Revealing the relationship between rational fatalism and the online privacy paradox

ABSTRACT

Previous research has revealed the privacy paradox, which suggests that despite concern about their online privacy, people still reveal a large amount of personal information and don’t take measures to protect personal privacy online. Using data from a national-wide survey, this study takes a psychological approach and uses the rational fatalism theory to explain the privacy paradox on the Internet and the social networking sites (SNSs). The rational fatalism theory argues that risks will become rational if the person believes he or she has no control over the outcome. Our results support the rational fatalism view. We found that people with higher levels of fatalistic belief about technologies and business are less likely to protect their privacy on the Internet in general, and the SNS in particular. Moreover, such relationship is stronger among young Internet users compared with older users.

KEYWORDS:

• Privacy
• self-disclosure
• fatalism
• social networking sites

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Wenjing Xie http://orcid.org/0000-0003-0575-7072

Notes

1. Tabachnick and Fidell (2007) suggested that oblique rotation such as direct oblimin be performed to decide between orthogonal and oblique rotation in EFA. According to them, the correlations of factor correlation matrix exceeding .32 would warrant oblique rotation. Otherwise, orthogonal rotation should be adopted. Therefore, we first performed the EFA with direct oblimin rotation method, which assumes that the factors are correlated with each other. The factor correlation matrix for correlations is lower than .32 (r = .305), suggesting ‘the solution remains nearly orthogonal’ (Tabachnick and Fidell 2007). Thus, we chose varimax rotation. Varimax rotation is an orthogonal rotation and maximises the variance of the squared loadings of a factor on all the variables. It has the effect of differentiating the original items by extracted factor, and can yield results that are easy to identify each variable with a single factor (Russell 2002).

2. The reason that we chose to create two new variables using weighted sum scores based on the factor loadings instead of using the saved factor scores is because the weighted sum scores can capture a greater proportion of the true score variance of the factors, and are less confounded by true scores from factors other than those they are supposed to be estimate of (DiStefano, Zhu, and Mindrila 2009; Grice and Harris 1998). Moreover, compared with using saved factor scores, the weighted sum scores are more stable across samples (Grice and Harris 1998). Weighted sum scores also have been the preferred scoring method among researchers (Alwin 1973; Ten Berge and Knol 1985; Gorsuch 1983; Grice 2001).

3. Cohen et al. (2014) suggested that linear regression be used first to determine if a linear or nonlinear regression fits the data. Therefore, although Kerwin’s (2012) original work showed non-monotonic relationship between fatalism and self-protective behaviour, this study started with testing the linear relationship between fatalistic beliefs and privacy protection behaviour. The scatterplots of residuals versus predicted values showed the linear pattern in the data. Thus, we chose linear regression instead of non-monotonic approach in our study.