192
Petek Akman Özdemir
of political communication and 44.8% of these participants were expressing opinion on
social media several times a day.
Civic engagement scores were analyzed by socio-demographic variables. Independ-
ent samples t-tests showed a significant difference between gender groups, t(453.38) =
-2.23, p < .05. Men (X̅
= 11.03, SD = 5.04) reported higher engagement in civic activities
than women (X̅
= 10.09, SD = 4.57). Correlational analysis revealed no significant relation-
ship between age and civic engagement. ANOVA showed that civic engagement scores
differed significantly among education groups, F(2, 546) = 8.22, p < .001, ƞ² = .03. To
compare differences among groups Games Howell Test was used as it did not require
homogeneity of variances (Field, 2009). Low education group (X̅
= 12.17, SD = 6.05) had
higher civic engagement scores than high education group (X̅
= 9.95, SD = 4.21).
A series of independent samples t-tests were run to analyze civic engagement by the type
of mass media used to follow news. Newspaper was identified as the only media type
that had an effect on civic engagement [t(547) = -4.84, p < .001]. The participants who
read newspapers (X̅
= 11.57, SD= 4.92) had significantly higher civic engagement scores
than the participants who did not (X̅
= 9,62, SD = 4.50).
Media groups were combined in order to examine effects of using more than one type
of media. Two new groups were formed, “newspaper group” and “television or internet
group”. “Newspaper group” consisted of participants who only read newspapers and
followed news via television or the Internet as well as reading newspapers. “Television or
internet group” included only television, only the Internet or both television and internet
groups. The participants using all types of media formed the third group. The final group
included participants who were not interested in news. ANOVA revealed a significant dif-
ference among these four groups for civic engagement, F(3, 545) = 10.75, p < .001, ƞ² =
.06. Civic engagement scores of “newspaper group” (X̅
= 13.07, SD= 5.48) and the group
using all types of media (X̅
= 11.11, SD= 4.66) were found to be significantly higher than
the scores of “television or internet group” (X̅
= 9.75, SD= 4.39) as well as the scores of
the group not interested in news (X̅
= 9.11, SD= 4.89). No additional significant differenc-
es were found between groups.
Civic engagement, following news and social media were found to be positively correlat-
ed. Correlations among research variables ranged from .19 to .38. Hierarchical regression
analyses were conducted to assess predictive power of following news and social media
use on civic engagement. In the first step, socio-demographic variables were entered.
They significantly predicted civic engagement, R² = .03, F(3, 545) = 6.35, p < .001. Edu-
cation level (β = -.16, p < .001) was the only socio-demographic variable predicting civic
engagement. In the second step, following news and social media use were added. The
model was found to be significant, R² = .09, F(5, 543) = 14.63, p < .001. After controlling
for socio-demographic variables, following news (β = .11, p < .05) and media use (β = .24,
p < .001) positively predicted civic engagement and explained additional 9% of variance.
