posc data

The Effect of Political Ideology on Media Preferences 26

The Effect of Political Ideology on Media Preferences 26

The Effect of Political Ideology on Media Preferences 26

The Effect of Political Ideology on Media Preferences During the 2016 Presidential Election

[Name Redacted]

Towson University

POSC 301, Political Science Research

December 11th, 2017

Abstract

This paper explores the effects of political ideology on an individual’s media preference for political news. Expansion of internet access has enabled information to become for selective in nature (Iyengar, Hahn 2009). Rapid adoption of the new technology drives research into the subject. Using data collected through a pre-test and post-test survey conducted by the American National Election Study during the 2016 Presidential Election we find that liberals more than conservatives have a greater preference for traditional media than the internet. The possible

explanations regarding shifting ideological preferences are examined.

Introduction:

The innovation of technologies that have expanded human communication and access to information have marked significant political and societal transitions throughout human history. In seventeenth century Europe, the development of the printing press helped fuel the Protestant Reformation. Later, in the nineteenth century, radio and television would bring nightly news broadcasts and images of global events into the American home. These innovations drastically changed the way in which people accessed information and understood the world around them. Today, the internet is revolutionizing the way humans learn, communicate and interact with the world around them. According to Pew Research Center data collected in 2016, the number of American adults who use the internet has increased to 88%, compared to 52% in 2000 (Pew Research Center, 2017). Through the expanded use and ownership of the smartphone, or cellular telephone capable of accessing the internet, daily internet usage has become synonymous with 21st century society. In 2016, over three-quarters (77%) of the adult population in the United

States owned a smartphone in 2016 (Smith, 2017). According to another Pew Research Study

(2016), 43% of American Adults use the internet as their primary source for news, compared to 50% who prefer television (Gottfried, Shearer, 2017). A similar study found that between 19962016, newspaper jobs had shrunk by 20,000 between – a 39% decrease (Barthel, 2017).

Despite, the narrowing gap between voter preference traditional media and the internet for political news, the ideological divide between Conservatives and Liberals in America continues to widen. According to data collected by the Pew Research Center, the percentage of Conservatives and Liberals who shared policy positions with those of the other, dropped to 32% in 2017, compared to 49% in 1994 and 2004 (Kiley, 2017). This phenomenon is not new; polarization has increased significantly since the introduction of the 24-hour news cycle and cable television, and increased media variety (Iyenghar and Hahn, 2009). Today, the expanded use of the internet has increased the variety and saturation of the political media market. With so many options, the inability to hear “point-counterpoint” news-analysis, or what Iyenghar and Hahn (2009) considered essential for neutral news consumption, may be limited by an individual’s preferred form media. Therefore, as the basis of my research is to answer the question: How does political ideology affect media use?

Literature Review:

This literature review will explore three hypotheses predicting media usage: age, education and degree of polarization. The first hypothesis, which believes age is the modulating factor of preferred media reliance and use, suggests that older Americans are more likely to select traditional news sources for political information, whereas younger Americans would expand use to include new media. The second hypothesis, Degree of Polarization, argues that a person’s political affiliation determines their media consumption, as people seek out news sources that align with their preexisting political beliefs. Third, a person’s level of education can impact their media consumption. Overall, I conclude that political ideology will be the strongest determinant of media consumption.

Hypothesis 1: Younger people will be less reliant on traditional media than older people

This hypothesis is supported by the Age school of thought, along with several previous studies. For example, Richey & Zhu (2015) find that despite increased internet availability due to falling prices and increased network availability, older Americans and those who became acquainted with the new technology, or “late adopters”, use the internet less than those who adopted its use earlier. Nimrod (2016) also showed that while internet usage is growing with older people, they still prefer the traditional forms of media and communication that they have been using for a majority of their lives. There a several theoretical explanations for why this is the case. Some studies, like Nimrod (2016) and Richey & Zhu (2015) argue that the lack of familiarity and steep learning curve older people face when using new forms of media, compared to younger people who have been raised using them, is the main explanation.

However, other studies show that the apparent relationship between age and media consumption may be a result of the real relationship between political affiliation and media consumption. Gelman & Ghitza (2014), suggest that generations often have a political identity, citing information gathered from political events since the 1940s. These generational differences in identity, and their preferences, might explain media use by generation, however, the availability and widespread use of new technologies will influence one’s familiarity, comfortability, and use of new media sources. A hybrid study conducted by Boxell, Gentzkow, Shapiro (2017), also found that polarization is increasing most among those aged 75 years or older, who predominately do not use the internet as opposed to those who are aged 18-39. The study highlights what likely drives older Americans (51 or older), especially the fact that they are predominately conservative.

While previous research has made it clear that age likely plays a role of media preferences, I believe that there will be a stronger relationship between political ideology and media preferences, and that much of the apparent relationship with age is really explained by the contingent relationship between media preferences and political ideology.

Hypothesis 2: People who identify as conservatives will show a greater reliance on traditional media than people who identify as liberals

This hypothesis is supported by the Degree of Polarization school of thought, which argues that as American politics and the news media has become more politically polarized, people have begun to seek out media which supports their more preexisting political beliefs. This school is based on the theory of cognitive dissonance, which describes the tendency of individuals to seek out beliefs or opinions that mirror their own (Festinger 1957). This “selectivity” (Iyengar, Hahn, 2009) of media choices has also led to individuals’ increased distrust of different media types, as well as simultaneously contributed to greater disparity between political poles. The distrust is derived from partisan messaging which seeks to use the individuals own perceive political ideological beliefs to reinforce American political party affiliation. An example of the dispersion of partisan dogma, is how conservative Republican party has been able to craft a narrative, which claims that the media is controlled by liberals (democrats) with the exception of only a few sources (all of which are, of course, conservative). As a result, previous research in this school has concluded that conservatives, more than liberals, have shown less variety in selection, choosing instead to follow only a few trusted news sources (Iyengar & Hahn, 2009). Whereas liberals, on the contrary, choose to collect political news from a variety of sources (Iyengar & Hahn, et al.)

An individual’s level of trust of specific news outlets based on their political views reinforced by polarization plays a larger role in the selection of news sources, while also reinforcing their reasons for doing so. Iyengar & Hahn (2009) found that liberals consume a greater variety of political news from a several sources, as opposed to conservatives who prefer to use a smaller variety of sources. Both conservatives, and liberals, however, selected their sources based on their ideological preferences. The study highlights and describes the underlying modulator of choice between ideologies in terms of their “selective exposure”, or their choice of sources in alignment with their way of thinking, as to reverberate own thoughts, in part due to a cognitive dissonance (Festinger, 1957). Given that conservatives see fewer forms of media as aligning with their political belfies, and the ones they do tend to trust tend to be traditional forms of media such as Fox news, it follows that they would favor traditional media.

I think the Degree of Polarization school of thought is the most appropriate school in describing the relationship between ideology and media selection and usage. This school better details the underlying factors of the relationship between ideology and media selection and that identification with the group itself, and the messages from that groups leadership, i.e. party elites, and candidates is often more important than specific policy positions. This in turn strengthens the desire for an individual to seek out information that corresponds, or agrees, with their perceived identity, and thus continually reinforces their beliefs. Although age correlates to the subject of media use, in that it shows although one may be less confident in their ability to use the internet or similar technology their age, income, or race will not adversely affect the levels in which someone is polarized. Given this, I believe that this hypothesis is true, and that the relationship between political ideology and media usage will be the strongest in my model.

Hypothesis 3: People who have achieved a higher level of education are less reliant of traditional media than those with a lower level of education

Another potential factor in a person’s media preferences is their level of education. Some previous studies have explored the role of education is political media preferences, such as Lee and Yang (2014) who found that those with a higher level of education were more knowledgeable about political events whether they actively sought out political news or actively avoided it. This result shows the high importance of education level in political literacy, and that consuming news media is not the same as being informed about politics. This result supports the idea that people with a higher level of education have more highly developed perspectives about politics, and are more interested and engaged in politics, which in turn impacts their media preferences. Lecheler and Vreese (2016) found support for this idea in an experimental study, which explored “which individuals can be encouraged to read an unfamiliar information-rich newspaper, and if using this newspaper, in turn, has effects on interest and knowledge.” (Lecheler, Vreese, 2016). They found that individuals with preexisting political interest and knowledge were more likely to read the unfamiliar newspaper. As people with higher levels of education tend to have more political knowledge, this result shows that more educated people are more likely to explore a variety of news sources, and not just depend on traditional media. Also, as republicans tend to be less educated than democrats, there could be some overlap between the relationships between political ideology, education level, and media preferences. While I think I will find support for this hypothesis, I think the relationship will be than that between political ideology and media preferences.

Through analyzing the literature related to my three hypotheses, I conclude that I should find support for each of my hypotheses, but that I will find the most support for hypothesis 2 and the relationship between political ideology and media preferences is the strongest in the model. This is because there was a relatively larger body of research supporting that hypothesis compared to the others. Also, both age and education level impact someone’s political ideology, so the relationship between political ideology and media preferences is more direct than the more overlapped relationships with age and education level.

Methodology:

I used data from the American National Election Study (ANES) which was collected during the 2016 U.S. Presidential Election. The survey was conducted in two parts, one before, and one after, the election on November 8th, 2016. The pre-test took place between September

7th and November 7th, 2016, and the post-test was then conducted from November 9th, 2016 and January 8th, 2017 with the greatest number of the same respondents as possible. The survey was administered by using in-person interviews and complemented by the use of the internet to gather respondent responses to two substantively similar yet separate questionnaires. Respondents were only allowed to complete the survey using one of the two modes. For those who did not have internet access, internet access and computers were provided. The approximate time to complete the survey was 80 minutes not counting additional time for preparation and screening, and was offered in both English, and Spanish. The units of analysis used in the survey were the respondents themselves and their answers to 355 questions pertaining to topics about political attitudes, social issues, emotions, and media use. The respondents had to be over 18 years of age, be United States citizens, and speak either English or Spanish. The sample for the internet study was gathered through the use of the United States Postal Service Sequence File; for the face-to-face interviews the sample was gathered using a stratified sample from the contiguous 48 U.S. states (excluding Hawaii and Alaska) and accounted for race, age and income. The survey was funded by the University of Michigan, Stanford University, and the

National Science Foundation, with both universities overseeing the collection. Westat Inc. helped in collecting the Data.

To study how political ideology affects media selection I selected two dependent variables, an independent variable, and two control variables from the ANES 2016 codebook. The independent measure selected was, “PRE: Ideological Scale: Liberal-Conservative selfplacement”, and was measured using a 7-point Likert-scale, which asked the respondent to place themselves along the ideological scale using the following responses: “Extremely Liberal”,

“Liberal”, “Slightly Liberal”, “Moderate – middle of the road”, “Slightly Conservative”, “Conservative”, or “Extremely Conservative”. The dependent measures used were the provided responses to the following questions concerning the respondent’s media usage: “How many times internet was used for information about the 2016 Campaign”, “How many TV programs watched about 2016 Campaign?”, and “How many stories read in newspaper about 2016 Campaign?”. These variables were measured using a four-point Likert Scale to determine the frequency of media use, and ranged from, “None”, “Just one or two”, “Several”, or “a good many”. The variables used as controls in the study were “Respondent Age Group” and “Respondents highest-level of education”. The ordinal variable used to measure respondent age, “Respondent Age Group”, asked the respondent to place themselves within 13 age-groups located in the frequency distribution in Table 6. The measure for “Respondents highest level of education” asked the respondent to identify their highest level of education achieved from the following list of 17 levels: “1. Less than 1st grade”, “2. 1st, 2nd, 3rd or 4th grade”, “3. 5th or 6th grade”, 4. 7th or 8th grade”, 5. 9th grade”, “6. 10th grade”, “7. 11th grade”, “8. 12th grade no diploma”, “9. High school graduate- high school diploma or equivalent (for example: GED)”,

“10. Some college but no degree”, “11. Associate degree in college - occupational/vocational program”, “12. Associate degree in college -- academic program”, “13. Bachelor's degree (for example: BA, AB, BS)”, “14. Master's degree (for example: MA, MS, MENG, MED, MSW, MBA)”, “15. Professional school degree (for example: MD, DDS, DVM, LLB, JD)”, and “16.

Doctorate degree (for example: PHD, EDD)”.

My greatest validity concern within my research design, is the issue of simultaneity:

meaning both media selection (DV) and political ideology (IV) simultaneously serve to reinforce the other (Slater, 2007). For example, when given the freedom to choose media from a variety of sources, an individual who identifies as ideologically liberal is more likely to choose media types that align with their beliefs and opinions, which therefore then reinforces their already held ideological beliefs. One possible test would be to create random experiment which tests for ideological preferences based on limited choice from only fabricated news sources created by the researcher. However, I lacked the adequate means to accurately test this complex relationship based on the data, the likely cost, and time required to conduct the study. The study could be repeated and produce similar results; however, media usage will likely be differently the further amount of time the attempt to recreate this study is conducted. The questions regarding the dependent variables are too vague, considering they do no measure the content being accessed accurately. I would instead advocate for the question to be modified and be more deterministic regarding frequency, i.e. “How many times per month…”, as well as, include data pertaining to the content being accessed, and questions that compare usage of different mead types, such as,

“For news about ____ what are you more likely to use?”. Possible answers that could be provided could be, “Facebook”, “Google”, “Cable News”, etc. and possible include a tiered system that asks additional fields with accompanying questions, such as “Fox”, “CNN”, etc. once you cleared the first field “Which cable news are you more likely to use?” The external validity of the ANES 2016 data being used for the study is also affected by respondents not returning to answer questions in the post test, the respondent lying during screening for the sample, or other inconsistencies with selection and recording of answers. This is hard to fix.

Using a pre-, and post-test method of surveying always incorporates the risk of “dropping-out”. Nonetheless, it is still a concern with reliability of data since substitutes were used to compensate for attrition during the study period. Some of the variables were measured consistently however the responses for the 4-point Likert scale used in measuring the independent variables for media usage, possibly compromised accurate reporting. Validity may have been improved had the question included a 5-point Likert scale and a fifth possible answer, “Daily”.

Analysis and Assessment:

Descriptive Statistics

Each of my independent variables, namely age, self-identified political ideology and highest attained education level, along with my dependent variable traditional media us are ordinal, as they are all coded as ranked categories. For example, age is ranked on a 13-point scale, with each category representing 3 years from ages 18-75. My dependent variable, traditional media use, is the sum of two 4-point Likert Scales.

The distribution for my dependent variable traditional media usage was even, with the most common response being 5, and the mean being 5.12. However, there were only 197 respondents that answered lowest option, 2, and 431 who answered the highest option, 8. The distribution for the age variable was also fairly even, though it skewed slightly old with the most common response being 9, meaning the respondent was 55-59. However, the mean was less, at 7.44 at the median as well at 8.00. The distribution for ideological self-placement was also mostly even, though slightly skewed conservative with a median of 5, which is slightly conservative and one point higher than moderate. 4, meaning moderate, was the most common response, with 895 responses. The distribution for education level was mostly even but slightly skewed towards more highly educated, with the most common response being 13, meaning a bachelor’s degree, and a mean of 11.74 on the 16-point scale. This could be explained by the fact that there were 9 categories for high school and below alone.

Bivariate Results

According to results of the bivariate regression (see Table 4), if Ideological self-placement on 7-point scale is 0, the expected value of traditional media use on a scale from 2 to 8 is 5.316 with higher values meaning more use of traditional media. A one-point increase on 7-point scale, with higher values meaning more conservative, leads to a -0.08 decrease in traditional media use. The relationship is significant, with a t-score of 11.012 which is greater than +/- 2 and a p-value of <.001, so p<.05. With an R-Squared of .032, 3.2% of traditional media use can be explained by Ideological self-placement. This result does not support my research hypothesis, as I expected conservatives to use traditional media more, and the results shows that liberals do.

According to Table 5, if a respondent’s age is 0 on a scale from 1-12, with 1 being 18-20 and 13 being 75 or older, the expected value of traditional media use on a scale from 2 to 8 is 4.075. A one-point increase in age on the 12-point scale leads to .142 point increase of traditional media use on the scale from 2 to 8. The relationship is significant, with a t-score of 17.888 which is greater than +/- 2 and a p-value of <.001, so p<.05. With an R-Squared of .083, 8.3% of traditional media use can be explained by Ideological self-placement. This result supports my research hypothesis that younger people will be more reliant on traditional media than older people.

According to Table 6, if a respondent’s level of education is 0 on a scale from 1 to 16, with 1 being less than first grade and 16 being a doctorate, then the expected value of traditional media use is 4.982 on a scale from 2 to 8. A one point increase in education on the 16-point scale leads to a 0.12 point increase in traditional media use on a scale from 2 to 8. . The relationship is significant, with a t-score of 3.064 which is greater than +/- 2 and a p-value of .002, so p<.05. This goes against my hypothesis that people who have achieved a higher level of education are less reliant of traditional media, as higher education leads to slightly higher traditional media usage.

Multivariate Results

If age on a 13 point scale, education on a 16 point scale, and self-identified political ideology on a 7 point scale are all 0, then the expected value traditional media use is 4.242, on a scale from 2 to 8. With R-Squared of .109, about 10.9% of a person’s use of traditional media use can be explained by their age, level of education and self-identified political ideology.

A one point increase in ideology on a 7-point scale where higher values mean more conservative leads to a -.007 decrease in traditional media use on a scale from 2 to 8. This relationship is not significant, with a t-score of .938, which is smaller than +/- 2 and a p-value of .348, so p>.05. This result matches my bivariate result and goes against my hypothesis that conservatives will be more reliant on traditional media, as it shows that liberal respondents are slightly more likely to have used traditional media.

A one-point increase in age on a 13-point scale with 1 being 18-20 and 13 being over 75 leads to a .136-point increase in traditional media usage on a scale from 2 to 8. The relationship is significant, with a t-score of 17.347 which is greater than +/- 2 and a p-value of .<.001, so p<.05. This result matches my bivariate result and further supports my hypothesis that younger people are less reliant on traditional media than older people.

A one point increase in education level on a 16-point scale leads to a -.007 decrease in traditional media usage on a scale from 2 to 8. The relationship is significant, with a t-score of -9.985 which is greater than +/- 2 and a p-value of .<.001, so p<.05. This matches my bivariate result and further counters my hypothesis that conservatives are more reliant on traditional media, as more liberal respondents tend to use traditional media slightly more.

Findings:

The results of my research did not support my hypothesis that conservatives would show a greater preference in the use of traditional media for political news, and instead caused me to accept the null hypothesis. The cross-tabulation between traditional media (DV1) and ideological self-placement (IV)(see Table 2) showed that liberals, not conservatives, displayed a higher preference for traditional media than conservatives. This is surprising, however I believe that the limited data set available to me for use in my research restricted the results, as well as other validity concerns mentioned earlier. The results of the multivariate-regression explaining the preferred use of traditional media (Table 12) revealed that respondent education level played no significant role in the preferred selection of traditional media for political news. Both of the multivariate regressions (Table 12, 13) for traditional media and internet media indicated a decrease in preference of at least -.006 for both forms of media, each positive degree traveled along the ideological scale.

Conclusions:

Although my hypothesis was not supported by the tests used in this study, the results have still revealed many interesting phenomena worthy of being researched in the future. Understanding the way in which people access information is an essential aspect of media marketing and advertising, and trends can help predict future behavior. The data suggests that the more conservative someone identifies as, the less likely they are to prefer either forms of media used in this study for political news. This could possibly be explained by increased distrust of among conservatives, or possibly a dramatic shift in the historic preferences of the conservatives to traditional gravitate towards a limited number of sources. Unfortunately, the simultaneity of the relationship between political ideology and media selection is continuing to widen the ideological divide in American society. The increasing polarization is a continuing threat to American security. Further studies should be conducted in an effort to address the vulnerabilities that were exploited by foreign governments in an effort to influence the 2016 Presidential Election.

References:

Barthel, M. (2017, June 01). Newspapers Fact Sheet. Pew Research Center, Washington, D.C.

Retrieved December 12, 2017, from http://www.journalism.org/fact-sheet/newspapers/

Boxell, L., Gentzkow, M., & Shapiro, J. M. (2017). Greater Internet use is not associated with faster growth in political polarization among US demographic groups. Proceedings Of

The National Academy Of Sciences Of The United States Of America, 114(40), 10612.

doi:10.1073/pnas.1706588114

Festinger, L. (1957). A theory of cognitive dissonance. Evanston, Ill., Row, Peterson [1957].

Ghitza, Y., & Gelman, A. (2014, July 7). The Great Society, Reagan’s Revolution, and Generations of Presidential Voting . Columbia. Retrieved December 11, 2017, from http://www.stat.columbia.edu/~gelman/research/unpublished/cohort_voting_20140605.p df

Gottfried, J., & Shearer, E. (2017, September 07). Americans' online news use is closing in on TV news use. Pew Research Center, Washington, D.C. Retrieved December 12, 2017, from http://www.pewresearch.org/fact-tank/2017/09/07/americans-online-news-use-vstv-news-use/

Internet/Broadband Fact Sheet. (2017, January 12). Pew Research Center, Washington, D.C. Retrieved December 11, 2017, from http://www.pewinternet.org/fact-sheet/internetbroadband/

Iyengar, S., & Hahn, K. S. (2009). Red Media, Blue Media: Evidence of Ideological Selectivity in Media Use. Journal Of Communication, 59(1), 19-39. doi:10.1111/j.14602466.2008.01402.x

Iyengar, S., Sood, G., & Lelkes, Y. (2012). Affect, Not Ideology. Public Opinion

Quarterly, 76(3), 405-431.

Kiley, J. (2017, October 23). In polarized era, fewer Americans hold a mix of conservative and liberal views. Pew Research Center, Washington, D.C. Retrieved December 12, 2017, from http://www.pewresearch.org/fact-tank/2017/10/23/in-polarized-era-feweramericans-hold-a-mix-of-conservative-and-liberal-views/

Lee, Hyunwoo, & Yang, Jungae. (2014). Political knowledge gaps among news consumers with different news media repertoires across multiple platforms. International Journal of Communication, 8(1), 597–617.

Lecheler, Sophie, & de Vreese, Claes H. (2017). News Media, Knowledge, and Political Interest: Evidence of a Dual Role From a Field Experiment: News Media, Knowledge, and Political Interest.  Journal of Communication,  67(4), 545–564. https://doi.org/10.1111/jcom.12314

Richey, S., & Zhu, J. (2015). Internet Access Does Not Improve Political Interest, Efficacy, and

Knowledge for Late Adopters. Political Communication, 32(3), 396-413.

doi:10.1080/10584609.2014.944324

Smith, A. (2017, January 12). Record shares of Americans now own smartphones, have home broadband. Pew Research Center, Washington, D.C. Retrieved December 12, 2017, from http://www.pewresearch.org/fact-tank/2017/01/12/evolution-of-technology/

Sophie Lecheler, Claes H. de Vreese, News Media, Knowledge, and Political Interest: Evidence of a Dual Role From a Field Experiment,  Journal of Communication, Volume 67, Issue 4, August 2017, Pages 545–564,  https://doi.org/10.1111/jcom.12314

Tewksbury, D., & Riles, J. M. (2015). Polarization as a Function of Citizen Predispositions and

Exposure to News on the Internet. Journal Of Broadcasting & Electronic Media, 59(3),

381-398. doi:10.1080/08838151.2015.1054996

Table

3

:

Traditional Medial Frequency Distribution

Table

4:

Ideological Scale: Liberal

-

Conservative Frequency Distribution

Table 5: Respondent Age Group Frequency Distribution

Table 6: Respondent Age Group Frequency Distribution

Table 7: Bivariate Regression (Traditional Media) Coefficients Table

Table

8

:

Bivariate Regression (Traditional Media DV) Model Summary

Table

9

Table

10

Table

11

Table

12