing Clinton's ability to be commander-in-chief);
and (4) GATT (treaty expansion which was up for vote in Congress during this
Time-5 (December 5): Survey III. The third questionnaire was handed out to all
subjects in lecture. It again measured information-seeking behavior and news
reception with items similar to the Survey I and Survey II measures.
One week later in lecture, marginal results of the questionnaire data were
presented with an explanation of the relationship of the class project to the
research literature on media use and political involvement.
Independent variables. Political involvement was manipulated in this
experiment by creating the temporary situation in which subjects participated in
a half-hour discussion about politics in general. Several studies, summarized by
Milbrath and Goel (1977, p. 36) suggest that persons participating in political
discussions are more likely than nondiscussants to become more interested and
active in the political process. This manipulation was designed to create a
situation that might generate in the students a sense of political involvement.
There was, however, no assignment to seek information, nor any suggestion that
there would be any further discussion of current politics. This was, then, a
mild manipulation, in contrast to other studies that manipulate involvement by
altering the personal relevance of a message (e.g. percent of course grade) or
by requiring subjects to engage in a specific task.
Frequency of topic discussion was measured by adding the number of times a
topic was mentioned during the session and dividing that sum by the total number
of times any student spoke about politics during the session. Content analysis
showed that during the pre-election sessions, more than half of the discussion
(53 percent) was about three topics that reflected current political news: (1)
the referendum to cut social services for illegal immigrants (23 percent); (2)
the referendum for a single-payer health care system (17 percent); and (3) the
U.S. Senate race between Dianne Feinstein and Michael Huffington (13 percent).
Other pre-election topics included general discussion of campaigns.
During the post-election treatment condition, 80 percent of the discussion
concerned just three current news topics: prayer in public schools (47 percent);
GATT (19 percent), and welfare reform (14 percent). Other post-election topics
included new Republican leaders.
During the discussion (treatment) sessions, subjects spoke on a voluntary basis
and were not graded or otherwise required to talk. Participation for purposes of
this experiment meant being in the room in which the discussion was taking place
-- some subjects talked, while others simply listened. In this way, all subjects
were considered a part of the situation, although not necessarily of the
Control condition. The control condition provided a baseline to assess the
effectiveness of the political involvement manipulation. Because this experiment
used a within-subjects design, each subject not only received the treatment
(i.e., participated in the discussion), but also "received" the control
condition. This made it possible to compare the effects of each person being in
the treatment condition with the effects on that same person when in the control
condition, eliminating individual differences in enduring political involvement,
media use habits, and other extraneous factors from the test of the main
proposition. Topics during the control meetings reflected the weekly class
lecture material and did not contain information being discussed in the
News reception measures. Following Price and Zaller (1993), news reception was
measured by recall of specific news content. For Survey I and Survey II, a total
of 28 items, shown in the Appendix, were developed to tap recall of California
1994 election coverage in the mass media. This included knowledge of (1)
candidate issue positions, (2) candidate personal information, (3) general
election news, (4) and information contained in television attack ads. These
items were divided into 14 pairs that dealt with similar current political
information, and then one question from each pair was randomly assigned to
either Survey I or Survey II. This procedure was designed to maximize
comparability of these two waves of measurement in terms of the overall mean,
variance, normality, and distribution of content. (This comparability could not
be achieved for Survey III as well, since it was impossible to anticipate at
Time 1 what the post-election news would be about). For example, items used to
assess knowledge of the different kinds of election coverage include: (1) "Who
sponsored a bill to create national parks in the California desert?" (2) "Who
offered to debate only on a PBS TV channel in Sacramento?" (3) "Who was accused
in TV ads of being 'a career politician' who rides to work in a limousine?" and
(4) "Who is married to an ordained minister?" The closed-ended choices were the
last names of the candidates running for California Governor and U.S. Senator
for California: "Brown," "Feinstein," "Huffington," and "Wilson." A "don't know"
option was also included to discourage random guesses. These indices were summed
and standardized to form measures of news reception at each time point.
The treatment sessions were audio-tape recorded and transcribed. The
transcripts were content analyzed by topic, enabling deletion from the recall
items on the questionnaires any topics that had been discussed during any
treatment session. Before data analysis, two items were dropped from the Survey
II index because they had been discussed during the Group 1 treatment
discussion. Any changes in this dependent variable should be influenced by
extra-curricular information-seeking behavior, and not by direct classroom
learning. To ensure that Survey I and Survey II would remain comparable, two
parallel items were dropped from the Survey I index. The overall reliability of
the Survey I news reception index (Cronbach's alpha) was .91. Reliability of the
Survey II index was .87.
News reception items for Survey III (shown in the Appendix) were based on
post-election political news, such as (1) "Who is the next U.S. House of
Representatives minority leader?" (2) "Who once said he would support GATT if
there were a cut in the capital gains tax?" and (3) "What has the United Nations
decided to do with Bosnia?" Four response choices were also given for these
items. This index was summed and standardized to form the Survey III measure of
news reception. Three items were dropped from the index because they had been
covered in a Group 2 treatment discussion.
Reliability for the Survey III reception index was lower than for the first two
surveys (Cronbach's alpha = .74). The decreased reliability may be due to fewer
items comprising the index (nine, compared to 12 for the first two scales), or
it may be due to the heterogeneity of the test, in that the Survey I and II
items dealt mainly with the upcoming election. Test items were based directly on
news from metropolitan daily newspapers and evening television news broadcasts.
Items would therefore appear difficult if a subject had not been attending to
and comprehending current news. Items likely seemed more difficult during Survey
III, a month after the election, because news about politics generally decreases
after an election. Another possibility for the lower reliability could be the
difference in response scale for Survey III. All response choices for Survey I
and Survey II media reception were the same: the names of four candidates. The
Survey III index contained several different response choices because less
political news at that time focused on different people. While these differences
in response choices could have decreased internal consistency in the Survey III
test, it is unlikely that they reduced validity. A subject could score
consistently well -- or consistently poorly -- on the more similar measures of
Surveys I and II simply because a given person might have been relatively
"expert" regarding the particular candidates who comprised the response set for
those two tests.
Newspaper exposure. On each of the three questionnaires, exposure was measured
by asking: "About how many days in the past week have you read a newspaper other
than [name of school newspaper]?" This was measured on an 8-point scale, ranging
from 0 to 7 days per week.
Newspaper attention. This was measured on all three questionnaires with a
5-point scale, in which 1 meant "no attention" and 5 meant "close attention." On
Surveys I and II, the question read, "How much attention are you paying to
newspaper articles about the election campaign?" Consistent with the change in
news context, the Survey III item was, "How much attention have you paid in the
past week to newspaper articles about politics?"
Exposure to television news. On all three questionnaires, items read: "About
how many days in the past week have you watched a national TV news show?" A
separate item was included about exposure to local TV news. Responses were
recorded on an 8-point scale ranging from 0 to 7 days.
Attention to television political news. This was measured on all three
questionnaires with a 5-point scale, in which 1 was labeled "no attention" and 5
"close attention." On Surveys I and II, the question read, "How much attention
are you paying to television news about the election campaign?" The Survey III
item was, "How much attention have you paid in the past week to television news
Frequency of political discussion. On Surveys I and II, the item read: "In the
past week, how many times do you estimate you have talked to other students or
family members about this year's election?" On Survey III, the item was: "In the
past week, how many times do you estimate you have talked to other students or
family members about current politics?" This was measured using an 11-point
scale ranging from 0 to 10 (or more) times in the past week.
The mean and standard deviation for each dependent variable (for each group and
on each survey) are shown in Table 1.
HYPOTHESIS AND PLAN FOR ANALYSIS
The working assumption of this research is that political involvement, defined
as interest in politics, can be stimulated by a situational experience. In this
study, the situation is the discussion in class about politics. Because prior
literature on the experimental manipulation of political involvement is not a
clear guide, the research hypothesis advanced here is general:
When people become involved in politics because of a situational experience,
they will seek more political information than at other times.
This experiment provides many tests of this proposition.
Research generally presumes that information seeking is evidence of a condition
of political involvement (Atkin et al. 1976; Perloff 1985; Verba and Nie 1972).
The design of this study enables us to compare the same person's
information-seeking behavior when he or she is situationally involved, and when
he or she is not.
This study uses multiple measures for information seeking behavior. This will
permit us to compare the news reception index with more traditional self-report
measures, and to compare information seeking via various channels.
The central empirical test in this study is the examination of change scores on
the media use dependent variables in both the treatment and control conditions.
Because there were three points of measurement, and because each person was in
both the treatment and control conditions, it is possible to calculate the
change in media use from pretest to posttest when a person was in the treatment
condition and when he or she was in the control condition. These short-term
changes can then be compared to one another. The hypothesis of this research can
be tested in two ways: whether the change in media use was significantly greater
when a person was in the treatment condition as compared to when he or she was
in the control condition, and, at a given time, whether those in the treatment
(discussion) condition subsequently sought and received more information than
did those in the control condition. Before computing change scores, all
dependent measures were standardized within times so that comparisons could be
drawn that are not confounded by different variances in a given dependent
variable at different times.
The first step is to examine the media use dependent measures using
between-groups analysis. This method permits examination of differences between
treatment and control groups on each dependent variable at each of the three
points in time. This in effect provides a replication of the experiment: In the
discussion manipulation that occurred before the election, Group 1 was the
treatment group and Group 2 was the control. In the discussion after the
election, Group 2 was the treatment group and Group 1 was the control. The
robustness of the experimental effect can be evaluated by first viewing these
two comparisons as separate replications.
Data Administration. Data were matched for each of the three surveys by
assigning an identification number for each subject. Attendance was taken during
the treatment sessions. Only students who completed all three questionnaires and
attended their section's treatment session were included in the analysis. This
resulted in some attrition; for Survey I (N = 161), for Survey II (N = 152), and
for Survey III (N = 142). Of the 89 subjects randomly assigned to Group 1, 80
attended the treatment session. Of the 76 subjects randomly assigned to Group 2,
65 attended the treatment session. And, 24 of those who attended treatment
sessions did not complete one of the three questionnaires. Therefore, the final
number of subjects included for data analysis (i.e. those for whom a complete
record is available) was 121.
Analyses Between Groups. Table 1 shows between-group differences in standard
scores on all dependent variables for both the pre-election and the
post-election experimental manipulations. The data in the Survey I column
indicate baseline measures for all subjects. Half the discussion sections were
randomly assigned to Group 1, the other half to Group 2. As expected, there were
no significant differences between these groups on mean responses to any media
dependent variable at Time 1. The tests of changes in each time period are
conservative tests of the research hypothesis, since they do not control for
within-subjects habits of political communication; those individual differences
are effectively removed from the within-subjects analyses reported below.
Examining the results of Surveys II and III, subjects in the October political
discussion condition had higher news reception scores than did subjects in the
October control condition (Table 1). This difference approached significance (p
Insert Table 1 about here
The control condition news reception scores actually decreased while the
treatment condition scores increased. In November, when the discussion took
place after the election, news reception for subjects in the control condition
again decreased, but scores for subjects in the treatment condition increased,
enough to achieve significant differences between groups on news reception (p <
.05). While neither difference is statistically strong, these results provide
overall support for the assumption that political discussion leads to
information gathering outside of the classroom situation.
Turning to the newspaper exposure measure, after the October discussion,
subjects in the treatment condition did not report reading the newspaper more
often than did subjects in the control condition, as shown in Table 1. However,
after the November discussion, subjects in the treatment condition did increase
their newspaper reading frequency compared to the control condition (p < .05).
This result suggests that the November manipulation may have been stronger than
the October version. It may have been less unusual to have a discussion about
politics preceding the election when casual conversations about politics are
more likely. Thus, the manipulation may not have been strong enough to alter
newspaper reading, usually a habitual, stable communication behavior. However, a
political discussion three weeks after the election -- in which election
consequences comprised the topics -- could be relatively more stimulating. In
this period, we find a significant increase in newspaper reading for the group
that participated in the discussion.
Regarding newspaper attention, the independent variable of political discussion
has a significant effect (p < .05) on the amount of attention subjects in the
October treatment said they gave to the newspaper stories they read about
politics. While the October treatment may not have led to an increase in
frequency of newspaper reading, subjects in the treatment group did increase
their self-reported attention to the political news they were reading. After the
November treatment, newspaper attention also increased in the predicted
direction for subjects in the discussion group, although this result did not
The political discussion treatment did not have a significant effect on mean
responses for either national or local television news exposure for subjects in
the October or the November manipulations, as compared with the control subjects
in each period. Television news use is apparently more habitual, and less
subject to short-term changes, than is the case with newspaper reading. In
addition, television viewing is often a group activity for most students at this
residential university, and therefore may not be subject to individual control.
The difference between treatment and control condition means on attention to
television news about politics approached significance (p = .09) in the November
manipulation. There were no significant differences between conditions for
television news attention following the October manipulation.
For the October manipulation, there were no significant differences in means
between treatment and control condition on reported frequency of political
discussion. For the November manipulation, subjects in the treatment condition
talked to others about politics more frequently than did subjects in the control
condition. This result approached significance (p < .08).
Overall the between-subjects results comparing treatment and control conditions
indicate significant effects on media use. This analysis suggests that the
experimental manipulation of political discussion did stimulate the
information-seeking behavior of subjects in the treatment condition. The
analysis of most general interest will now be considered -- whether the same
subject sought more information when he or she was in the treatment condition,
compared to the control condition.
Analyses Within Subjects. Data relevant to within-subjects analysis for all
dependent variables are shown in Table 2. Each subject was in the treatment
condition in one time period and in the control condition at the other time. The
first column of Table 2 data shows the means for each dependent variable when
subjects were in the treatment condition (following class discussion about
politics). The second column shows the means for each dependent variable when
subjects were in the control condition. These means represent change scores
between measures before and after each subject was in these contrasting
conditions. The third column shows the difference between treatment and control
change scores, and the fourth column indicates statistical significance of these
change scores. These are highly sensitive tests of the research hypothesis,
since the net results are not affected by individual differences in habitual
political communication. The comparison of each subject with himself reduces the
error variance, and also controls for shifts in news content over time, such as
campaign news compared to post-election news.
Insert Table 2 about here
Did the one-time, in-class political discussion stimulate involvement in
politics as evidenced by subsequent information seeking in the mass media?
Results in Table 2 are presented from strongest to weakest net difference, as
indicated by t-values. All net differences are presented in relation to the
direction hypothesized, so that a positive sign indicates support for the
News reception. As shown in Table 2, after subjects participated in the class
discussion about politics, they recalled more information about current
political news than when they did not participate in the discussion (p < .01).
Since news reception requires news communication, we assume that the latter
class of behaviors was significantly affected by the experimental manipulation,
even if we should fail to find differences on self-report measures (below) in
the survey. That is, according to the best measure -- the one least contaminated
by social desirability and other frailties of self-reports -- we have clear
evidence that the experimental treatment stimulated the acquisition of
information about current news. Just how it might have been acquired is the
subject of our remaining analyses.
Newspaper attention. After subjects were in the discussion treatment, they said
they had been paying more attention to political information in newspapers than
after they were in the control condition (p < .01). This result cross-validates
the results for news reception, and explains why reception increased without a
change in newspaper reading habits. Subjects may not have increased their
frequency of daily newspaper reading, but they apparently changed in intensity
or focus on political news as they encountered it.
There were no significant net differences, as an effect of participating in
the treatment condition, in mean responses for political discussion frequency,
local or national television news exposure, television news attention, or
newspaper exposure. It is of course possible that some of those behaviors were
affected, but our self-report measures were inadequate to detect the minor
shifts that we did pick up with the news reception tests.
Tests of Hypothesis. The central hypothesis tested in this study holds that
when people become involved in politics because of a situational influence (in
this case, a discussion about politics), they will seek more political
information than at other times. This hypothesis is supported by the overall
pattern of results in Table 2; all are in the predicted direction, suggesting
that participation in the discussion did have a positive effect on
information-seeking behavior in general. This hypothesis also is supported by
the news reception and newspaper attention results, suggesting that the social
situation stimulated an active information search in the newspaper in
particular, one that resulted in information acquisition, and that subjects were
at least partly aware of this subtle shift in their media behavior.
This hypothesis was not supported by the information-seeking items that
measured television use or interpersonal communication. This suggests that when
people are situationally involved in politics, they may not seek information on
television or through interpersonal channels. There is, though, the possibility
of Type II error, in that these items may simply not have been sensitive enough
to detect effects on the behaviors in question. It should be noted that these
measures were modeled on standard survey questions, which are generally designed
to assess stable traits rather than situational activities.
This study demonstrates the power of even a minor situation to affect
communication behavior. The central purpose was to test empirically whether
political involvement could be influenced by situational factors, beyond
enduring personal traits and structural characteristics such as education and
income. This research offers experimental support for this proposition.
Classroom discussion did stimulate students' involvement in politics and
consequent use of newspapers, at the least, to gain information. Even if
subjects were not fully aware they had temporarily changed, and even if they did
not modify all their communication habits, the information was there when we
tested them on it. Knowledge is not a measure a person can fake, nor is it
subject to ingratiation or other extraneous factors that render self-reports
dubious. The treatment worked, exactly as we expected it to. And it was a most
minor intervention in these students' lives.
The situational variability of political involvement is underscored by the fact
that this experiment produced statistically detectable effects despite the
minimal nature of the "treatment" condition. The discussion lasted only about
half an hour, and students were under no compulsion even to participate. Topics
that were explicitly discussed were removed from the dependent measures. There
was no instruction or implication that information should be sought. And yet the
predicted result was replicated in the two time periods for several measures.
Prior studies have defined political involvement as habitual interest in or
concern about politics. This research, by contrast, supports the view that
political involvement, stimulated by a particular situation, may be conceptually
different from mere interest or concern -- terms that suggest a relatively
passive or routinized orientation toward politics. This traditional
conceptualization may be suitable for "predicting" habitual levels of media use
and political discussion in undisturbed field situations. But when involvement
refers to a motivational process, a more active conceptualization is more
fitting. The data in Table 2 suggest that the discussion situation stimulated
political communication activity in individuals regardless of their underlying
predispositions toward politics. Situational political involvement, thus, can be
thought of as a process of communication activation that manifests itself in
information seeking, and perhaps in other ways, as well.
A recent field experiment by Chaffee, Moon and McDevitt, with Pan, McLeod and
Eveland (1995) similarly found that a curriculum intervention can lead to a
state of political activation that, in turn, stimulates political communication.
After intensive elementary and secondary classroom instruction about voting
during an election campaign, significant increases were found in current news
knowledge, newspaper reading, television news viewing and interpersonal
discussion about politics. These effects had subsided six months later, although
there were residual effects on newspaper reading and public affairs knowledge
(Chaffee, Moon and McDevitt 1996). The authors suggest that political
socialization should be reconceptualized in terms of stimulation of
communication behavior, rather than passive learning and adherence to systematic
How do we know that discussion stimulated political involvement, specifically
as a process of communication activation? First, the strongest results
supporting this proposition were found with the least questionable measures
(i.e., the news reception index). Significant effects were found in both time
periods for the news reception measures of current news knowledge. This kind of
measure overcomes the social desirability problems, as well as uncertainties of
recall and quantitative estimation, that are inherent in self-reported media
use. While subjects may overestimate or otherwise erroneously describe their
media use in self reports, a high news reception score can only be obtained if
the person has been following recent events as reported in the news media. There
may, of course, be differences in recall due to intelligence, say, but these
factors were controlled in our design because each subject contributed equally
to the treatment and control data in Table 2.
The news reception measure can capture slight increases in information seeking
that may go unnoticed even by the subject. Given the subtlety of the discussion
manipulation, a subject might well not have been aware either that involvement
had been stimulated or that communication activity had increased. Nor was there
any indication that such a change was desired, or expected, by the instructor
leading the discussion. Because of the unobtrusive nature of the knowledge quiz
(relative to its true purpose), subjects may not have been aware that monthly
changes in their media use were being measured. The news knowledge quizzes
occurred in the context of a course on mass media, and it is likely that
students simply tried to do their best at this task. If students answered more
news reception items correctly after participating in the discussion, we can
infer that political awareness and active media use had been stimulated. These
"news reception" measures combine the results of each student's attention,
comprehension, and recall regarding political news. The short-term increase in
these indicators suggests purposive information seeking driven by an activated
awareness of politics.
The newspaper was the channel apparently most affected by situationally induced
information seeking, perhaps because newspaper reading may be more amenable to
situational changes in media behavior. It is subject to more individual control
than is television: newspaper readers can set their own pace, skip some articles
and re-read others. Television news controls both the topic and the pace of news
presentation, while others in the same household often control whether news will
be seen (or heard) at all. And, it is likely that to someone stimulated to learn
about current political affairs, the newspaper offers a richer volume and
variety of relevant information.
External Validity. How generalizable are these findings? The result indicating
the newspaper to be the channel used most for situationally induced information
seeking may be particular to college students at a residential campus. These
students, who mostly live in dormitories, have less access to television than
does the general population, and students who do watch television usually do so
in groups. Newspapers, in contrast, are accessible to all dormitory residents,
and can be read individually on one's own schedule. While newspapers might be
regarded as, on average, a more "costly" medium than television (Atkin 1973),
the expense may be quite small for a university student who reads regularly,
leads an independent life and may value a channel that is printed and somewhat
But our strong newspaper findings may also indicate less about a subject's
lifestyle and more about the nature of purposive information seeking. Surveys of
general household populations show that individuals who are actively seeking
political content also tend to go to the newspaper for it. This result has been
found, for instance, with Lubbock, Texas voters (Tan 1980); Dane County,
Wisconsin voters (Pettey 1988); and 5th through 12th grade students in San Jose,
California (Chaffee et al. 1995).
Measurement Issues. Significant information-seeking effects were found using
the news reception index, which measured recall of specific news stories. Not
only does this measure have high reliability, it also has high face validity: by
testing subjects on information that could only have come from the news media,
we are confident that an increase in "news reception" indicates that media use
did occur. News reception tests the acquisition of knowledge, and in turn
captures, by implication, the motivational component of active information
seeking. The present study certainly enhances the construct validity of this
measure. If media use had only been measured in this research by self-reported
exposure (the standard in many studies), our results would have been "anemic and
understated," as Price and Zaller (1993) predict. Still, the self-report
measures were not worthless by any means. They were almost all affected to some
extent, not significantly, but in the predicted direction. And they were
sufficiently sensitive to detect the expected difference between information
seeking via newspapers versus other channels. Self-report continues to be
necessary for this kind of distinction, since it would be difficult to measure
news reception that is specific to each channel.
Future Research. This experiment supports a causal model linking situational
activity, political involvement, and information seeking and acquisition. But
what particular aspect of the social situation stimulated involvement? Did peer
pressure cause the students to become more attentive to politics and to seek
information so they would appear well informed to their friends? Did subjects
suspect that another political discussion might take place in class (although
they were told otherwise), and did this expectation lead to information seeking
so they could participate the next time? Or, was involvement in politics --
regardless of peer pressure or social expectation -- a condition that followed
them out of the classroom situation? Future studies may be able to address these
questions about the nature of the phenomenon we have documented here.
Future research may also investigate whether other types of discussion
situations can stimulate political involvement -- discussions that are less
structured or not part of a classroom situation. For now, we can suggest that
political involvement is not based on stable personal characteristics alone, but
also on factors that are subject to situational change. Political involvement is
more than an abiding interest in politics. When it is stimulated by a situation,
involvement may describe a "process of activation" that can lead to
communication outcomes such as media use and political learning.
CONSTRUCTION OF NEWS RECEPTION INDICES
(1) Survey 1. One point each was scored for answering the correct candidate
(Kathleen Brown, Dianne Feinstein, Michael Huffington, Pete Wilson) for the
(a) "Who endorses a law to end illegal immigrants' access to public school and
government-financed health care?" (answer: Wilson)
(b) "Who wants to prohibit different prices for men and women for services such
as dry cleaning and haircuts?" (answer: Brown)
(c) "Who says the Endangered Species Act favors animals' needs over those of
people?" (answer: Huffington)
(d) "Who opposed the North American Free Trade Agreement (NAFTA)?" (answer:
(e) "Who is self-described as "President Clinton's No. 1 cheerleader" in
California?" (answer: Brown)
(f) "Who offered to debate only on a PBS TV channel in Sacramento?" (answer:
(g) "Who is accused in TV ads of mismanaging the state budget?" (answer: Wilson)
(h) "Who is accused in TV ads of being out of step with public sentiment against
illegal immigrants?" (answer: Brown)
(i) "Who is accused in TV ads of approving federal judges who are advocates of
criminal rights." (Feinstein)
(j) "Who is accused in TV ads of being called "secretive, threatening and
greedy" in "government documents"?" (Huffington)
(k) "Who is a former mayor of San Diego?" (Wilson)
(l) "Who is the current state Treasurer of California?" (Brown)
(m) "Who gave an entire year's salary to charity?" (Huffington)
(n) "Who was defeated in a previous campaign for Governor of California?"
(2) Survey 2. One point each was scored for answering the correct candidate
(Kathleen Brown, Diane Feinstein, Michael Huffington, Pete Wilson) for the
(a) "Who sponsored a bill to create national parks in the California desert?"
(b) "Who advocated replacing the welfare system with charitable contributions?"
(c) "Who suggested that all Americans should be required to show identification
cards for government services?" (answer: Wilson)
(d) "Who supported continuation of the CLAS system of standardized essay tests
in California schools?" (answer: Brown)
(e) "Who took out a second mortgage on the family home to buy more TV ads?"
(f) "Who was the subject of recent Doonesbury cartoons?" (answer: Huffington)
(g) "Who was accused in TV ads of opposing the death penalty "even for drive-by
killings"?" (answer: Brown)
(h) "Who was accused in TV ads of being a "Texas millionaire voters can't
trust"?" (answer: Huffington)
(i) "Who was accused in TV ads of being "a career politician" who rides to work
in a limousine?" (answer: Feinstein)
(j) "Who was accused in TV ads of supporting tax breaks for the rich, at the
expense of students, seniors and the middle class?" (answer: Wilson)
(k) "Who spoke at Richard Nixon's funeral?" (answer: Wilson)
(l) "Who is a former mayor of San Francisco?" (answer: Feinstein)
(m) "Who is married to an ordained minister?" (answer: Huffington)
(n) "Whose father and brother were Governors of California?" (answer: Brown)
(3) Survey 3. One point each was scored for the correct answer for the following
(a) "Which one of the following has opposed GATT?
High-tech firms; California agriculture; The Sierra Club; Less-developed
countries; Don't Know" (answer: The Sierra Club)
(b) "Who once said he would support GATT if there were a cut in the capital
Newt Gingrich; Richard Gephardt; Jesse Helms; Bob Dole; Don't Know" (answer: Bob
(c) "What has the United Nations decided to do with Bosnia?
Side with the Serbs; Take over; Side with the Muslims; Pull out "peace-keeping"
troops; Don't Know" (answer: Pull out troops)
(d) "Who is the next U.S. House of Representatives minority leader?
Newt Gingrich; Richard Gephardt; Jesse Helms; Bob Dole; Don't Know" (answer:
(e) "Who is the next U.S. Speaker of the House?
Newt Gingrich; Richard Gephardt; Jesse Helms; Bob Dole; Don't Know" (answer:
(f) "Who is the next majority leader of the U.S. Senate?"
Newt Gingrich; Richard Gephardt; Jesse Helms; Bob Dole; Don't Know" (answer: Bob
(g) "Who is the United Nations Secretary-General?
John Shalikashvili; Yasir Arafat; Boutros Boutros-Ghali; Jean Aristide; Don't
Know" (answer: Boutros Boutros-Ghali)
(h) "What is not part of the 'Contract with America'? Increases in defense
spending; Increases in federal food assistance programs; Cuts in the capital
gains tax; Increases in spending to build more prisons; Don't Know" (answer:
Increases in federal food assistance programs)
(i) "Jim Bakker...Is out of prison on parole; Has been pardoned; Was denied
parole; Was killed in a prison restroom; Don't Know" (answer: Is out of prison
(j) "Who is the new president of Mexico? Luis Donaldo Colosio; Ernesto Zedillo;
Silvio Berlusconi; Carlos Menem; Don't Know" (answer: Ernesto Zedillo)
(k) "Michael Huffington refuses to concede to Dianne Feinstein because he
alleges vote fraud by: Dianne Feinstein; Illegal immigrants; Organized labor;
The California Secretary of State; Don't Know" (answer: The California Secretary
(l) "Who said President Clinton was "unfit to be Commander in Chief" of the U.S.
Armed Forces? Newt Gingrich; Ross Perot; Jesse Helms; John Shalikashvili; Don't
Know" (answer: Jesse Helms)
Atkin, Charles K. (1972). Anticipated communication and mass media information
seeking. Public Opinion Quarterly, 36:188-189.
Atkin, Charles K. (1973). Instrumental utilities and information seeking. In
Peter Clarke (ed.), New Models for Mass Communication Research (pp. 205-242).
Beverly Hills CA: Sage.
Atkin, Charles K., John Galloway and Oguz B. Nayman (1976). News media exposure,
political knowledge and campaign interest. Journalism Quarterly, 53:231-237.
Barbato, Carole A. (1986). Uses of interpersonal communication. Paper presented
at the meeting of the Speech Communication Association, Chicago, September.
Berelson, Bernard, Paul F. Lazarsfeld, and William N. McPhee (1954). Voting: A
Study of Opinion Formation in a Presidential Campaign. Chicago: University of
Blumler, Jay G. and Elihu Katz (1974). The Uses of Mass Communications: Current
Perspectives on Gratifications Research. Beverly Hills, CA: Sage.
Bybee, Carl R. (1978). Testing a process model of involvement. Communication
Campbell, Angus, Philip E. Converse, Warren E. Miller and Donald E. Stokes.
(1960). The American Voter. New York: John Wiley and Sons.
Chaffee, Steven H. (1982). Mass media and interpersonal channels: Competitive,
convergent, or complementary? In Gary Gumpert and Robert Cathcart (eds.),
Inter/Media: Interpersonal communication in a media world (2nd ed.). New York:
Oxford University Press.
Chaffee, Steven H. and Jack M. McLeod. (1973). Individual vs. social predictors
of information seeking. Journalism Quarterly, 50:95-120.
Chaffee, Steven H., Youngme Moon and Michael McDevitt, with Zhongdang Pan, Jack
M. McLeod and William P. Eveland. (1995). Stimulation of communication:
Reconceptualizing the study of political socialization. Paper presented at the
annual meeting of the Association for Education in Journalism and Mass
Communication, Washington, D.C., August.
Chaffee, Steven H., Youngme Moon and Michael McDevitt. (1996). Immediate and
delayed effects of an intervention in political socialization: A
disequilibration- restabilization model. Paper presented at the annual meeting
of the International Communication Association, Chicago, May.
Chaffee, Steven H. and Joan Schleuder. (1986). Measurement and effects of
attention to media news. Human Communication Research, 13:76-107.
Chaffee, Steven H. and Seung-Mock Yang. (1990). Communication and political
socialization. In Orit Ichilov (ed.), Political Socialization, Citizenship
Education, and Democracy (pp. 137-157). New York: Columbia University Teachers
Chew, Fiona. (1994). The relationship of information needs to issue relevance
and media use. Journalism Quarterly, 71:676-688.
Cronbach, Lee J. (1951). Coefficient alpha and the internal structure of tests.
Chaiken, Shelley. (1980). Heuristic vs. systematic information processing and
the use of source vs. message cues in persuasion. Journal of Personality and
Social Psychology, 39:752-66.
Conway, M. Margaret. (1985). Political Participation in the United States.
Washington, D.C.: CQ Press.
Culbertson, Hugh and Guido H. Stempel III (1986). How media use and reliance
affect knowledge level. Communication Research, 13:579-602.
Cundy, Donald T. (1989). Televised political editorials andthe low-involvement
viewer. Social Science Quarterly, 70:911-922.
Donohew, Lewis and Leonard Tipton. (1973). A conceptual model of information
seeking, avoiding and processing. In Peter Clarke (ed.), New Models for Mass
Communication Research (pp. 243-268). Beverly Hills CA: Sage.
Erbe, William. (1964). Social involvement and political activity: a replication
and eleboration. American Sociological Review, 29:198-215.
Eulau, Heinz and Peter Schneider. (1956). Dimensions of political involvement.
Public Opinion Quarterly, 20:128-142.
Eulau, Heinz, J.W. Siegel and S. Weatherford (1979). Interpersonal contexts of
political behavior: Network analysis. (Technical report). Stanford, Calif.:
Stanford University Department of Political Science.
Faber, Ronald J., Albert R. Tims and Kay G. Schmitt. (1993). Negative political
advertising and voting intent: The role of involvement and alternative
information sources. Journal of Advertising, 22:67-76.
Gandy, Oscar H. Jr., Paula W. Matabane and John O. Omachonu (1987). Media use,
reliance and active participation: Exploring student awareness of the South
African conflict. Communication Research, 14:644-663.
Gantz, Walter, Michael Fitzmaurice and Ed Fink. (1991). Assessing the active
component of information-seeking. Journalism Quarterly, 68:630-637.
Grunig, James E. (1976). Communication behaviors occurring in decision and
nondecision situations. Journalism Quarterly, 53:252-263.
Katz, Elihu, Michael Gurevitch & Hadassah Haas. (1973). On the use of mass media
for important things. American Sociological Review, 38:164-181.
Katz, Elihu, Jay Blumler and Michael Gurevitch (1974). "Utilization of mass
communication by the individual" in Jay Blumler and Elihu Katz (eds.) The Uses
of Mass Communication. Beverly Hills, CA: Sage.
Kinsey, Dennis F. and Steven H. Chaffee. (1994). Communication behavior and
presidential approval: The decline of George Bush. Paper presented to the 1995
convention of the International Communication Association, Albuquerque, New
Krugman, Herbert E. (1965). The impact of television advertising: learning
without involvement. Public Opinion Quarterly, 29:349-356.
Lazarsfeld, Paul F., Bernard Berelson, and Hazel Gaudet. (1944). The People's
Choice. New York: Columbia University Press.
Lo, Ven-hwei. (1994). Media use, involvement and knowledge of the Gulf War.
Journalism Quarterly, 71:43-54.
Martinelli, Kathleen A. and Steven H. Chaffee. (1995). Measuring new-voter
learning via three channels of political information. Journalism and Mass
Communication Quarterly, 72:18-32.
McCombs, Maxwell E. (1972). Mass communication in political campaigns:
Information, gratification, and persuasion. In F. Gerald Kline and Phillip J.
Tichenor (eds.), Current Perspectives in Mass Communication Research (pp.
169-194). Beverly Hills, CA: Sage.
McLeod, Jack M. and Lee B. Becker. (1981). The uses and gratifications approach.
In Dan D. Nimmo and Keith R. Sanders, (eds.), Handbook of Political
Communication (pp. 67-100). Beverly Hills, CA: Sage.
McLeod, Jack M. and Daniel G. McDonald. (1985). Beyond simple exposure: media
orientations and their impact on political processes. Communication Research,
Milbrath, Lester W. and M.L. Goel. (1977). Political Participation: How and Why
do People Get Involved in Politics, Second Edition. Chicago: Rand McNally.
Morley, Donald Dean. (1984). Voter involvement and the potential for media
influence. Communication Research Reports, 1:134-140.
Olsen, Marvin E. (1972). Social participation and voting turnout: a multivariate
analysis. American Sociological Review, 37:317-333.
Pederson, Johannes T. (1978). Political involvement and partisan change in
presidential elections. American Journal of Political Science, 22:18-30.
Perloff, Richard M. (1985). Personal relevance and campaign information seeking:
A cognitive response-based approach. In Sidney Kraus and Richard M. Perloff
(eds.), Mass Media and Political Thought: An Information Processing Approach.
Beverly Hills, CA: Sage.
Petty, Richard E. and John T. Cacioppo. (1979). Issue involvement can increase
or decrease persuasion by enhancing message-relevant cognitive responses.
Journal of Personality and Social Psychology. 37:1915-26.
Petty, Richard E. and John T. Cacioppo. (1984). The effects of involvement on
responses to argument quantity and quality: central and peripheral routes to
persuasion. Journal of Personality and Social Psychology 46:69-81.
Pettey, Gary R. (1988). The interaction of the individual's social environment,
attention and interest, and public affairs media use on political knowledge
holding. Communication Research, 15:265-281.
Price, Vincent and John Zaller. (1993). Who gets the news? Alternative measures
of news reception and their implications for research. Public Opinion
Robinson, John P. and Mark R. Levy. (1986). Interpersonal communication and news
comprehension. Public Opinion Quarterly 50:160-175.
Roper, Burns W. (1983). Trends in attitudes toward television and other media: A
twenty-four year review. New York: The Roper Organization.
Roser, Connie. (1990). Involvement, attention and perceptions of message
relevance in the response to persuasive appeals. Communication Research 17(5):
Roser, Connie and Margaret Thompson. (1995). Fear appeals and the formation of
active publics. Journal of Communication, 45:103-121.
Rosengren, Karl-Erik. (1973). News diffusion: An overview. Journalism Quarterly
Ross, Lee and Richard E. Nisbett. (1991). The Person and the Situation. New
York: McGraw Hill.
Rothschild, Michael L. and Michael L. Ray. (1974). Involvement and political
advertising effect: an exploratory experiment. Communication Research 1(3):
Rubin, Rebecca B., Elizabeth M. Perse and Carole A. Barbato (1988).
Conceptualization and measurement of interpersonal communication motives. Human
Communication Research 14:602-628.
Salmon, Charles T. (1986). Perspectives on involvement in consumer and
communication research. Progress in Communication Sciences, 7:243-268.
Tan, Alexis S. (1980). Mass media use, issue knowledge and political
involvement. Public Opinion Quarterly, 44(2): 241-248.
Verba, Sidney and Norman H. Nie. (1972). Participation in America: Political
Democracy and Social Equality. New York: Harper and Row.
Differences Within Subjects When In Treatment and Control Conditions1
(Cell entries are standard scores indicating gain from pretest to posttest)
Treatment Control Difference t
t (s.d) c (s.d) t- c
News Reception .168 (.90) -.198 (1.03) .366 2.86**
Newspaper Attention .228 (.91) -.134 (1.00) .362 2.43**
Frequency .094 (.94) -.012 (1.05) .106 .70
TV News Exposure .055 (.96) -.048 (1.04) .103 .63
TV News Attention .079 (.96) .020 (.90) .059 .42
TV News Exposure .033 (1.11) -.010 (1.05) .043 .26
Newspaper Exposure .016 (.72) .026 (.77) -.010 -.10
1When in treatment condition, subjects discussed politics between pretest and
posttest; when in control condition, they did not. Each subject (N = 121)
participated in each condition but at different periods of time. Data are for
treatment minus control condition, regardless of time period. Entries are
presented in standardized scores.
*p<.05, **p<.01, ***p<.001, one-tailed
1 Situational political involvement refers specifically to social situations
that stimulate involvement. This is distinct from research that examines message
involvement (Petty and Cacioppo 1979, 1984; Chaiken 1980; Krugman 1965),
including studies that manipulate involvement with political messages (Roser and
Thompson 1995; Lo 1994; Cundy 1989; Bybee 1978).