The Effects of Age on the Interplay Between News Exposure, Political Discussion, and Political Knowledge
Abstract
Abstract. Using the theory of fluid-crystallized intelligence, we argue that with growing age, political discussion becomes less important as a complement to news exposure in political knowledge building. We applied moderated mediation analyses to the survey data of N = 69,125 German respondents. The data supported the hypothesis that news exposure influences political discussion, which in turn leverages political knowledge. As expected, we showed that news exposure is more strongly associated with political discussion for younger age groups. The results are discussed with regard to how to integrate a psychological lifespan perspective into further research on knowledge acquisition.
The acquisition of knowledge varies tremendously across the lifespan (Baltes, 1987; Beier & Ackerman, 2005; Salthouse, 2003). The ways in which differential forms of knowledge are acquired also vary with age (Ackerman, 2008). The theory of fluid-crystallized intelligence refers to age-related differences in knowledge: It implies that knowledge is generated and stored differently across the lifespan (Ackerman, 2008; Horn, 1982; Horn & Cattell, 1967). Crystallized intelligence is one of two main factors of the theory of fluid-crystallized intelligence. It refers to intelligence as cultural knowledge and represents a person’s experiences (Baltes, 1987). Crystallized abilities also imply a type of knowledge that is referred to as declarative knowledge, namely, the knowledge that is usually learned in school (e.g., European history as a part of political knowledge), the knowledge that we need to survive (e.g., how to behave during a thunderstorm), or the knowledge that is necessary to sustain well-being and physical health (Ackerman, 2008).
Crystallized abilities have been shown to increase linearly with age, and in previous research, a peak in crystallized intelligence was observed at age 60–70 (Rönnlund, Nyberg, & Bäckman, 2005; Schaie, 1996). However, positive effects of age on crystallized intelligence can also be found in younger age groups and in samples with smaller age ranges (Beauducel & Kersting, 2002; Beauducel, Liepmann, Felfe, & Nettelnstroth, 2007).
Political knowledge can be considered one domain of declarative knowledge and is significantly related to crystallized intelligence (Beauducel & Kersting, 2002). Political knowledge has been shown to be an important outcome variable because it is bound to participation and deliberation (Delli Carpini, 2004). The most important factor that influences political knowledge acquisition is exposure to news media in general and to print news in particular: Countless studies have shown that people who more frequently read the news score higher on tests of political knowledge (e.g., Yang & Grabe, 2011).
Another factor that influences knowledge acquisition is political discussion: Particularly when the news is repeated and collaboratively elaborated, individuals become able to learn from it (Eveland, 2004). For example, in a longitudinal panel survey, Eveland and Thomson (2006) found that the frequency of political discussion caused significant changes in political knowledge over the course of 1 year. Eveland and Hively (2009) conducted a telephone survey of 600 adults and revealed that frequency of discussion was positively related to knowledge about candidates’ stances on issues.
Previous studies have investigated the question of how political news reading and political discussion work their way into political knowledge. The relations between news exposure, discussion, and knowledge were suggested to be sequential such that people who are exposed to political news discuss the information they have read online or in the printed press, and by engaging in these two activities, they acquire knowledge (Eveland, 2004). For example, Nisbet and Scheufele (2004) showed that political discussion strengthens the effects of campaign exposure. In the 2000 American Election Study, they demonstrated that only those who were exposed to both online campaigns and political discussions benefitted in terms of political knowledge. These results suggest that political discussion is related to news exposure and provides a powerful means for generating political knowledge. In other words, individuals who tend to read the news more frequently also discuss politics more frequently and, in turn, have a higher chance to generate political knowledge. Therefore, we predicted the following:
Exposure to (a) print as well as (b) online news will positively influence political discussion, which, in turn, will positively influence political knowledge.
Different habits, capacities, and needs across the lifespan might influence whether and how news exposure and discussions leverage political knowledge. As stated above, knowledge acquisition and knowledge testing cannot be considered without referring to the differential effects of age. Knowledge acquisition varies tremendously across the lifespan and with age (Ackerman, 2008; Horn, 1982; Horn & Cattell, 1967). Here, we suggest that people from different age groups benefit differently from political discussions. In other words, we expected to find a moderated mediation with age moderating the relations between both (a) news exposure and political discussion and (b) political discussion and knowledge.
Thus, the interplay between news exposure, political discussion, and knowledge may be subdivided into two relations, both mediated by age: one referring to the relation between news exposure and political discussion and the other addressing the relation between political discussion and political knowledge.
With regard to the first relation between news exposure and political discussion, it seems likely that younger individuals – more than older persons – tend to combine news exposure with political discussions. According to Ackerman’s (2008) interpretation of the theory of fluid-crystallized intelligence, younger individuals actively seek interpersonal experiences (e.g., political discussion) to develop domain knowledge. Fluid abilities such as concentration, attention, and active elaboration are invested with the aim of converting these experiences into crystallized abilities (Salthouse, 2003). As many gaps in their cognitive skeletons need to be filled with relevant experiences and knowledge, young people in particular converse about political issues in order to make sense of the news. Then, at an older age with advancing cognitive development, the two activities – news exposure and political discussion – become increasingly detached. Knowledge structures become much more established (Ackerman, 2008). Moreover, with growing age, political schemas become developed further, and the need for communication decreases (Fiske & Kinder, 1986). A meta-analysis provided support for this notion by showing that age was negatively related to elaborative processing by discussing the news (Eveland, 2005). Therefore, we predicted the following:
Age will negatively moderate the effect of (a) print and (b) online news exposure on political discussion.
With regard to the second relation describing the interplay between political discussion and political knowledge, one might assume that political discussion positively influences political knowledge because it delivers additional factual information. However, referring to the fluid-crystallized theory of intelligence and previous research on declarative, domain knowledge (Ackerman, 2008), it seems plausible to argue that growth and decline may make this extra information less effective for knowledge building in older age groups (Baltes, 1987). With regard to growth, more experience produces less incremental knowledge (Smith & Baltes, 1990). With regard to decline, it has been shown that large amounts of novel learning that must occur quickly and that might not be controlled by the learner, such as in political discussions, work better for younger than for older individuals (Schaie, 1996). As age increases, political discussions may be important for exchanging attitudes and evaluations, but they are no longer important for crystallizing knowledge and schemas (Lau & Redlawsk, 2008). On the basis of this rationale, we predicted:
Age will negatively moderate the positive effect of political discussion on political knowledge.
Method
Procedure
The data used in this analysis originated from a broadly conceived online survey conducted in 2009 in Germany to investigate general knowledge.1 A total of 36.8% of the participants were recruited via links on a German social network site, which was the most prominent social network site at the time of the study. The other 63.2% of the participants were recruited via links on the German online news magazine spiegel.de, which was one of the top-ranked news magazines in Germany at the time of the study.
The online questionnaire comprised five different categories of knowledge, namely, political knowledge, history, economics, culture, and the natural sciences. After analyzing the items and scales with a pretest administered to N = 6,224 respondents, a total of 180 items were chosen for the main study. Four parallelized item sets were assessed for each of the five knowledge categories. Each respondent was randomly assigned to one of the parallel item sets that comprised questions from all of the knowledge categories. In sum, each participant answered nine questions on each of the five topics, adding up to a total of 45 questions per participant. In addition to being tested on their knowledge, participants were randomly assigned to fill out questionnaires that referred to their discussion habits with regard to one of the five categories: politics, history, economics, culture, or the natural sciences. In addition, demographics and exposure to print and online news were assessed.
In the study presented here, we will refer to political knowledge, political discussion, online and print news exposure, as well as sociodemographics (age, gender, educational level). The other knowledge categories described above were regarded elsewhere (Trepte & Verbeet, 2010).
Sample
For the present analysis, a subsample of n = 69,125 persons ranging from 18 to 70 years of age was chosen from the overall sample of N = 692,215. Participants were considered if they provided valid data on the following variables: print and online news exposure, political discussion, political knowledge, age, gender, and educational level. As explained above, only one fifth of the sample was asked to fill out questions on political discussion habits. Thus, the overall sample was reduced in this respect.
Previous literature was consulted to define valid age thresholds, which were supposed to include the span of adulthood from young adulthood to older adulthood. In line with studies on cognitive aging, the sample was restricted to the ages between 18 and 70 because at age 60, the peak of declarative knowledge is reached with slow attenuation in the 70s and 80s (Ackerman, 2008; Baltes, 1987). The mean age was 26.4 years (SD = 8.28), and 61.3% of respondents were male; 43% of the respondents had achieved at least a high school diploma. In terms of occupational status, the sample consisted of 41.0% college or university students, 32.2% full or part-time employees, 8.8% pupils, 7.2% participants in vocational training, and 10.4% retired, unemployed, or other; 0.4% of the data on occupational background was missing.
As participants were self-selected to participate in the survey online (social network site and online magazine), the sample was not representative of the overall German population. In terms of age and educational background, the sample was biased as we had a large proportion of younger participants due to the recruiting strategy. However, concerning educational level, the sample was very similar to the German online population (37% of German “onliners” have at least a high school diploma; AGOF, 2015).
Measures
Political Knowledge
The political knowledge items were generated with a Delphi method by 12 German journalists working in different fields of interest at the German news magazine Der Spiegel. The items covered political events, the political system, countries, capital cities, and political candidates. The procedure of generating the knowledge items and selecting them according to item and scale analyses is described in detail in Trepte and Verbeet (2010). Multiple-choice items (e.g., “What is the name of the current UN Secretary General?”; answer options: Ban Ki Moon, Kofi Annan, José Manuel Barroso, and Joseph Blatter) as well as free-choice items (e.g., provide a candidate’s name to go with a photo; no answer options were available) were used. A time limit of 30 s per question reduced the possibility that participants would use additional help in answering the questions. All answers were coded as 0 (= false or not answered) or 1 (= correctly answered). The political knowledge score ranged from zero to nine correct answers (M = 4.70, SD = 2.23).
Print News Exposure
Participants were asked how many of the 12 recently published issues of the three most read German news magazines (Der Spiegel, Focus, Stern) they had read in the last 3 months. Answer options ranged from 0 (= no issues) to 12 (= 12 issues). The score for print exposure represents the arithmetic mean of the exposure to all three magazines (M = 1.63, SD = 1.93, min = 0, max = 12).
Online News Exposure
Participants were asked how often they read news online from the same three German news magazines. Answer options for every magazine ranged from 0 (= never/almost never) to 6 (= daily/almost daily). The score for online news exposure was computed by using the arithmetic mean of these three items (M = 1.30, SD = 1.33, min = 0, max = 6).
Political Discussion
Political discussion was measured with an additive index of four items that addressed discussion frequency (e.g., “How often do you talk with your friends about politics in general?”) as well as discussion intensity (e.g., “How much information do you share in discussions about politics with your friends?”). Each of the items had to be rated on a 5-point scale with smaller values indicating fewer political discussions (less intensity) and higher scores indicating more political discussions (greater intensity; M = 3.10, SD = 0.93, min = 1, max = 5; α = .85).
Control Variables
As control variables, we included gender (0 = male, 1 = female) as well as education (0 = no high school diploma, 1 = high school diploma).
Results
In order to test the proposed hypotheses, we computed multivariate path analyses in the statistical environment R (R Core Team, 2015). The model was estimated with the package lavaan. As recommended by Hayes (2012), we mean-centered the predictor variables prior to the analyses and tested for the significance of indirect effects with bootstrapped confidence intervals with 1,000 samples (Hayes & Scharkow, 2013). We controlled for gender and education. To illustrate the conditional effects, we used the package ggplot2. The estimated model fit the data well, χ2(4) = 216.16, p < .001 (CFI = .99; TLI = .98; RMSEA = .03, 90% CI [.025, .032], SRMR = .006, see also Figure 1 and Table 1). The predictors included in the model were able to explain 21% of the variance in political discussion and 31% of the variance in political knowledge.
Hypothesis 1 indicated a positive indirect effect of (a) print and (b) online news exposure on political knowledge through political discussion. The path analysis revealed that print news exposure (b = .09, 95% CI [.088, .095], β = .19) as well as online news exposure (b = .16, 95% CI [.153, .164], β = .23) positively influenced political discussion. Political discussion, in turn, was positively related to political knowledge (b = .62, 95% CI [.605, .638], β = .26). We found significant indirect effects of print news exposure (b = .06, 95% CI [.054, .059], β = .05) and online news exposure (b = .10, 95% CI [.094, .103], β = .06) on political knowledge through political discussion. Hypotheses 1a and 1b were supported by the data. Individuals who read online or print news more intensely tended to discuss politics more and, in turn, possessed higher political knowledge. With this result, we replicated previous research that had shown that news exposure and political discussion sequentially influence political knowledge.
Hypothesis 2 proposed that age would moderate the relations between (a) print and (b) online news exposure and political discussion. For H2a, evidence of moderation by age was found in a significant negative interaction between print news exposure and age on political discussion (b = −.001, 95% CI [−.002, −.001], β = −.03). Figure 2 shows that with decreasing age, the magnitude of the coefficient of print news exposure on political discussion also decreased considerably. Hypothesis 2a was supported by the data.
For H2b, a significant negative interaction between online news exposure and age on political discussion was found (b = −.002, 95% CI [−.003, −.002], β = −.03). Figure 3 shows that with decreasing age, the magnitude of the coefficient of print news exposure on political discussion also decreased considerably. Hypothesis 2b was supported by the data.
With these results on how age was found to moderate the relation between news use and political discussion, we complement current research. Younger people tended to combine online or print news exposure with political discussions. By contrast, older news readers did not relate their reading of the news to political discussions as often.
Hypothesis 3 predicted that age would moderate the relation between the mediator variable political discussion and the outcome variable political knowledge. We found a significant negative conditional effect (b = −.001, 95% CI [−.002, −.001], β = −.001). Figure 4 shows that with decreasing age, the magnitude of the effect of political discussion on political knowledge decreased considerably. Hypothesis 3 was supported by the data. The older people grow, the less their political discussions affect their political knowledge.
To examine the influence of different levels of the moderator on the mediation, we calculated the conditional effects for one standard deviation above and below the arithmetic mean. To further depict how the interplay between news exposure, political discussion, and political knowledge develops with age, we estimated the indirect effects of different values of the moderator variable age (18, 25, 35, 45, 55, 65, 70). As can be seen in Table 1, the indirect effects were consistently positive but decreased with increasing age. The indirect effect of print news exposure through political discussion to political knowledge was nearly two times larger for 18-year-olds (β = .057) than for 45-year-olds (β = .033) and four times larger for 18-year-olds than for 70-year-olds (β = .013). Thus, the probability that an 18-year-old person will link print news exposure with political discussions to finally generate political knowledge is considerably higher than for a 45-year-old or a 70-year-old. The same applies for online news use. The indirect effect was nearly two times larger for 18-year-olds (β = .067) when compared with 55-year-olds (β = .033) and three times larger when compared with 70-year-olds (β = .020). The older people grow, the more they learn directly from news exposure and the less through engaging in political discussion.
In line with previous research, we found gender as well as educational level to be relevant covariates. People who are male or highly educated are more likely to have more intense and frequent discussions about politics (Table 1). In addition, they are more likely to know more about politics than less educated or female respondents.
Discussion
Supporting previous research, we found that news exposure influences political discussion, which, in turn, influences political knowledge. Further, our results demonstrate that this mediation is moderated by age. In particular, younger news users combine print and online news with political discussions to generate political knowledge. We suspect that younger news readers acquire knowledge from news more easily when they get the chance to discuss it with others because in a political discussion, they are more intensely exposed to political content and get the chance to complement their lack of experience in the field of politics (Eveland, 2004).
As predicted by the theory of fluid-crystallized intelligence and its current applications, not only knowledge and crystallized abilities but also learning experiences differ significantly across the lifespan (Salthouse, 2003; Schaie, 1996). For example, in our sample, 18-year-olds gave considerably more correct answers to knowledge questions when they reported reading and discussing political topics. The picture was very different for news readers in their 60s for whom previous news reading significantly affected their political knowledge, but for whom the sequence of news reading and political discussion was only marginally important when political knowledge was required.
Limitations
Of course the data impose some limitations on our findings. Large sample sizes such as the one applied here are known to inflate the Type I error rate and to increase the chances of achieving statistical significance (Mallinckrodt, Abraham, Wei, & Russell, 2006). However, a large sample can also be seen as a strength as it allows researchers to detect small effects that could not have been found with a smaller sample (Button et al., 2013; Cohen, 1992). On the one hand, some of the effects in our sample were small and should therefore be interpreted with caution. On the other hand, small effect sizes do not necessarily indicate that an effect does not exist or is not important.
Due to the cross-sectional nature of the survey, causality could not be measured, and our results could also represent differences between cohorts. Longitudinal studies will be needed to investigate causal effects and the differential evaluation of age versus cohort effects.
Another limitation of our study is that we did not measure important covariates such as motivation, political interest, or attention to the news. In previous research, these have been shown to be significantly related to political knowledge (Eveland, 2005).
Conclusion
Our research underscores the idea that it is important to apply a lifespan perspective to questions of learning and knowledge acquisition. In particular, we demonstrated for the first time that the previously assumed sequence of news use, political discussion, and political learning considerably applies to younger but not older members of the population.
Our research complements the current understanding of news learning. The development of political knowledge is a core task of governments and educational policy. To inspire political learning and to guarantee the sustainability of political knowledge, a psychological lifespan perspective seems crucial. Young adolescents with only marginal experience in political and media landscapes need political discussion as a means to learn the news.
1Respondents were offered a personalized knowledge test score sheet as an incentive. Comparisons of how incentives work in online studies have shown that personal feedback that is based on study results significantly increases response quantity (Marcus, Bosnjak, Lindner, Pilischenko, & Schütz, 2007; Singer & Ye, 2013). Response quality was checked with a number of tests that were published in Trepte & Verbeet (2010).
References
(2008).
Knowledge and cognitive aging . In F. I. M. CraikT. A. SalthouseEds., The handbook of aging and cognition (3rd ed., pp. 445–489). New York, NY: Psychology Press.. (2015). Bildungsstand der Internetnutzer im Vergleich zur Gesamtbevölkerung in Deutschland im Juli 2015
[Educational level of online users compared with the general German population in July 2015] , Hamburg, Germany: Statista GmbH. Retrieved from http://de.statista.com/statistik/daten/studie/72319/umfrage/internetnutzer-in-deutschland-nach-bildungsstand/(1987). Theoretical propositions of life-span developmental psychology: On the dynamics between growth and decline. Developmental Psychology, 23, 611–626. doi: 10.1037/0012-1649.23.5.611
(2002). Fluid and crystallized intelligence and the Berlin model of Intelligence Structure (BIS). European Journal of Psychological Assessment, 18, 97–112. doi: 10.1027/1015-5759.18.2.97
(2007). The impact of different measurement models for fluid and crystallized intelligence on the correlation with personality traits. European Journal of Psychological Assessment, 23, 71–78. doi: 10.1027/1015-5759.23.2.71
(2005). Age, ability and the role of prior knowledge on the acquisition of new domain knowledge. Psychology and Aging, 20, 341–355. doi: 10.1037/0882-7974.20.2.341
(2013). Power failure: Why small sample size undermines the reliability of neuroscience. Nature Reviews Neuroscience (pp. 1–12. (online first). doi: 10.1038/nrn3475
(1992). A power primer. Psychological Bulletin, 112(1), 155–159. doi: 10.1037/0033-2909.112.1.155
(2004).
Mediating democratic engagement: The impact of communications on citizens’ involvement in political and civic life . In L. L. KaidEd., Handbook of political communication research (pp. 395–429). Hillsdale, NJ: Erlbaum.(2004). The effect of political discussion in producing informed citizens: The role of information, motivation and elaboration. Political Communication, 21, 177–193. doi: 10.1080/10584600490443877
(2005).
Information processing strategies in mass communication research . In S. DunwoodyL. B. BeckerG. KosickiD. McLeodEds., The evolution of key mass communication concepts: Honoring Jack McLeod (1st ed., pp. 217–248). New York, NY: Hampton Press.(2009). Political discussion frequency, network size, and “heterogeneity” of discussion as predictors of political knowledge and participation. Journal of Communication, 59, 205–224. doi: 10.1111/j.1460-2466.2009.01412.x
(2006). Is it talking, thinking or both? A lagged dependent variable model of discussion effects on political knowledge. Journal of Communication, 56, 523–542. doi: 10.1111/j.1460-2466.2006.00299.x
(1986).
Involvement, expertise, and schema use: Evidence from political cognition . In N. CantorJ. F. KihlstromEds., Personality, cognition and social interaction (pp. 171–190). Hillsdale, NJ: Erlbaum.(2012). PROCESS: A versatile computational tool for observed variable mediation, moderation, and conditional process modeling, [White paper]. Retrieved from http://www.afhayes.com/public/process2012.pdf
(2013). The relative trustworthiness of inferential tests of the indirect effect in statistical mediation analysis: Does method really matter? Psychological Science, 24, 1918–1927. doi: 10.1177/0956797613480187
(1982).
The theory of fluid and crystallized intelligence in relation to concepts of cognitive psychology and aging in adulthood . In F. I. M. CraikS. TrehubEds., Aging and cognitive processes (pp. 847–870). New York, NY: Plenum Press.(1967). Age differences in fluid and crystallized intelligence. Acta Psychologica, 26, 107–129. doi: 10.1016/0001-6918(67)90011-X
(2008). Older but wiser? Effects of age on political cognition. Journal of Politics, 70, 168–185. doi: 10.1017/S0022381607080127
(2006). Advances in testing the statistical significance of mediation effects. Journal of Counseling Psychology, 53, 372–378. doi: 10.1037/0022-0167.53.3.372
(2007). Compensating for low topic interest and long surveys: A field experiment of nonresponse in surveys. Social Science Computer Review, 25, 372–383. doi: 10.1177/0894439307297606
(2004). Political talk as a catalyst for online citizenship. Journalism & Mass Communication Quarterly, 81, 877–896. doi: 10.1177/107769900408100410
. (2015). R: A Language and Environment for Statistical Computing [Computer Software]. Vienna, Austria: R Foundation for Statistical. Retrieved from https://www.R-project.org
(2005). Stability, growth, and decline in adult life span development of declarative memory: Cross-sectional and longitudinal data from a population-based study. Psychology and Aging, 20, 3–18. doi: 10.1037/0882-7974.20.1.3
(2003).
Interrelations of aging, knowledge and cognitive performance . In U. M. StaudingerU. LindenbergerEds., Understanding human development. Dialogues with lifespan psychology (pp. 267–287). Norwell, MA: Kluwer Academic.(1996). Intellectual development in adulthood: The Seattle longitudinal study. New York, NY: Cambridge University Press.
(2013). The use and effects of incentives in surveys. The ANNALS of the American Academy of Political and Social Science, 645, 112–141. doi: 10.1177/0002716212458082
(1990).
A life-span perspective on thinking and problem-solving . In M. SchwebelC. A. MaherN. S. FagleyEds., Promoting cognitive growth over the life span (pp. 47–69). Hillsdale, NJ: Erlbaum.(2010). Der Studentenpisa-Test 2009: Idee, Entwicklung, Validierung
[The Students-Pisa-Test: Idea, development, validation] . In S. TrepteM. VerbeetEds., Allgemeinbildung in Deutschland. Erkenntnisse aus dem SPIEGEL-Studentenpisa-Test (pp. 55–71) Wiesbaden: VS Verlag.(2011).
Knowledge acquisition gaps: A comparison of print versus online news sources . New Media & Society (13, pp. 1211–1227). doi: 10.1177/1461444811401708
