Abstract
This study examines whether youth’s goals for their future societal engagement affect their future political activities. We hypothesized, first, that youth’s political interest would be a central predictor of their future political activities and, second, that their goals for future societal engagement would moderate their political interest. The greatest future political activity would be among youth who were both politically interested and had set high goals. A community sample of about 400 16-year-olds (Mage=16.62, SD=0.71, 50.8% females) was followed over four years. The theory of planned behavior was applied. Different types of political action were predicted: offline political activities, online political activities, membership of a political/civic organization, and taking a stance in public for a party pre and post voting in a national election. Age-16 political interest significantly predicted all age-20 political activities, except for being a member of political/civic organizations and actual voting. In line with the second hypothesis, significant interactions were found between political interest and setting societal goals for all the future political activities except voting. It was concluded that youth’s early-developed agency in setting goals for their societal engagement has widespread predictive utility for their political activities over a lengthy period.
Keywords
The political interest of youth is perhaps the most powerful predictor of a wide range of political variables (Prior, 2019). But studies also show that having a political interest per se may not be enough to activate young people politically. Youth may, for example, act according to what is expected of them, or because an action is personally important or momentarily beneficial to them, rather than being in line with their intrinsic political opinion (Koestner et al., 1996; Lopes et al., 2009). Also, political self-efficacy, relationships and attitudes towards others, trust in political institutions and democratic ideals, and existing political opportunities appear to matter (Dahl et al., 2018; Vromen, 2017; Oser & Hooghe, 2018). What is lacking is knowledge of the early motivational conditions that might turn youth’s political interest into political action. We examine such conditions here.
Why are some, but not all, youth willing to prepare themselves for future political activity? Perhaps they have a strong political interest and have already set goals for societal engagement. That the first antecedent of choosing to become politically and civically engaged in the future is to have a strong political interest is well supported in the literature (Prior, 2019). That there is a second antecedent, that of making plans for future societal engagement, is less well-established. If young people’s goals for their societal engagement affect their future political activities, it is a strong theoretical indication that youth play an active role in their own political socialization. That is, they envision their own future political activism. What the present study attempts to contribute to the literature is the idea that youth’s motivations for future societal engagement should not be seen separately from their political interests.
A General Theory of How Youth’s Goals Affect their Future Societal Engagement
The model that most explicitly expresses the idea that people’s goals determine their behaviors is in Fishbein and Ajzen’s theory of planned behavior (1975; Ajzen, 1985, 1991), which was later to emerge in the theory of reasoned action (Fishbein & Ajzen, 2010). This social psychological theory of why people behave as they do is broader than other theories that make behavioral predictions, is more general across ages, and extends across their performance goals (c.f., Locke & Latham, 2002). The fundamental idea underlying the theory is that the most proximal predictors of people’s behaviors lie in their intentions to perform particular actions. Given this study’s focus on predictions over lengthy periods, setting goals is a more adequate predictive label than having immediate intentions. In the theory of reasoned action, a person’s behavioral attitudes or beliefs about a behavior are presumed to affect his/her goals or intentions to perform a behavior. Also, the social norms of important persons and the person’s behavioral control are assumed to affect future behavior. In this study, we operationalize these theoretical aspects.
Transferring the theory of planned behavior into the political development domain, youth’s political interest can be said to represent the personal significances of politics to them, i.e., their political attitudes. In addition to high political interest, youth also need to have clear goals for their future societal engagement. Youth’s political interest can be hypothesized to make some of them committed to goals about their political future. Setting goals may act as a moderator of political interest, in that it will lead youth to future political activities, particularly if their political interest is associated with high goal setting. In short, high goal setting may enable the future societal engagement of politically interested youth.
Up to now, only one longitudinal study built on the theory of planned behavior has examined the role early intentions regarding societal engagement play in youth’s later political and civic activities (Eckstein et al., 2013). That study covered the main predictors in the theory in this context: youth’s attitudes to politics, their perceptions of the norms of important others, their perceived behavioral control, and, subsequently, their intentions to engage in politics. The model was tested over three consecutive years and, in line with the theory, youth’s behavioral intention to participate in politics was the main predictor of their actual political activities; further, this intention was predicted by their political attitudes and behavioral control.
Eckstein et al.’s study (2013) is an appropriate starting point for understanding youth’s political engagement in the future. But it might also mean revisiting the theoretical foundations of the theory of planned behavior as applied to youth’s political activities. The current study differs in two aspects from Eckstein et al.’s. First, in the earlier study, youth’s attitudes to politics were measured by items covering their preferred political activities rather than their factual attitudes (e.g., by asserting “Somebody who complains about political parties should join a party to change it”, and “There are not too many, but too few people politically active in Germany”). By contrast, we argue that youth’s political interest can be regarded as the most sensitive indicator of their attitudes to politics. Political interest has been described as a self-driven process, and it predicts a wide variety of political phenomena (Prior, 2019). Hence, we consider that the interest in politics that youth express is the most pertinent measure of their attitudes to political issues.
Second, in the Eckstein et al. study (2013), youth’s attitudes to politics were treated as a unique predictor of their intention to engage themselves politically. By contrast, it is proposed here that youth’s goals for their societal engagement do not operate independently of their political interest. Rather, we propose that setting societal goals acts as a moderator between their political interest and future political activities. Especially frequent political activities can be expected among youth who are high on both political interest and goals for future societal engagement. There are few reasons to expect differences in future political activities due to societal goal setting among youth who are politically uninterested. In short, political interest and setting goals for future engagement jointly should be more important for understanding youth’s future political attitudes than political interest and goals for future societal engagement separately.
In line with Eckstein et al. (2013), we use youth’s internal political efficacy, defined as young people’s beliefs in their capacity to participate effectively in politics (Craig et al., 1990), as a measure of perceived behavioral control, or perceived capability to perform political activities. Both youth’s political interest and political efficacy are strong predictors of political participation according to earlier studies (Levy & Akiva, 2019; Strömbäck & Shehata, 2010). Therefore, youth’s beliefs in their capacity to understand and participate effectively in politics (Craig et al., 1990; Miller et al., 1980) may be predictors of their future political activities that rival the goals they set for their future societal engagement. Finally, in line with our definition of youth’s attitudes to politics in terms of their political interest, we measured youth’s perceptions of important others’ social norms as their perceptions of the political interests of their parents and peers.
The Present Study
This study follows longitudinally the political activities of a sample of 16-year-old high-school (upper-secondary) students over four years. We focus on this age group because four years later, although they could still engage in about the same types of political activities as they could as 16-year-olds, there was one exception; they could vote. The year the 16-year-olds reached 20 was the year of a national election in Sweden, and they were eligible to vote for the first time. Before the election, the participants were given different options in a questionnaire to find out if they would take an active stance in public for a particular party. Subsequently, after the election, the participants were asked the same questions retrospectively, and also if they had voted. A key issue is whether these 16-year-old’s goals for societal engagement would be restricted to the political activities at age 20 that they were already familiar with at age 16 or would extend to voting and related activities four years later.
The current study is based on the theory of planned behavior and is modeled on the study by Eckstein et al. (2013). The issue is whether youth’s political interest, perceptions of the social norms of parents and peers, internal political efficacy, and goal setting at age 16 relate to their political and national-election-related activities four years later. We present two hypotheses. First, we hypothesize that youth’s political interest is the main predictor of future political activities. Note that, in most previous studies, political interest has been found to explain a wide variety of political issues and is a good predictor of engaging in political activities (Prior, 2019). Second, we hypothesize that youth’s goals for their societal engagement moderate the effects of political interest on their different future political activities. High political activity is expected among youth who are both politically interested and have set high goals for their future societal engagement, whereas high goal setting is not likely to affect the future political activities of youth who are not politically interested. This expectation differs from that of the theory of planned behavior, which treats people’s intentions to act as being caused by their attitudes or beliefs. Further, we examine the roles that youth’s behavioral control, defined as their internal political efficacy, and the perceived political interests of important others, play in explaining youth’s future political activities over and above their goals for future societal engagement.
It is unclear what predictive power individual and social characteristics like gender and immigrant status have for youth’s future political engagement. These characteristics were therefore included as active variables in the main analyses. Further, youth’s perceptions of economic resources in the family were used as a proxy for family socio-economic conditions.
Method
Sample
An age cohort of 16-year-old adolescents (Mage=16.62, SD=0.71, 50.8% females) was followed-up in a longitudinal project on the political socialization of youth (Amnå et al., 2009). The study was conducted in a mid-sized city in Sweden of about 130,000 inhabitants from 2010 to 2015. Here, we used information from a follow-up of these adolescents when they were 20 years old (T2). The target sample at age 16 (T1) was 1052. The analytic sample, with data on the measures in the study at T1, was 869 (83% of the target sample). Of the analytic sample, there were data from the age-20 collection, when the adolescents had left high (upper-secondary) school, for 463 persons (53%), and data at post-election for 385 (44%). The age-20 data collection was in April and May of the election year, while the post-election questionnaire was distributed in October. The national election took place in September.
The proportion of participants for whom both parents were born outside Scandinavia was 17%, similar to that in the national proportion. The city is like the country as a whole regarding its immigration rate, income, educational level, unemployment rate, and the percentage of people voting in the previous election (Statistics Sweden, 2010).
Procedure
The cohort was recruited from three high (upper-secondary) schools in the city, and strategically chosen to represent the social and demographic characteristics of its adolescents. There was an opt-out procedure, where parents of the youth in the two youngest cohorts were informed about the study prior to the first assessment and could opt their children out of the study. The assessments at age 16 took place in the participants’ classrooms during regular school hours. Trained test administrators distributed a self-report questionnaire to the participants with their teachers absent. At the first assessment, 1,000 SEK (about 120 USD) was contributed to the class fund, and, at subsequent assessments, the participants were compensated with movie tickets. At the follow-up at age 20, the participants were mailed a questionnaire together with a personalized link to an online version of the questionnaire. They received a 28 € gift card for their participation. One of the six local ethics review boards of the Swedish Ethics Review Authority approved all aspects of the procedure (Dnr 2010/115).
Measures
The study’s measures were developed in the research project (Amnå et al., 2009).
Political Interest
The level of political interest was measured using two questions. The first was: “How interested are you in politics?” (see ANES, 2014). The response options ranged from 1 (totally uninterested), to 5 (very interested). Second, the participants’ emotional reactions to politics were investigated by the question: “People differ in how they feel about politics. What are your feelings?” Responses were on a scale from (1) loath to (6) great fun. The feeling question was rescaled into a five-point rating scale and the two items aggregated. They had a correlation of .77, p < .001.
Setting Goals for Societal Engagement
Setting goals for societal engagement was a one-item measure (henceforth also called goal setting), “Some people have set goals for themselves concerning their engagement in societal issues. Have you?”, with the response options: (1) I’m not interested in issues like this, rather the opposite. I will certainly not commit myself to doing anything about them, (2) I’m not interested in issues like this and, as far as I can see, I will not commit myself to getting engaged in them, (3) I will definitely work actively (in an organization or by other means) on issues like this but have not yet committed myself, and (4) I will work actively in an organization, and I’m already a member of such an organization. Note that the last response option includes being a member of any such organization. Given that this is a sensitive measure, it should differentiate between youth who were already members of political/civic organizations at T1 and those who were not (see the definition of membership of a political/civic organization below
Social Norms
As a measure of social norms among important people, the youth responded to the statements: “My parents (friends) are interested in what is going on in the world” and “My parents (friends) keep in touch with the news”. The response scale ranged from 1 (doesn’t apply at all) to 4 (applies perfectly). The correlation between the two items about what happens in the world was .59, p<.001 for parents’ interest, and .53, p<.001, for peers’ interest.
Political Activities – Offline
The participants were asked if they, during the previous year, had been involved in ten political activities, such as “Taken part in a legal demonstration or strike”, and “Collected signatures”. The response scale ranged from (1) never to (3) several times. Cronbach’s alpha was .85 at T1 and .82 at T2.
Political Activities – Online
The participants were asked if they had engaged in fifteen different political activities online in the last 12 months, such as “Written about politics or societal issues on my own blog or website”, and “Taken part in an internet-based protest” (we excluded items depicting a search for political information on the internet). The response options ranged from 1 (no) to 3 (yes, several times). Cronbach’s alpha was .88 at T1 and .86 at T2.
Membership of a Political/Civic Organization
The participants were asked if they were a member (Yes, No) of any of three types of organizations: a political organization (4%), an environmental organization (1.4%), or an organization for peace or human rights (the Peace Movement, Amnesty International, etc.) (2.3%). Memberships of these organizations were summed.
The following questions were asked before (in April, May) and after (in October) the national election, which took place in September:
Taking a Stance in Public in the National Election
Before the election, the participants were asked if they would take an active stance in public in various ways: “Use the internet to support the parties I like”, “Follow and ‘like’ parties on social networks, such as Twitter and Facebook”, “Work pro-bono for a party in the national election, for example by handing out leaflets for the party”, and “Support a party by wearing their badge or t-shirt”. The response items were: 1 (I will definitely not do that), 2 (I will probably not do that), 3 (I will probably do that), and 4 (I will certainly do that). In the post-election questionnaire, the participants were asked again about taking an active stance in public. Here the stem question was “What did you do before the national election?” for the same items, and the response options were 1 (no) and 2 (yes). The alpha reliabilities before and after the election were .87 and .73. The correlation between taking a stance in public for a party before and after the election was .60, p < .001.
Finally, there was a question about whether the participants had voted in the election (Yes, No) in the post-election questionnaire.
Control Variables
We controlled for family economic conditions, immigrant status and gender. Economic conditions in the family were measured by five items, such as “If you want things that cost a lot of money (e.g., a computer, skateboard, cell phone), can your parents afford to buy them?” with responses scale from (1) absolutely not to (5) yes, absolutely, and “How often do you and your family go on vacation abroad?” with responses from (1) seldom or never to (4) we travel once or several times a year. Cronbach’s alpha at T1 was .82. Immigrant status was a dichotomous measure, differentiating between students both of whose parents were born outside the Nordic countries and other students. Gender was coded 1 = female, 2 = male.
Attrition Analyses
Logistic regression was used to compare the T1 participants who took part in the data collection at age 20 and those who did not on all the study variables at T1. Significant results at age 20 emerged for gender, indicating that males were less interested in taking part at age 20 (before the election: OR = 0.46, p < .001; after the election: OR = 0.43, p < .001). Nagelkerke R2 was .06 in both these attrition analyses. The significant odds ratios could be transformed into Cohen’s ds, where they were both 0.43, which is close to a medium effect size. Hence, we regard the age-20 and the post-election data as weakly gender-biased. Note that there was no bias for political interest, goal setting, political efficacy, or any of the other measures at age 16.
Further, when data were collected in high (upper-secondary) school, over 80% of participants responded each year. However, when sending questionnaires by mail to participants after they had left school, only 50% returned their questionnaire. We attribute attrition primarily to the type of data collection employed.
Plan for Analyses
To test our two hypotheses, we ran moderated regressions using Hayes’s PROCESS macro (Hayes, 2017), predicting youth’s political activities at age 20 from their political interest, goal setting and political efficacy, peers’ and parents’ political interest, and the control variables at age 16. The regressions also included the interaction between political interest and setting goals for societal engagement. All measures were standardized before the calculations.
Following customary longitudinal procedure, the three types of political activities at age 20 – political activities offline, political activities online, and membership of political/civic organizations – were controlled for by the measures at age 16 so as to focus attention on what happened between 16 and 20. Now, because the measure of goals included the option of already being politically engaged (response option 4: I will work actively in an organization and I’m already a member of such an organization), these predictions would, strictly speaking, include persons who were already members of such organizations at age 16. To take this into account, we controlled for age-16 membership of a political/civic organization when predicting age-20 offline and online political activities. The organizations concerned were among the three most common types of organizations mentioned by youth in the ages of the analytic sample in pilot tests.
Before proceeding, we performed a cross-sectional validation at age 16 of the measure of setting goals for societal engagement to examine if setting goals interacted with political interest or at least was a unique predictor of their preferred level of political activities at this age (important to. . . . “Be involved so that politicians get to know what people think”, “Actively try to influence societal issues” and “Make my voice heard on societal issues”). We predicted these items from all the study variables and the covariates. These results are reported in Supplementary Online Material Table S1 and show that both setting goals for societal engagement and political interest were significant predictors at the .001 level. A significant interaction was not found. Although this cross-sectional prediction does not support a moderation hypothesis concurrently, it shows that youth’s goals for societal engagement is a unique predictor over and above political interest.
In the Supplementary Online Material we report descriptive statistics for all the study variables at ages 16 and 20 (Table S2) and correlations between the measures at age 16 and at age 20 (Table S3).
Results
Correlations among the Study Variables at Age 16
Table 1 reports on the correlations among the study variables at age 16. Four features stand out. First, there were moderate to high correlations (.37 to .52) between political interest, setting goals for societal engagement, internal political efficacy, and online political activities. Second, there were lower correlations between these measures and membership of a political/civic organization. Third, youth both of whose parents were born abroad were slightly more engaged in online and offline political activities than youth with at least one parent born in Sweden. Finally, unexpectedly, good economic conditions in the family were more commonly perceived by males than females. Political online activities were also more common among males. Further, setting goals for societal engagement produced somewhat higher correlations with the three political activities at T1 (offline and online political activities, and being a member of a political/civic organization) than did youth’s internal political efficacy (average of .39 versus .27).
Correlations among the study variables at age 16.
Note. *p <.05, **p <.01, ***p <.001.
Predictions of Future Political Activities and Participation
Table 2 reports on predictions of offline and online activities, and membership of a political/civic organization, four years after the original data collection at age 16. The predictors were political interest, setting goals for societal engagement, the interaction between interest and goal setting, political efficacy, peers’ and parents’ political interest, and the control variables. All analyses controlled for the same political activity measures at age 16.
Predictions of changes in offline and online activities, and membership of a political or civic organization, from age 16 to age 20.
Note. CI = confidence interval [lower bound, upper bound].
The first hypothesis was that age-16 political interest is a significant predictor of the three age-20 political activities. As reported in Table 2, political interest was a significant predictor at the .05 level of offline and online political activities, but not of membership of a political/civic organization. The second hypothesis was that setting goals for societal engagement is a moderator of political interest. As shown in the table, the interaction between political interest and goal setting was significant for all three dependent measures. The moderating role of goal setting for political interest is reported in figures 1, 2, and 3. In all three cases, a high level of political activities was particularly evident for participants who had both high political interest and high goal setting. Follow-up simple slope tests indicated that political interest was a significant predictor of political activities at high levels (+1 SD) of goal setting (the coefficients ranged from .22 to .30), but not at low levels (- 1 SD) (the coefficients ranged from -.04 to .03). In sum, setting goals for future societal engagement moderated the effect of political interest for offline and online political activities, and for being a member of a political/civic organization. Setting goals as a main effect was a significant predictor of offline political activities and membership of a political/civic organization but not of online political activities. Political efficacy was a significant predictor at the .05 level only of offline political activities. Finally, peers’ and parents’ political interest, family economy, immigrant status, and gender had little predictive impact on the outcome measures.

Interactive effect of political interest and goal setting on offline activities at age 20.

Interactive effect of political interest and goal setting on online activities at age 20.

Interactive effect of political interest and goal setting on being a member of a political or civic organization at age 20.
Predictions of Activities in Connection with a National Election and Voting
Youth’s first national election is a unique event the experience of which cannot be predicted from earlier direct experiences. The same type of regression as earlier was applied to the predictors of publicly supporting a political party pre- and post-election. The results are shown in Table 3. Concerning the first hypothesis, political interest was a significant predictor of publicly supporting a party both before and after the election, and setting goals for one’s societal engagement was a significant predictor of taking a public stance for a party before the election. Concerning the second hypothesis, the interaction between political interest and goal setting was a significant predictor of both outcomes. Political efficacy did not predict publicly supporting a party before or after the election. Peers’ and parents’ political interest and the control variables were, with one exception, non-significant predictors. As reported in figures 4 and 5, high support in public particularly applied to participants who had both high political interest and high goal setting. Follow-up simple slope tests indicated that political interest was a good predictor of support in public of parties at high levels of goal setting, with coefficients of .37 and .38; significant at the .001 level, but not at low levels of goal setting, with coefficients as low as -.03 and -01. In sum, setting goals for societal engagement moderated the effect of political interest on support of a party in public, reported before and after the election. Finally, with 94% of participants reporting that they had voted, a logistic regression indicated that none of the predictors of voting were significant. Six percent of the variance was explained.
Predictions of support for parties in public at pre- and post-election at age 20.
Note. CI = confidence interval [lower bound, upper bound].

Interactive effect of political interest and goal setting on support for a party in public at pre-election.

Interactive effect of political interest and goal setting on support for a party in public at post-election.
Control Analyses
The highest scale value of the item about setting goals for societal engagement referenced both an aspiration for future societal engagement and already being a member of an association for this purpose; “I will work actively in an organization, and I’m already a member of such an organization.” Accordingly, in the multivariate analyses, when making predictions of future political activities, we controlled for the participants’ membership of political organizations, environmental organizations and organizations for peace or human rights. Now, it can be argued that these three types of organizations do not cover all the potential organizations and engagements of the participants at T1. Therefore, it was decided to exclude all youth who reported the highest scale value on the measure of setting goals for societal engagement and to reanalyze tables 2 and 3. If the findings that we had already reported are robust, we should obtain about the same results when reducing the four scales to three. We report on these analyses in tables S4 and S5 in the Supplementary Online Material. For three of the outcome measures, namely offline political activities, membership of a political or civic organization, and taking a stance publicly for a party reported before the national election, the interaction between political interest and goal setting was significant. However, it did not reach a conventional level of significance for online activities and taking a public stance reported after the election. This loss of significance was expected because removing the highest scale value of a particular measure will most likely reduce existing associations with other measures. Overall, we conclude that the findings held up reasonably well.
Discussion
That youth have rational plans for their societal engagement has been little discussed in the political socialization literature. But the present study, built on the theory of planned behavior (Ajzen, 1985, 1991; Fishbein & Ajzen, 1975), hypothesized that whether youth’s political interest predicts their future political activities depends on these youth’s goals for their future societal engagement. As predicted, a high level of future political activities is seen particularly among youth who are politically interested and have high goals for their future societal engagement. This idea of a moderating effect of societal goal setting on the relation between political interest and future activities received ample support. Generally, this suggests that youth are active agents in their political development. Some highly politically interested youth have long-term plans for their societal engagement by age 16, and they also show a high level of political activity four years later.
Using a sample of 16-year-old high (upper-secondary) school students, we examined three different types of political activities four years after recruitment and controlled for the same political activities at baseline: offline political activities, online political activities, and membership of a political/civic organization. At age 20, we found significant effects of age-16 political interest for offline and online political activities, but not for being a member of a political/civic organization. In line with the study hypothesis, we also found significant interaction between political interest and setting goals for societal engagement at age 16 for all three age-20 measures, showing that political activities were particularly common among youth who had both high political interest and high goal setting.
The year of follow-up was the first time the participants could vote in a national election. We measured whether they would take a stance in public for a political party. For taking a stance in public for a party, both before and after the election, we found significant effects of political interest and a significant interaction effect at age 16 between political interest and setting goals for societal engagement. Overall, in line with our hypothesis, setting goals for societal engagement seems to facilitate politically interested youth’s future political and civic engagement and taking a stance in public for a party. The presence of moderation appears consistent and robust and applies even after controlling for relevant social and individual characteristics.
The study made two major predictions. First, age-16 youth’s political interest was expected to significantly predict their political activities and willingness to take a stance in public for a political party four years later. This proved to be the case, with the exception of membership of a political/civic organization. Second, we predicted that youth’s goals for their societal engagement would moderate the influence of their political interest. A significant interaction between political interest and goal setting was found for all age-20 political activities and for taking a stance in public. The interactions confirm that the level of age-20 political activities was highest among the age-16 youth with high political interest and high goal setting. Concerning voting, we found no significant effect of political interest or any moderation effect with goal setting.
Taken together, the results seem to encourage and support various initiatives to channel youth’s political interest and long-term commitments during their politically formative years. It addresses how the politically oriented organizations of civil society, the political parties and their youth unions can give a voice to young people (Pickard, 2019). But it also applies to the public administration’s offering of forums that take advantage of and deliberately develop young people’s involvement in school, community planning, participatory budgeting, citizen dialogues and more (Nabatchi, 2014).
The study was built on a previous longitudinal study (Eckstein et al., 2013) about youth’s future political activities, which was founded in one of the best known theories of individual behavior, the theory of planned behavior (Ajzen, 1985; Fishbein & Ajzen, 1975). In this theory, peoples’ attitudes to a particular behavior are said to be directly linked to their behavioral intentions, which subsequently affect their actual behaviors. In this study, youth’s attitudes to politics were measured by their interest in politics, and their behavioral intentions by their long-term goals for engaging in societal issues. In contrast to the theory of planned behavior, which suggests that people’s attitudes are directly linked to their intentions, we proposed that young people’s goals for their societal engagement moderate their interest in politics, and that this interaction affects their future political attitudes. We found empirical support for this idea. Confirmations of interaction effects in psychological research are rare. Here, however, we found significant interactions between political interest and setting future societal goals for most of the measured political activities.
We advanced an alternative hypothesis that it might not be primarily the goals that youth set for their societal engagement that affect their future political activities but rather their internal political efficacy, i.e., their perceptions of their capacities to successfully affect political issues. This alternative hypothesis had little support. The longitudinal analyses showed that youth’s internal political efficacy was a unique predictor of only one of the six outcomes considered. This indicates that youth’s political efficacy may not have long-term consequences and does not predict future political activities. In fact, political efficacy was not even a strong concurrent predictor. Analyses not reported shows that the impact that political efficacy had concurrently and longitudinally largely disappeared when the participants’ goals for societal engagement entered the predictions. This insight is relevant to the wide literature built on internal political efficacy.
For proper interpretation of the findings, information is needed about how political interest and setting goals for societal engagement are associated with the other measures, and how they all relate, both cross-sectionally and longitudinally. First, youth’s political interest and their goal setting were correlated with political activities in high (upper-secondary) school to about the same extent (Table 1). Second, cross-sectional analyses showed that youth’s political interest and their goals for societal engagement at T1 were related, with a correlation of .52, p < .001. Third, the associations between political interest and goals setting and the age-20 outcomes were on the same level (between .22 and .42, p < .001; see Table S3). This indicates that youth’s goals for societal engagement do not live lives of their own, and do not necessarily operate longer-term, independently of youth’s general political interest.
Due to the skewed distribution, we were not successful in predicting actual voting. In the study, 94% of participants stated that they had voted in the national election. This is higher proportion than the 85% who, according to Statistics Sweden (2018), voted among youth aged 18 to 24 in the region. It is not uncommon to find a higher rate of reported than actual voting (Dahlgaard et al., 2019). An interpretation of the difference found here is that the over-reporting of first-time voting is particularly high because of a lack of routines among young voters (Gerber et al., 2003). Some of the youth might have felt pressurized to report in a socially desirable manner, and to report incorrectly in the questionnaire that they had voted.
A developmental perspective is needed to view progression in youth’s political activities. Over the age period covered, 16 to 20 years, the studied youth became more politically active (see Online Appendix Table S2). Political activities online and offline and being engaged in a political/civic organization were significantly more common among the 20-year-olds than when they were 16-year-olds. But there are also continuities over time. To be politically active in mid-adolescence, at the age of 16, whether it concerns offline or online activities, or to be engaged in a political/civic organization, is connected with being involved in the very same political activities four years later after the adolescents had finished upper-secondary school or had started to work or a university education. Over-time correlations are in the .30 range (see Online Appendix, Table S3). Perhaps what is most noteworthy is that being politically active on the internet at age 16 tends to be closely related to also being active on the internet four years later. Being a member of a political/civic organization at age 20 was three times more common among the youth who were engaged in these types of organizations in upper-secondary school (36.8%) as among those who were not members of such an organization (12.7%). Overall, there seems to be substantial continuities of being politically active by age 16 into early adulthood.
Finally, it should be added that we found only small predictive effects of the covariates over the four years. Moreover, peers’ and parents’ political interest, as measures of the social norms surrounding the adolescents, showed little predictive value.
Strengths and Limitations
The overriding strength of this study is that it empirically confirms the idea that young people rationally set goals for their future societal engagement, and that these goals aid politically interested youth in their future political activities. But there are questions left unanswered. We have leaned on central features of the theory of planned behavior (Ajzen, 1985; Fishbein & Ajzen, 1975), and have included political interest, behavioral control, social norms and setting future goals for one’s own societal engagement as factors to illuminate the hitherto understudied role that young people’s longer-term goals play in their societal engagement. But we have not attended to other features that ultimately need to be considered, such as youth’s perceptions of challenges to their goals, their beliefs concerning whether important persons approve or disapprove of their behaviors, and differences in the actual opportunities for carrying out political actions available to individuals with high goals for their engagement. In future studies, these factors need to be incorporated. In addition to individuals’ perceptions of the strategic opportunities for personal actions (i.e., for themselves), there is also a need to acknowledge that these opportunities depend on the more global structural factors that promote or hinder people’s political activism, as emphasized in opportunity structure theory (Meyer & Minkoff, 2004).
Further, the youth’s specific motivations for setting goals for their societal engagement in this study are not known. It is still an open question whether youth’s specific political motivations in mid-adolescence remain unstable (Ekström & Svenningsson, 2019) or have already crystallized (Schmid, 2012). Also, it is unknown to what extent these motivations reflect the civic motivations (i.e., persuade others, express a political view, learn new information, form an opinion) or the social motivations (i.e., talk, enjoy, relate) that have been associated with political discussion with others among adults (Zúñiga et al., 2016). Also, we do not know the extent to which youth’s goals in engaging in societal activities are based on the more stable moral and humanistic/ecological principles that have been previously described (Flanagan et al., 1998; Malin et al., 2015; Schmid, 2012; Wray-Lake & Syvertsen, 2011). Future research is needed to establish the specific political motivations that adolescents have, and their connections with later political activism.
A further limitation of the study is that the sample was found to be slightly gender biased. About the same proportions of males and females took part in the data collection at age 16, but the youth who answered the questionnaire four years later were under-representative of males. However, our attrition analysis did not find significant effects of any of the other measures used in the study.
Conclusions
Political interest is one of the most potent explanations for youth’s engagement in political activities. It was found that youth’s goal setting for their future societal engagement moderated the impact of political interest on their future political and civic activities. Highly politically interested youth with high goals for their societal engagement were particularly likely to activate themselves later on. These findings highlight the role that youth’s plans to engage have in their actual future political activities and indicate that youth are active agents in their political development. The message of the present study is that youth’s political interest increases their political activism if they also have clear goals for their societal engagement. Another theoretical contribution of the study is that it suggests that youth’s goals for their future societal engagement seem to be more important to attend to when examining both their adolescent and their early adult political engagement than their political efficacy.
Supplemental Material
sj-docx-1-jbd-10.1177_01650254221113465 – Supplemental material for Setting societal engagement goals during adolescence amplifies the impacts of political interest on political activities during young adulthood
Supplemental material, sj-docx-1-jbd-10.1177_01650254221113465 for Setting societal engagement goals during adolescence amplifies the impacts of political interest on political activities during young adulthood by Håkan Stattin, Erik Amnå and Silvia Russo in International Journal of Behavioral Development
Footnotes
Acknowledgements
This study was made possible by access to data from the Political Socialization Program, a longitudinal research program at YeS (Youth & Society) at Örebro University, Sweden.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The data collection was supported by grants from Riksbankens Jubileumsfond.
Supplemental material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
