Narrowing the Knowledge Gap

Abstract

Using a national survey of citizens during Singapore’s general election, this study finds that alternative online media filled information gaps and narrowed the knowledge gap between social status groups. Contrary to the original knowledge gap hypothesis, lower educated voters gained knowledge more rapidly from increased alternative media use than more highly educated voters. This indicates that the press system itself influences the knowledge gap. In authoritarian press systems, where mainstream news is regulated and framed to promote ruling party’s elite perspectives, alternative online information produced from the ground up embodies greater relevance and authenticity, facilitating greater knowledge gain among lower educated voters.

Keywords

knowledge gap authoritarian press elections Internet

An informed electorate is essential for democracy. Higher level of political knowledge is associated with greater political participation (Delli Carpini & Keeter, 1996). More knowledgeable voters can make decisions consistent with their interests and beliefs and better reap the benefits provided by the social system (Cho & McLeod, 2007; Delli Carpini & Keeter, 1996). Conversely, deprivation of knowledge may lead to relative deprivation of power. Ignorant voters will find it difficult to control public policy and may be manipulated by elites (Somin, 2004). By examining how knowledge is distributed and acquired, scholars hope to address how elites develop and maintain social power over lower social classes through differential acquisition of information (Donahue, Tichenor, & Olien, 1973).

Gaps in political knowledge occur between social classes in part due to the influence of the mass media (Tichenor, Donahue, & Olien, 1970). The knowledge gap hypothesis states that as mass media information increases, there will be an increasing gap between higher and lower social economic status (SES) segments of the population. The theory was formulated in the United States, where the intensity of coverage by diverse media is proportional to the degree of polarization and disputes between competing political groups (Hindman, 2012). The knowledge disparity results under intense media coverage as better educated segments of society are able to acquire information faster from a mass media that tends to cater selectively to their interests and needs. Compounding this problem is how the knowledge gap is more prevalent for in-depth and difficult topics such as politics (Gaziano, 1983). Studies find that education-based knowledge gaps remain the same or even widen over time with increased complexity of issues and increased media use (Holbrook, 2002; Kim, 2008; McCann & Lawson, 2006). Furthermore, education is a powerful resource that can undermine political propaganda. Better educated people are more media literate, are more active in seeking out information, and can draw on a greater store of knowledge and information sources to recognize self-serving political facts and counter propaganda (Bonfadelli, 2002; Johansen & Joslyn, 2008). They possess greater prior knowledge of topics, are better able to use and interpret media information, and have more relevant social contacts that provide additional information (Johansen & Joslyn, 2008).

Although research on the knowledge gap and political campaigns is extensive, it has an important limitation. This body of literature is based mainly on media consumers in the United States, with its democratic and liberal political and press systems. Even the few exceptions outside the United States are situated in democratic press systems (Kim, 2008; McCann & Lawson, 2006; Nisbet, 2008; Yang & Grabe, 2011). These studies, while providing support for the correlates of knowledge gap and education, do not allow us to understand how the knowledge gap will be influenced in a different media system. In an authoritarian regime, for example, political groups that overtly oppose the government and receive objective mainstream media coverage are unlikely. As the government owns a large share of media outlets and controls the media industry and news content, the media would toe the government’s line and present its case regularly to the public while vilifying dissenting groups. Citizens are more politically ignorant and apathetic, not knowing enough to monitor or reprove effectively politicians who may eschew accountability (Leeson, 2008). Information is less likely to be challenged, even when inaccurate, preventing knowledge gain (Johansen & Joslyn, 2008). Increased media coverage thus may not necessarily increase the knowledge gap as neither the educated nor the non-educated will gain knowledge.

Furthermore, existing research is based mainly on mainstream media use. Though researchers in the late 1990s had studied the influence of what was then considered “non-traditional media” on knowledge gain (Johnson, Braima, & Sothirajah, 1999), to date, there has yet to be a study on the influence of alternative media on knowledge gap. Research shows that contextual factors including community structure and types of media available and accessed by residents interacted with education to influence knowledge differentials (Gaziano, 2010; Tichenor, Rodenkirchen, Olien, & Donahue, 1973; Viswanath, Finnegan, Hertog, Pirie, & Murray, 1994; Viswanath, Kosicki, Fredin, & Park, 2000).

It is thus the purpose of this study to examine the knowledge gap hypothesis in the context of an authoritarian press system to determine whether knowledge disparities exist along socio-economic lines. It focuses specifically on the role of alternative news online in influencing knowledge gaps left by government-controlled media. Without mainstream media to offer diverse coverage of opposition politics, alternative media online may be voters’ only constant and readily accessible source of news, information, discussions, and opportunities for political engagement with opposition parties, their candidates, and their issues. Nonetheless, elites with more resources use the Internet more effectively to gain political knowledge (Bonfadelli, 2002). This raises the question on whether knowledge gaps between social classes will also result from accessing alternative political information online.

Singapore is an optimal case for examining how the Internet influences knowledge differentials in authoritarian press systems where political information is controlled in mainstream media outlets but alternative political information exists online. Although it is a democratic republic, mainstream news in Singapore remains regulated by the ruling political party and thus strongly supports government and party’s positions through greater and more positive coverage (Cenite, Chong, Han, Lim, & Tan, 2008). In 2014, Reporters Without Borders ranked Singapore 150 out of 180 on its World Press Freedom Index. Freedom House labeled Singapore’s press freedom as “not free,” placing the country at 153 on its Global Press Ranking.

Because Internet penetration and usage in Singapore is high, there is often little doubt that the digital divide in terms of physical access has almost been overcome, and that Singaporeans can easily engage in political activities online. Although the number of alternative political news sites and blogs has mushroomed in Singapore (Lee & Kan, 2009), little is known about Singaporeans’ political activities online and the effect of such use on their political knowledge. A national survey indicates that social status is a significant predictor of online political information consumption, with male, younger, and more educated voters more likely to go online (T. H. Tan, Mahizhnan, & Ang, 2011).

Understanding how different media contributes to voters’ knowledge gap on opposition parties, candidates, and issues is especially necessary in Singapore, where voting is compulsory. If people of higher social status continue to benefit from online access, this may further widen disparities between social groups, with lower status people at risk of further marginalization. Singapore’s 2011 general election was the first in its history where the Internet was liberalized to allow election campaigning. This resulted in a surge of alternative views on Singapore politics as resource-strapped opposition parties went online to campaign and citizens criticized government policies on social media. The context offers a rich setting for studying the relationship between alternative media and knowledge gap. This study thus examines whether a knowledge gap exists between voters who consumed news on government-regulated mainstream media during Singapore’s 2011 general election, and those who went online seeking alternative information about opposition parties.

Literature Review

The Internet and Knowledge Gap

In the present information society that is premised on the idea that power and control lie in the hands of those with knowledge, cyber-optimists hold high hopes that the Internet can foster democracy, empowerment, and equality by breaking barriers to political knowledge. This benefits citizens as multiple sources of information enhance their understanding of alternative electoral choices, their evaluation of those in authority, and enable them to make informed decisions about the possible consequences of their votes (Norris, 2001). The Internet facilitates political discussion and online activism, and enables rapid news and information dissemination to mass audiences at relatively lower cost and ease than traditional media, providing a platform for opposition movements and minor or fringe parties to challenge authoritarian regimes and level the political playing field (Norris, 2001). The Arab Spring exemplifies how social media successfully undermined and circumnavigated autocratic media control to disseminate information to citizens, thus shaping their decisions on the organization and participation of the protests (Tufekci & Wilson, 2012). In China, increasingly, bloggers and activists are using the Internet to disseminate information and promote civic discourse on social and political problems, albeit within limits delineated by the Chinese government (MacKinnon, 2008). Although many scholars have examined the democratic influence of the Internet in authoritarian regimes (Kalathil & Boas, 2003), none has examined its effect on political knowledge. There is also little research on how alternative media influence knowledge gain. This article can thus only draw from research on mainstream, democratic press to discuss the relationship between Internet use and knowledge gap.

Scholars have argued that online information is more accessible as it is more casual, colloquial, and accompanied with more photos (Graber, 1994). Studies comparing knowledge gain between print news and other media find that newspaper reading does not always correlate or predict greater knowledge (Jeffres, Atkin, & Fu, 2011; Jerit, Barabas, & Bolsen, 2006; Nisbet, 2008). After controlling for demographics and SES, these studies show that a variety of media including televisions news, campaign ads, televised debates, and local community websites influenced knowledge gain (Drew & Weaver, 1991, 2006; Holbrook, 2002; Jeffres et al., 2011; Norris & Sanders, 2003; Weaver & Drew, 1995, 2001). Television coverage may be more simplistic and less substantive compared with print but is more easily comprehended, making political information more accessible to the least educated or to low information voters, providing them the best opportunities to learn about political issues and gain ground on high information voters (Holbrook, 2002; Jerit et al., 2006). Print news either involves more framing and is more compactly written, or includes in-depth discussion with expert commentary, history, political, and social background. Both forms require higher cognitive skills associated with higher education while those with low education are least able to digest this type of information (Jerit et al., 2006).

Moreover, for ignorant publics, even marginal exposure to political information online will increase their political knowledge (Drew & Weaver, 2006; Norris & Sanders, 2003), suggesting that regardless of SES, knowledge gaps can be reduced with exposure to online political information. Furthermore, being more educated does not guarantee immunity to biased reporting. Scholars, however, remain ambivalent over the effectiveness of the Internet in promoting democracy. They worry of a democratic divide, where Internet politics would bypass the disengaged and engage those who are already engaged. Although the Internet makes it easier for citizens to learn about public affairs and be politically engaged, it does not automatically make them more inclined or able to do so (Norris, 2001). A number of studies that examined the influence of social status on information gaps across old and emerging media provide evidence that education is a determining factor for online information gaps. The higher educated gain more political knowledge than the lower educated from online sources such as news and political websites (Yang & Grabe, 2011). One recent study even found that online news failed to contribute to political learning (Tran, 2013). Several reasons explain this.

First, knowledge, economic resources, and computer and media literacy skills remain important prerequisites for effectively engaging digital content in an increasingly complex media environment (Servaes, 2003; Warschauer, 2002). Online content is heterogeneous and unlimited, and unlike old media that have journalists to filter and frame information, effective use of the Internet requires users to possess new skills like purposeful searching strategies, assessment of source credibility, and construction of meaningful frames for interpretations (Bonfadelli, 2002). People with lower income and education are those most likely to face barriers in gaining such digital literacy skills, forming a class of “new illiterates” (Servaes, 2003). Such inequalities in knowledge can further lead to exclusion of social resources and inequalities in social power, risking further marginalization of the underprivileged or apathetic (Bonfadelli, 2002).

Second, the knowledge gap has been found to be influenced by people’s motivations for acquiring information and their political beliefs (Bonfadelli, 2002; Ettema & Kline, 1977; Hindman, 2009, 2012). People acquire knowledge better when they have interest in the topic (Weaver & Drew, 2001). Such patterns of use could create a knowledge gap between online and print news users. Higher educated people are more likely to use the Internet for informational and service-oriented purposes, while lower educated users consume more entertainment and sports content than public affairs (Bonfadelli, 2002; Tewksbury, 2003; Wei & Hindman, 2011). Informational uses of the Internet are preferable as they increase users’ political knowledge and participation while entertainment use weakened them (Wei & Hindman, 2011).

Structural characteristics of the Internet compound the problem of disinterest. Unlike print media, where news producers provide editorial cues such as headlines and size of coverage to influence exposure to content, online media give equal prominence to a wide range of news stories (Tewksbury & Althaus, 2000; Yang & Grabe, 2011). As the Internet allows readers greater control over news selection, readers will pick news relevant to their interests. Moreover, readers are more likely to access news at the article level through hyperlinks from different sources, including news aggregators and social media, rather than from the homepage (Yang & Grabe, 2011). These factors may lower the chance of accidental exposure and incidental learning. Optimists hope that new, alternative sources of local political information and community websites might supplant traditional local news outlets, but the most popular sites tend not to be about community issues and often have no geographical basis that can contribute to local public affairs knowledge (Jeffres et al., 2011; Shaker, 2009). Studies that compared knowledge gain from reading print news and online news find that Internet users read less national, international, and political news, and that print news users still gained more knowledge, regardless of education level (Kim, 2008; Tewksbury & Althaus, 2000; Yang & Grabe, 2011).

Opposition Information Gap in Singapore

Singapore is a parliamentary republic with a single house parliament. The parliament has a term of 5 years, and general elections were historically held every 4 or 5 years. Voting is compulsory for citizens aged 21 years and older if their constituencies are being contested. The People’s Action Party (PAP) has been the ruling party since the country’s independence in 1965. In the 2011 general election, the PAP won 81 of 87 parliamentary seats. In Singapore, the Internet is an essential source for filling information gaps about opposition parties and views contrary to government policies. This is due to a highly regulated mainstream media known for its limited and often critical and negative coverage of opposition parties and their politicians (Gomez, 2008; Kluver, 2004). The mainstream media views itself as the government’s partner in progress and nation building, and is keenly in tune with the government’s agenda (Y. S. Tan & Soh, 1994). The bias against the opposition persists right up till the 2006 election (Cenite et al., 2008), and there has been no reason to expect the Singapore Government–mainstream media relationship to change for the 2011 general election.

Up till 2010, the government maintained stringent control on political content and discourse online. Individuals and groups engaging in political discussions must register with the authorities, and during elections, only political parties, candidates, and election agents could campaign online using a limited range of web tools (Shanmugam, 2011). Opposition party politicians also were wary of posting content online following site shut downs, defamation suits, and other legal run-ins with the government (Gomez, 2008). Such overt and subtle control of Internet content had a chilling effect, limiting political discourse on the Singapore blogosphere. Political bloggers who persist became ambivalent, carefully negotiating between their pursuit of free speech and exercising self-regulation and censorship (Lee & Kan, 2009).

In 2010, the government liberalized the Internet, eliminating the requirement for individuals to be registered to engage in online political activities. Under the new ruling, election candidates could also engage new media for campaigning including websites, social networking sites, social media apps, and text messaging on mobile phones, podcasts, and online videos. The new legislations brought optimism for greater dissemination of information from opposition parties and alternative political views online during the 2011 election, which saw the most number of seats contested since independence. Lower- and middle-income Singaporeans, faced with stagnant incomes, escalating costs of living, and widening income gaps, were increasingly frustrated with the PAP government. Labor policy promoting the influx of foreign workers in Singapore triggered heavy criticisms from the electorate. Nonetheless, Singapore’s mainstream media carried perspectives that were mostly consistent with the pro-foreign labor policies of the state. Political and thought leaders argued that foreign labor was critical to sustain a vibrant economy, which helped to create new business opportunities. These benefits collectively contributed to a higher standard of living. Online, a different picture emerged. In blogs, critiques and comments reflecting unhappiness with the influx of foreign labor abound as netizens broadcasted their views on overcrowding on public transport, inflated property prices, and competition for jobs (Pang & Goh, in press).

Opposition parties and their supporters too enhanced their campaigning efforts online, using their websites and social media sites to publicize their campaign, detail their positions on issues and arguments sidelined by the mainstream media, and engage voters in discussion of election issues (Goh & Pang, in press). The growth of alternative political content online in the past decade shows that the Internet may be serving as a public sphere for dissenting political views that are restricted offline. Singapore is well equipped for Internet politics. The country is among the most wired and computer-literate nations in the world. In 2010, 84% of Singaporean households have access to at least one computer at home (52% have two or more computers), and 82% have broadband Internet access at home (Infocomm Development Authority of Singapore, 2013). A survey found that use of the Internet for political information during the 2011 election period has also increased to 30% of electorate, compared with just 13% the previous year (T. H. Tan et al., 2011).

Nevertheless, voters who went online for political information tended to be male, younger, and more educated, and have higher household income. These mirror variables found to influence both the digital divide and knowledge gap, suggesting that political knowledge gaps will exist along these lines in Singapore. Furthermore, the mainstream media continue to be used by 86% of voters for election information (T. H. Tan et al., 2011). The heavy dependence on mainstream media is likely a spillover effect from a long history of control on political information consumption and discourse. In addition, discursive attempts by ruling party politicians and the mainstream media to vilify online political discourse cast doubts on its accuracy and reliability. The quality of online political information and level of participation also remained constrained and limited by opposition politicians and netizens’ overcaution, self-censorship, and lack of financial and human resources (Goh & Pang, in press; Gomez, 2008; Kluver, 2004; Lee & Kan, 2009).

Hypotheses and Research Question

The literature indicates that exposure to political information online may be effective in increasing political knowledge. This may be more likely when political content is limited in traditional media. Exposure to alternative news online may also mitigate ideological influence of authoritarian-controlled information. Although the mainstream media in Singapore do report about the opposition, coverage is limited and often critical. It can be expected that knowledge gain about opposition parties will be greater for alternative media users than mainstream media users. This prompts the first hypothesis as follows:

H1: Alternative media users will be more knowledgeable about opposition parties in Singapore than mainstream media users.

However, the literature also established that there are limitations to how much knowledge gain can result from Internet news consumption. Studies indicate that mainstream media still play a significant role in contributing knowledge. This may be particularly relevant in Singapore, where mainstream news is the main source of political information and has strong pro-PAP ideological influence that could impede opposition knowledge gain. There is likely an interaction effect, where knowledge gain is dependent on using both forms of media. However, the relationship between intensity of use of both media and knowledge is unclear. The following research question thus asks,

RQ1: What is the effect of different intensity of mainstream and alternative media use on knowledge about opposition parties?

There is evidence that differences in social status prevail online, playing a determining role in how likely people can acquire knowledge. It is likely that the Internet will widen knowledge gaps between users from different social classes, with the more highly educated voters gaining more knowledge than low-educated voters from alternative media use. The second hypothesis is proposed as follows:

H2: Education will moderate the effect of alternative media use on knowledge such that as alternative media use gets more intensive, knowledge gain will be much greater for highly educated voters than for low-educated voters.

Method

Sample

Data for this study were derived from a nationwide survey to determine media use, media attitudes, political knowledge, and political activities during Singapore’s 2011 General Election (T. H. Tan et al., 2011). A random sample of 2,000 Singapore citizens aged 21 years and above were interviewed using computer-assisted telephone interviewing between May 24 and July 17, 2011, by trained interviewers from Nielsen, a commercial research company. Administering the survey immediately after the election ensured exposure to election information occurred mostly during campaigning. To reduce respondent fatigue and satisficing behavior that would compromise data quality, the questionnaire was capped at 50 questions, with an average interview length of 35 min. The sample constituted a response rate of 19%. Quotas were set for race, age, gender, and education levels based on Singapore’s Census 2010. Majority of the calls were made on weekday evenings and nights, and throughout the day on weekends. Interviews were conducted in three of Singapore’s official languages—English, Mandarin, and Malay.

Measurements

Political knowledge of opposition parties

Knowledge gap researchers have measured knowledge in terms of respondents’ ability to identify facts, awareness of events or issues, or knowledge of arguments (Hwang & Jeong, 2009). Studies consistently find that factual knowledge is more strongly correlated with education than political beliefs about contentious issues (Hindman, 2009, 2012; Hwang & Jeong, 2009). Delli Carpini and Keeter (1993) argue that factual political knowledge is the best single indicator of political sophistication and identified a knowledge-measurement index comprising factual questions about politicians, their parties, and civic knowledge. As the knowledge gap hypothesis focuses on knowledge gain, researchers applying knowledge gap theory to elections focused specifically on audience learning of campaign issues and candidates’ positions, measuring learning of new political information that could only be acquired during campaigning (Jenssen, 2012; Johnson et al., 1999; Holbrook, 2002; Moore, 1987). These studies typically test respondents’ knowledge on campaign-specific information, including factual knowledge about candidates and their positions on issues (Drew & Weaver, 1991, 2006; Moore, 1987; Weaver & Drew 1995, 2001). Norris and Sanders also pointed out that in making voting decisions, people need practical knowledge, including prospective policy proposals of major political parties on election issues (Norris & Sanders, 2003). Such domain-specific measures are preferable as campaign information is more likely acquired through the media during the election period, compared with general political information, which may be derived from different sources at different times (Johnson et al., 1999). Domain-specific measures capture the impact of the information environment and indicate information gain during the course of campaigning (Jerit et al., 2006; Nadeau, Nevitte, Gidengil, & Blais, 2008). Moreover, Delli Carpini and Keeter’s analysis of knowledge indices indicate that measures of national political knowledge in one domain provide reasonably good measures of overall knowledge about national politics.

Following these earlier studies, this study measures factual knowledge about Singapore’s 2011 general election, focusing specifically on respondents’ knowledge about opposition party members and their position on issues. Focusing on opposition party knowledge allows for the measurement of the effects of differential coverage in mainstream and alternative media. Singapore law legislates that campaigning can only occur from the official date of nomination of candidates (April 27, 2011) to the day before polling day (May 6). Outside the election period, mainstream media coverage of opposition parties is almost non-existent. Opposition parties too were dormant between elections, with little publicity on members, issues, and activities. Thus, voters were more likely to encounter any information about opposition party candidates and their positions on issues only during the 10-day election period.

Two questions comprising 11 items asked respondents to identify opposition candidates according to their parties. To determine voters’ knowledge on opposition parties’ positions on issues, two questions focused on two campaign issues put up by the opposition parties (see appendix). Unlike a number of election issues raised by the opposition that failed to gain mainstream coverage, these two issues received more than average coverage in both mainstream and alternative online media and thus possess content validity.

The first issue was raised by Tan Jee Say, a candidate from the opposition Singapore Democratic Party, who proposed a restructuring of Singapore’s manufacturing sector. Tan received much attention during the election because he had previously held a high-ranking position in the government service as principal private secretary to the former prime minister. His credentials and proposal attracted much flak from PAP candidates, whose criticism of Tan were widely carried in the mainstream press. In response to such mainstream media treatment, Tan and his supporters posted his credentials and the full report of his proposal on alternative news sites and blogs (Loh, 2011). The second issue pertained to the opposition Workers’ Party’s (WP) manifesto, which received PAP criticisms that were also carried in the mainstream media. The WP’s defense of its manifesto was disseminated through its party website and alternative online news sites and blogs.

Each issue question asked respondents to choose from four statements the one that was the opposition party’s position. The four statements, developed from a thorough reading of mainstream and alternative news reports on the two issues and refined in consultation with political experts in Singapore, comprised three that presented PAP criticisms of the opposition and one based on the opposition parties’ declaration of their position on the issue.

For the first question, one point was awarded for correctly identifying each candidate, as well as for not selecting a wrong candidate. The total score for the first question is 10. One point each was awarded for correctly answering the second question and for selecting the opposition party’s position on the two issues. Measuring factual knowledge involves asking questions that can be verified with facts and not an opinion about reality (Hindman, 2012). Picking the opposition party’s statement will indicate that respondents had gained knowledge of the opposition parties’ position, regardless of their own stand on the issue.

The 13 items were combined into an index for political knowledge of opposition parties, with a total score of 13 points (Cronbach’s α = .87). The correlations of the knowledge index with measures of political behavior—political interest, r(1850) = .42, p < .01; participation in rally, r(1950) = .33, p < .01—are significant and consistent with theory and prior research, indicating that although the knowledge index is short, the items measured political knowledge with good reliability and validity. The mean score was 5.35 (SD = 3.80), indicating that overall knowledge of opposition is low. A total of 68.5% scored 7 points or lower. This means two thirds of respondents did not even get half the total points.

Control group

Demographics, political interest, and political rally participation were measured to control for their influence. Questions were asked to determine respondents’ gender, age, education level, and income. This study used education level as a measure of SES. Income was initially added to the regression model but was dropped due to high multicollinearity with education. The sample consisted of 52.8% female and 47.3% male. The mean age was 45 years old, with 16.7% in the 21 to 29 age group, 16.3% in 30 to 39, 24.8% in 40 to 49, 25% in 50 to 59, 11% in 60 to 69, and 6.4% who were over 70 years old. Of the 2,000 respondents, 19.9% had bachelor’s degree or higher, 17% had polytechnic diplomas, 3.9% had vocational certificates, 7.2% had junior college (A-level) certificates, 32.1% had secondary school certificates, 16.4% had primary school certificates, and 3.7% had no formal education.

As political interest and participating in political rally are factors found to influence political knowledge (Delli Carpini & Keeter, 1996), two questions measuring these (I am interested in political issues and I attended a political rally during the election) were also included in the model to control for their influence.

Mainstream media use

Respondents were asked to indicate, in 5-min intervals, how many minutes they spent daily reading newspapers (M = 24.7, SD = 35.6), listening to radio (M = 12.7, SD = 51.7), watching television (M = 32.2, SD = 50.7), and reading websites of mainstream media (M = 12.4, SD = 29.5) for election news. These items were summed to create an index (α = .60) measuring mainstream media use (M = 82.0 min, SD = 102.4). The means show the average time interval respondents spent on mainstream media (e.g., the average time spent on all mainstream media is between 80 and 85 min). There were 171 voters, or 8.9%, who did not use any mainstream media. The median time spent on mainstream media was 55 min, and the range was 0 to 1,050 min, resulting in a positively skewed distribution. After logarithmic transformation, the variable’s distribution approached normality (skewness = −1.183, SE = .056). The large percentage of non-users caused the distribution to be slightly negatively skewed.

Alternative online media use

One question asked respondents to indicate, in 5-min intervals, how many minutes they spent daily reading Singapore blogs and Internet-only news sites for election information (M = 6.3, SD = 19.6). There were 425 voters, or 21.4% of respondents who used alternative media. The large percentage of non-users (78.6%) resulted in a severely skewed positive distribution. An inverse transformation was thus applied to reduce skewness of the alternative media use variable (skewness = 1.371, SE = .056).

Interaction variables

Knowledge gaps are tested empirically as interactive effects between media use and education (Eveland & Scheufele, 2000; Nisbet, 2008). Education and the transformed mainstream media use and alternative media use variables were centered to avoid multicollinearity problems, and the product terms were computed to develop two interaction variables: Education × Mainstream media use and education × Alternative media use. The analysis of distributions indicated that the centered and interaction variables met the assumptions of normality.

Results

A two-way ANOVA was used to determine whether there are knowledge differences resulting from different levels of mainstream and alternative media use. As the purpose of the study was to determine whether there were differences in knowledge based on different intensity of media use, the media use variables were recoded into discrete variables. Recoding also addressed the negative skewness problem. The scale for mainstream media use was developed by first excluding non-users and finding the mean for the transformed mainstream media variable for only mainstream media users. Non-users were coded as 0, cases with less than or equal to the mean were recoded as low use (1), and cases greater than the mean were recoded as high use (2). Alternative media use was recoded into a dichotomous variable (0 = no use, 1 = use) due to the large percentage of non-users.

The ANOVA reveals a small but significant interaction between mainstream media use and alternative media use on knowledge, F(2, 1907) = 6.639, p = .001. This indicates the knowledge score will differ according to different levels of mainstream media and alternative media use. There are significant main effects from the two independent variables. The main effect of mainstream media is moderate, F(2, 1907) = 27.60, p < .001, while the main effect of alternative media is large, F(2, 1907) = 75.69, p < .001. Partial Eta Squared for the corrected model is .218, indicating that the two independent variables and the interaction effect contribute to 22% of the variability in knowledge scores.

Table 1 shows the mean knowledge scores and the differences in scores according to media use. Media non-users were least knowledgeable (M = 1.90). Alternative media non-users were also less knowledgeable, even with high mainstream media use (M = 5.96). Intensive use of both media produced the highest score (M = 8.14). The first hypothesis is thus supported.

Table 1.

Mean Knowledge Score for Mainstream and Alternative Media Users (N = 1,913).

Alternative media	Don’t use		Use		Score difference
Alternative media	M	SD	M	SD	Score difference
Mainstream media
Don’t use	1.90	2.60	7.00	3.28	5.10
Low use	4.06	3.30	7.44	3.55	3.38
High use	5.96	3.60	8.14	3.37	2.18

To determine the effect of different intensity of mainstream and alternative media use on knowledge, one-way ANOVAs and post hoc tests were conducted to determine differences in the simple main effects between cells for each media use level (e.g., differences in means between non-mainstream media users who use and do not use alternative media). The findings indicate that online alternative media use significantly increased knowledge about opposition parties for non-mainstream users, F(1, 169) = 31.90, p < .001; low mainstream users, F(1, 797) = 111.41, p < .001; and high mainstream media users, F(1, 941) = 75.31, p < .001. Different intensities of mainstream media use produce significantly different scores, F(2, 1491) = 114.96, p < .001, for non-users of alternative media but not for alternative media users, F(2, 416) = 2.22, p = .110. Nevertheless, knowledge difference became smaller as mainstream media use increased.

The regression provides some answers. To test whether the gaps resulted due to SES and the interaction with media use, knowledge score was regressed against control variables, education, media use, and the interaction terms between education and media use variables. Variables were entered in blocks based on their assumed causal priority (Kwak, 1999). Table 2 shows the control block, which comprised age, gender, political interest, and participation in political rally, accounting for 24.3% of the variance on the dependent variable (p < .001). Political interest made the strongest contribution (β = .35), followed by rally participation (β = .23). Gender and age were also significant predictors, with men and younger voters more likely to have higher scores. Education, when it was added to the model, became the strongest predictor (β = .31, p < .001) accounting for 6.5% of the R² change.

Table 2.

Summary of Hierarchical Regression Analysis for Variables Predicting Opposition Party Knowledge (N = 1,852).

Variables	Model 1			Model 2			Model 3
Variables	B	SE	β	B	SE	β	B	SE	β
Age	−0.013	.005	−.048*	0.027	.006	.099**	0.030	.006	.111**
Gender	−0.796	.156	−.105**	−0.763	.149	−.100**	−0.778	.143	−.102**
Political interest	1.261	.077	.345**	1.004	.076	.275**	0.703	.076	.193**
Rally attendance	2.072	.187	.233**	1.924	.179	.216**	1.570	.174	.176**
Education				0.787	.060	.309**	0.648	.059	.255**
Mainstream media use							1.223	.124	.203**
Alternative media use							1.696	.237	.173**
Mainstream × Education							0.094	.076	.024
Alternative × Education							−0.432	.147	−.066**
Adjusted R²		.243			.308			.365
Adjusted R² change		.244			.065			.059
F for change in adjusted R²		149.215			174.132			42.747

p < .05. **p < .001.

In the full model (R² = .365), education was the strongest predictor of knowledge, followed by mainstream media use, political interest, and rally participation. Alternative media use was the fifth strongest predictor, followed by age, in favor of seniority. Men were also more likely to be knowledgeable. The interaction effect between alternative media use and education was significant, but it was in the opposite direction as hypothesized. Education significantly moderated the relationship between alternative media use and knowledge (β = −.066, p = .003). The negative association indicates that the correlation between alternative media use and knowledge score was weaker among those with higher education. As Figure 1 shows, knowledge gain with increased alternative media use was greater for lower educated voters than highly educated voters, thus narrowing the knowledge gap. The effect of mainstream media use on knowledge did not depend on education.

Figure 1.

Knowledge gap between high and low education groups using alternative media.

Discussion

This study confirms that in an autocratic media system where political information in the mainstream media is controlled in favor of the government, alternative online media played an important role in helping voters gain political knowledge of opposition parties. Mainstream media only provides its users with limited information. Overall, alternative media users, including those who did not use mainstream media, were more knowledgeable about opposition parties. Furthermore, contrary to the original knowledge gap hypothesis, this study found that increased alternative media use enabled the lower educated to learn more about opposition parties, narrowing the knowledge gap. Studies have shown lower educated voters can catch up in political knowledge when information is more accessible and easily comprehended (Holbrook, 2002; Jerit et al., 2006). An analysis of Singapore political blogs showed that issues discussed online were relevant and authentic, and presented with colloquialism and contextualization (Pang & Goh, 2015). This might have resonated with lower education groups, resulting in larger knowledge gain.

However, the influence of alternative media was also limited. Overall, knowledge of opposition parties among voters was low. Although alternative media use did significantly increase political knowledge, the amount of knowledge gained decreased as mainstream media use increased. The diminishing influence of alternative media use on mainstream media users suggests that the Internet, while necessary in providing alternative political perspectives, likely still plays a secondary role to traditional media. In Singapore, mainstream media remain the main source of political information. This explains why the regression shows mainstream media as the second strongest predictor of knowledge. Intensive users of mainstream media were likely supporters of the government, and hence, they might have been resistant to alternative information online. Research shows that even with increased media coverage of scientific evidence asserting particular truths, ideology still served as a barrier against acceptance of the information, influencing how people interpret and recall information in ways congruent with their existing beliefs (Hindman, 2009, 2012).

There may also be confusion and uncertainty in accepting alternative information online due to voluminous pro-PAP news and limited but negative coverage of the opposition parties in the mainstream media. The PAP also repeatedly disputed alternative assertions online as untruths. The distrust in online content is compounded by the lack of quality online political discourse. Very few alternative political sites are maintained by people with professional journalism skills or political astuteness. Opposition parties’ online campaigning efforts were also limited due to resource constraints. Hence, these sites were easy and frequent targets of criticisms and even defamation suits by the PAP.

The rudimentary nature of online political information also likely explains why the lower educated could gain knowledge, while the higher educated did not learn much more. Online political discussions are often just that—discussions and opinions on what the mainstream news reported about government policies or comments by government officials. Knowledge gain is contingent on prior knowledge as better informed citizens are more aware of election issues and hence better prepared to understand them when they appear in the media. However, with the dearth of reporting on opposition issues in the mainstream media, going online may only be useful in providing basic and not sophisticated political knowledge. This may mean that when online political content becomes more sophisticated, it may be very likely that the traditional knowledge gap will manifest as the higher educated can better process and engage in the discourse.

Before further discussion, this study’s limitations should be noted. First, the measures of knowledge were based on political knowledge, with only two issues selected for the study. Although these issues were the most debated, they were still just two among many election issues. Expanding the issues to include more elections and general knowledge topics will enhance the significance of this study in determining whether alternative media use can narrow the general knowledge gap. This is an area that warrants further research. Second, only 30% of Singaporeans used online alternative information during the election. The findings can thus only be interpreted for this group of voters, who tended to be younger, have higher income, and are more educated (T. H. Tan et al., 2011).

Conclusion

With these shortcomings in mind, the findings in this study indicate that political and media systems do influence the knowledge gap differently. In Singapore, where political information is highly regulated, the Internet provided a means for citizens, particularly those with lower education, to gain political knowledge. Higher education did not widen the knowledge gap with increased media information. Contrary to criticism of online political content as inaccurate, this study shows that the Internet had been a reliable and effective, albeit limited, source in helping citizens become more knowledgeable about alternative parties, their candidates, and positions on policies and issues.

Footnotes

Appendix

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by the Institute of Policy Studies, National University of Singapore.

Author Biography

Debbie Goh is an assistant professor at the Wee Kim Wee School of Communication and Information, Nanyang Technological University. She received her PhD in mass communication from Indiana University Bloomington. Her research focuses on digital inequalities, media framing, and the processes that influence how marginalized communities engage with new media technologies.

References

Bonfadelli

(2002). The Internet and knowledge gaps: A theoretical and empirical investigation. European Journal of Communication, 17, 65-84.

Cenite

Chong

S. Y.

Han

T. J.

Lim

L. Q.

Tan

X. L.

(2008). Perpetual development journalism? Balance and framing in the 2006 Singapore election coverage. Asian Journal of Communication, 18, 280-295.

Cho

McLeod

D. M.

(2007). Structural antecedents to knowledge and participation: Extending the knowledge gap concept to participation. Journal of Communication, 57, 205-228.

Delli Carpini

M. X.

Keeter

(1993). Measuring political knowledge: Putting first things first. American Journal of Political Science, 37, 1179-1206.

Delli Carpini

M. X.

Keeter

(1996). What Americans know about politics and why it matters. New Haven, CT: Yale University Press.

Donahue

G. A.

Tichenor

P. J.

Olien

C. N.

(1973). Mass media functions, knowledge and social control. Journalism Quarterly, 50, 652-659.

Drew

Weaver

(1991). Voter learning in the 1988 presidential election: Did the debates and the media matter? Journalism Quarterly, 68, 27-37.

Drew

Weaver

(2006). Voter learning in the 2004 presidential election: Did the media matter? Journalism & Mass Communication Quarterly, 83, 24-42.

Ettema

J. S.

Kline

G. F.

(1977). Deficits, differences, and ceilings: Contingent conditions for understanding the knowledge gap. Communication Research, 4, 179-202.

10.

Eveland

W. P.

Scheufele

D. A.

(2000). Connecting news media use with gaps in knowledge and participation. Political Communication, 17, 215-237.

11.

Gaziano

(1983). The knowledge gap: An analytical review of media effects. Communication Research, 10, 447-486.

12.

Gaziano

(2010). Notes on “revisiting the knowledge gap hypothesis: A meta-analysis of thirty-five years of research.” Journalism & Mass Communication Quarterly, 87, 615-632.

13.

Goh

Pang

(in press). Untapped potential: Internet use by political parties. In Mahizhnan

Tan

T. H.

Ang

P. H.

(Eds.), Battle for hearts and mind: The impact of new media on Singapore General Election 2011. Singapore: World Scientific Publishing.

14.

Gomez

(2008). Online opposition in Singapore: Communications outreach without electoral gain. Journal of Contemporary Asia, 38, 591-612.

15.

Graber

(1994). Why voters fail information tests: Can the hurdles be overcome? Political Communication, 11, 331-346.

16.

Hindman

D. B.

(2009). Mass media flow and differential distribution of politically disputed beliefs: The belief gap hypothesis. Journalism & Mass Communication Quarterly, 86, 790-808.

17.

Hindman

D. B.

(2012). Knowledge gaps, belief gaps, and public opinion about health care reform. Journalism & Mass Communication Quarterly, 89, 585-605.

18.

Holbrook

T. M.

(2002). Presidential campaigns and the knowledge gap. Political Com-munication, 19, 437-454.

19.

Hwang

Jeong

(2009). Knowledge gap hypothesis: A meta-analysis of thirty-five years of research. Journalism & Mass Communication Quarterly, 86, 513-532.

20.

Infocomm Development Authority of Singapore. (2013). Annual survey on infocomm usage in households for 2010. Retrieved from http://www.ida.gov.sg/~/media/Files/Infocomm%20Landscape/Facts%20and%20Figures/SurveyReport/2011/ASInfocommIndustry11pdf.pdf

21.

Jeffres

L. W.

Atkin

(2011). Knowledge and the knowledge gap: Time to reconceptualize the “content.” Open Communication Journal, 5, 30-37.

22.

Jenssen

A. T.

(2012). Widening or closing the knowledge gap? The role of TV and newspapers in changing the distribution of political knowledge. Nordicom Review, 33, 19-36.

23.

Jerit

Barabas

Bolsen

(2006). Citizens, knowledge, and the information environment. American Journal of Political Science, 50, 266-282. doi:10.1111/j.1540-5907.2006.00183.x

24.

Johansen

M. S.

Joslyn

M. R.

(2008). Political persuasion during times of crisis: The effects of education and news media on citizen’s factual information about Iraq. Journalism & Mass Communication Quarterly, 85, 591-608.

25.

Johnson

T. J.

Braima

M. A. M.

Sothirajah

(1999). Doing the traditional media sidestep: Comparing the effects of the Internet and other nontraditional media in the 1996 presidential campaign. Journalism & Mass Communication Quarterly, 76, 99-123.

26.

Kalathil

Boas

(2003). Open networks, closed regimes: The impact of the internet on authoritarian rule. Washington, DC: Carnegie Endowment for International Peace.

27.

Kim

(2008). Testing the knowledge gap hypothesis in South Korea: Traditional news media, the Internet, and political learning. International Journal of Public Opinion Research, 20, 193-210.

28.

Kluver

(2004). Political culture and information technology in the 2001 Singapore General Election. Political Communication, 21, 435-458.

29.

Kwak

(1999). Revisiting the knowledge gap hypothesis. Communication Research, 26, 385-413.

30.

Lee

Kan

(2009). Blogospheric pressures in Singapore: Internet discourses and the 2006 general election. Continuum: Journal of Media & Cultural Studies, 23, 871-886.

31.

Leeson

P. T.

(2008). Media freedom, political knowledge, and participation. Journal of Economic Perspectives, 22, 155-169.

32.

Loh

(2011, April 24). Tan Jee Say’s paper on new economy. The Online Citizen Blog. Retrieved from http://www.theonlinecitizen.com/2011/04/tan-jee-says-paper-on-new-economy-reproduced-here-in-full/

33.

MacKinnon

(2008). Flatter world and thicker walls? Blogs, censorship and civic discourse in China. Public Choice, 134, 31-46.

34.

McCann

J. A.

Lawson

(2006). Presidential campaigns and the knowledge gap in three transitional democracies. Political Research Quarterly, 59, 13-22.

35.

Moore

D. W.

(1987). Political campaigns and the knowledge-gap hypothesis. Public Opinion Quarterly, 51, 186-200.

36.

Nadeau

Nevitte

Gidengil

Blais

(2008). Election campaigns as information campaigns: Who learns what and does it matter? Political Communication, 25, 229-248.

37.

Nisbet

(2008). Media use, democratic citizenship, and communication gaps in a developing democracy. International Journal of Public Opinion Research, 20, 454-482.

38.

Norris

(2001). Digital divide: Civic engagement, information poverty, and the Internet worldwide. Cambridge, UK: Cambridge University Press.

39.

Norris

Sanders

(2003). Message or medium? Campaign learning during the 2001 British general election. Political Communication, 20, 233-262.

40.

Pang

Goh

(2015). Pro, anti, neutral: Political blogs and their sentiments. In Mahizhnan

Tan

T. H.

Ang

P. H.

(Eds.), Battle for hearts and mind: The impact of new media on Singapore General Election 2011. Singapore: World Scientific Publishing.

41.

Servaes

(2003). Digital citizenship and information inequalities: Challenges for the future. In Servaes

(Ed.), The European information society: A reality check (pp. 231-238). Bristol, UK: ECCR Book Series.

42.

Shaker

(2009). Citizen’s local political knowledge and the role of media access. Journalism & Mass Communication Quarterly, 86, 809-826.

43.

Shanmugam

(2011, October 12). Second reading speech: Parliamentary elections (amendment) bill 2010. Retrieved from http://www.elections.gov.sg/mediarelease/Second_Reading_by_Law_Minister_K_Shanmugam_on_the_Parliamentary_Elections_Amendment_Bill.pdf

44.

Somin

(2004, September 22). When ignorance isn’t bliss: How political ignorance threatens democracy. Policy Analysis, 525. Retrieved from http://object.cato.org/sites/cato.org/files/pubs/pdf/pa525.pdf

45.

Tan

T. H.

Mahizhnan

Ang

P. H.

(2011). Media myths and realities: Findings of national survey of media use in the general election [PowerPoint slides]. Retrieved from http://lkyspp.nus.edu.sg/ips/wp-content/uploads/sites/2/2013/06/S1_1_Tan-Tarn-How_0410.pdf

46.

Tan

Y. S.

Soh

Y. P.

(1994). The development of Singapore’s modern media industry. Singapore: Times Academic Press.

47.

Tewksbury

(2003). What do Americans really want to know? Tracking the behavior of news readers on the Internet. Journal of Communication, 53, 694-710.

48.

Tewksbury

Althaus

(2000). Differences in knowledge acquisition among readers of the paper and online versions of a national newspaper. Journalism & Mass Communication Quarterly, 77, 457-479.

49.

Tichenor

P. J.

Donahue

G. A.

Olien

C. N.

(1970). Mass media flow and differential growth knowledge. Public Opinion Quarterly, 34, 159-179.

50.

Tichenor

P. J.

Rodenkirchen

J. M.

Olien

C. N.

Donahue

G. A.

(1973). Community issues, conflict, and public affairs knowledge. In Clarke

(Ed.), New models for mass communication research (pp. 45-79). Beverly Hills, CA: SAGE.

51.

Tran

(2013). Does exposure to online media matter? The knowledge gap and the mediating role of news use. International Journal of Communication, 7. Retrieved from http://ijoc.org/index.php/ijoc/article/view/1550

52.

Tufekci

Wilson

(2012). Social media and the decision to participate in political protest: Observations from Tahrir Square. Journal of Communication, 62, 363-379.

53.

Viswanath

Finnegan

J. R.

Hertog

Pirie

Murray

D. M.

(1994). Community type and the diffusion of campaign information. International Communication Gazette, 54, 39-59.

54.

Viswanath

Kosicki

G. M.

Fredin

E. S.

Park

(2000). Local community ties, community-boundedness, and local public affairs knowledge gaps. Communication Research, 27, 27-50.

55.

Warschauer

(2002). Reconceptualizing the digital divide. First Monday, 7(7). Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.5210%2Ffm.v7i7.967

56.

Weaver

Drew

(1995). Voter learning in the 1992 presidential election: Did the “nontraditional” media and debates matter? Journalism & Mass Communication Quarterly, 72, 7-17.

57.

Weaver

Drew

(2001). Voter learning and the Internet in the 2000 presidential election: Did the media matter? Journalism & Mass Communication Quarterly, 78, 787-798.

58.

Wei

Hindman

D. B.

(2011). Does the digital divide matter more? Comparing the effects of new media and old media use on the education-based knowledge Gap. Mass Communication and Society, 14, 216-235.

59.

Yang

Grabe

M. E.

(2011). Knowledge acquisition gaps: A comparison of print versus online news sources. New Media & Society, 13, 1211-1227.