DATA_Social Media, Fake News and Influence on Elections in NIGERIA

This data set is from a cross-sectional online survey sample of 290 social media users who voted during Nigeria’s 2019 Elections in response to the dearth of evidence on the impacts of fake news on attitudes and beliefs, which can impact an election as a democratic event. Two hypotheses were proposed: (1) Fake news awareness will have no significant negative influence on voters due to its limited reach; (2) Inoculation against fake news will protect voters against the influence of fake news as postulated by the Inoculation The instrument adapted concepts and scales from previous studies to ensure the reliability of measurement. Fake news was operationalized from previously cited extant literature (See: Allcott & Gentzkow, 2017; Siapera, 2018; Uberti, 2017) as misleading and false information made to look like a true or factual story; that may be intentionally fabricated or manipulated with the intention of influencing public opinion on a certain reason for certain benefit, in this case, political gains or advantage. The data set contains Likert and non-Likert Scale questions. The Likert Scale components comprised 5-points (1=Strongly Disagree to 5=Strongly Agree, and 1=Very Infrequently to 5=Very Frequently). Items for measuring “awareness on fake news” was adapted from Fletcher et al. (2018); “influence of fake news” from Allcott and Gentzkow (2017); and “inoculation against fake news” from Cook et al. (2017), Roozenbeek and van der Linden (2018) and Van der Linden et al. (2017). The non-Likert Scale component was used to collect relevant demographic data (age, education, and occupation), as well as descriptive data on prevalence and sources of fake news on the elections, which required no reliability of scales. The adapted Likert-scale survey instrument initially contained 21 items. In the following pre-test with a convenience sample of 30 different sets of respondents, reliability of scales for internal consistency of measures helped in eliminating 10 items with low Cronbach Alpha (α) values. Three scales were finally measured using an 11-item 5-point Likert-scale instrument as follows: Awareness on fake news (4 items), Influence of fake news (3 items), and Inoculation against fake news (4 items). The Cronbach Alpha (α) values ranged between 0.712 and 0.938. Spearman Rho Correlation, the non-parametric statistical tested correlation between “awareness of fake news and influence of fake news”; and “influence of fake news and inoculation against fake news” on voters. Spearman Rho Correlation was tested instead of Pearson’s product-moment Correlation because the data set violated normality assumptions. Mean scores, frequencies, and percentages (descriptive statistics) were used to report the participants’ demographic statistics and prevalence of fake news. The results revealed the effects of fake news on voters, its prevalence on three social media platforms, and recent evidence on the Inoculation Theory regarding the election.