Cultural Differences in Students' Privacy Concerns in Learning Analytics across Germany, South Korea, Spain, Sweden, and the United States

Olga Viberg, René F. Kizilcec, Ioana Jivet, Alejandra Martínez-Monés, Alice Oh, Chantal Mutimukwe, Stefan Hrastinski, Maren ScheffelarXiv 2023 · DOI · arXiv

Abstract

Learning Analytics (LA) offers transformative potential for higher education, yet its implementation is often hindered by student privacy concerns. This study examines how cultural values influence these concerns across five countries: Germany, South Korea, Spain, Sweden, and the United States. By integrating the SPICE model of privacy with individual-level cultural dimensions (CVSCALE), the researchers analyzed survey data from 762 students. The results reveal that while institutional trust is high, cultural dimensions like power distance and uncertainty avoidance are critical predictors of privacy perceptions. Notably, individualism was not a significant factor, suggesting that the educational context alters traditional privacy dynamics. These findings underscore the need for "culture-sensitive" LA policies that move beyond universal privacy-by-design approaches.

Introduction

Learning Analytics (LA) has the potential to revolutionize education by leveraging data to personalize learning pathways and improve institutional decision-making. However, the success of LA initiatives hinges on student participation, which is frequently obstructed by privacy concerns. Historically, privacy research in LA has adopted a Western-centric, "one-size-fits-all" approach, assuming that data ethics are universal.

This study challenges that assumption by investigating privacy as a social construct shaped by cultural nuances. By examining students across five diverse nations—Germany, South Korea, Spain, Sweden, and the U.S.—the researchers aim to understand how national and individual cultural dimensions influence the "SPICE" of privacy: Perceived risk, perceived control, privacy concerns, trusting beliefs, and non-self-disclosure behavior.

Methodology

The researchers employed a survey-based approach ($N = 762$) utilizing two primary theoretical frameworks. The first is the SPICE Model, a validated instrument designed to measure information privacy concerns specifically within the LA context. It maps the relationships between antecedents (perceived risk and control) and outcomes (trusting beliefs and disclosure behavior).

The SPICE model of privacy concerns in Learning Analytics.

The second framework involves Hofstede’s Cultural Dimensions, focusing on Power Distance, Individualism vs. Collectivism, Masculinity vs. Femininity, and Uncertainty Avoidance. Crucially, the authors used the CVSCALE to measure these values at the individual level rather than relying solely on national averages. This approach accounts for the fact that personal worldviews can vary significantly within a single country. Data was collected between 2021 and 2022 from students in Germany ($n=247$), South Korea ($n=59$), Spain ($n=121$), Sweden ($n=156$), and the U.S. ($n=179$), primarily representing STEM and HASS disciplines.

Results and Discussion

The study uncovered several key insights into the intersection of culture and privacy. A primary finding was the "Institutional Trust Advantage." Unlike the "privacy paradox" often observed in commercial social media—where users express concern but share data freely—students demonstrated high inherent trust in their universities. They generally perceived the risks associated with LA as manageable compared to those of commercial technology.

Hofstede’s Cultural Dimensions across the five studied countries.

Cultural dimensions played a significant role in shaping these perceptions. Power Distance emerged as a "double-edged sword": students who valued hierarchy reported higher perceived control but also higher perceived risk and a greater likelihood of refusing data disclosure. This suggests that perceived power imbalances make students more cautious about sharing information with institutional authorities. Conversely, high Uncertainty Avoidance was a strong positive predictor of trusting beliefs. When institutions provide clear, transparent guidelines, students who fear ambiguity are more likely to trust the system.

Interestingly, Individualism was not a significant predictor of privacy concerns. This contradicts previous research in general information systems and suggests that in higher education, the collective benefits of data-driven insights might outweigh individualistic desires for total data autonomy.

Geographically, Sweden and Germany were the most trusting and least concerned. U.S. students, however, reported the highest perceived risk (17% higher than other samples) and the lowest perceived control. Despite their different national scores in Hofstede’s original index, South Korea and Spain showed remarkably similar responses across all privacy dimensions.

Conclusion

While this study provides valuable cross-cultural insights, it is limited by non-probability sampling and a focus on technical universities, which may affect generalizability. Future research should include qualitative studies to explore the underlying reasons for these cultural variations. Ultimately, the findings suggest that LA policies must be "culture-sensitive." Moving beyond technical "privacy-by-design," institutions should adopt value-based designs that respect local levels of institutional trust and power dynamics to foster a more inclusive and effective data-driven educational environment.

Summary updated

BibTeX
@misc{viberg2023cultural,
  author = {Olga Viberg and René F. Kizilcec and Ioana Jivet and Alejandra Martínez-Monés and Alice Oh and Chantal Mutimukwe and Stefan Hrastinski and Maren Scheffel},
  title = {{Cultural Differences in Students' Privacy Concerns in Learning Analytics across Germany, South Korea, Spain, Sweden, and the United States}},
  year = {2023},
  eprint = {2312.02093},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2312.02093},
}
Figure