Privacy Literacy on Social Media: Its Predictors and Outcomes

Abstract

This study focuses on privacy literacy on social media with respect to how it functions as a marker for reflecting and alleviating the privacy divide. In doing so, the paper investigates the influence of objective measures such as characteristics of user populations (e.g., ethnicity, gender, and privacy experience) on the degree of privacy literacy on social media. Then it examines processes by which privacy literacy influences advanced privacy practices (privacy rule application and adaptation) through the perceived technology usability and information co-ownership. A structural equation model is proposed to test the hypotheses. Results show that ethnicity does not significantly impact privacy literacy, but gender does: more women than men are privacy literate. Experiences with privacy features are found to increase privacy literacy. Findings show a mediated effect of perceived usability of technology on the association between privacy literacy and privacy rule application. Another outcome of privacy literacy, perceived information co-ownership, was proposed to mediate the association between privacy literacy and privacy rule adaptation, and the prediction is supported. The final model, with added paths based on modification indices, suggests the following: a direct positive impact of gender (women) on the perceived usability of technology and information co-ownership, a direct negative impact of ethnicity (Asian) on the adaptation of privacy rules, and a direct positive path from experiences with privacy features to perceived information co-ownership. The study viewed privacy literacy on social media as an indicator of hardwired divide in technology use. This perspective of privacy literacy then functions to expand the contextual understanding of the role of privacy literacy in explaining advanced privacy practices.

Acknowledgements

I sincerely thank Drs. Craig R. Scott and Sun Kyong (Sunny) Lee for feedback and advice throughout the whole process of this project.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Author note

The author is now in The State University of New York at New Paltz.

Notes

1 About 1% of participants responded that others have never posted on their Facebook timeline.

2 None of participants responded “never” on how often others tagged them into the others’ posts.

3 The adequacy of statistical power was met by the sample size (N = 322) (Boomsma & Hoogland, 2001; Kyriazos, 2018)

4 Model fit was slightly better when a path was added from privacy rule application to privacy rule adaptation. Due to the theoretical focus of the model that treated these two variables distinctively and that was based on the framework of digital divide, a priority for adding a path was put on meeting the theoretical assumption.

5 To further verify that privacy literacy has the measurement validity for the mediation analyses in the model, the original model was compared to a constrained model where there is no mediation assumed. In doing so, the paths from the perception measures (perceived usability of technology and perceived information co-ownership) to the advanced privacy practices were fixed and constrained. This model comparison proved that the model fit was statistically significantly better for the original, p < .05, (χ²(16) = 38.97, CFI = .94, NFI = .90, RMSEA = .07) than for the constrained model (χ²(18) = 46.93, CFI = .92, NFI = .88, RMSEA = .07).

Additional information

Notes on contributors

SoeYoon Choi

SoeYoon Choi is an Assistant Professor of Communication in The State University of New York at New Paltz. Her research interest covers privacy, self-disclosure, technology affordance, and the impact of psychological distance on emotional disclosure on social media.