The Danger of Censorship: Why I Am Glad That Freedom of Speech Still Exists (And Why We Must Defend It)

I invite all of you to try to forget major aspects of your life for a moment, including your work, your values, and all the related ethical considerations.

Let's also try to define dis-misinformation. One possible way to do it comes from Spitzberg [1]: “[...] any message or a set of messages that represent a meaning complex discrepant from or incompatible with a sender’s intent and/or a relatively informed or expert consensual evidentiary state.” Therefore, sharing any paper that contains even a single improper methodological aspect or unclear communicative content, regardless of the motivations behind the sharing, generates a certain type of dis-misinformation. As alien observers, we must then try to establish varying degrees of impact of dis-misinformation (infodemiological objective). This means methodologically studying both the modes of propagation and the consequences of such propagation, in order to identify the forms of infodemic (information overabundance, which includes dis-misinformation) that most threaten the preservation of collective health.

As alien observers, we can then notice how scientists often oversimplify this aspect based on their beliefs and/or limited perceptions. Due to psychological mechanisms such as availability heuristics, biases (in the epidemiological sense, including ideological values and conflicts of interest), and expertise, a scientist S1 who is more exposed to - or naturally more sensitive to - an issue X will be more adept than scientist S2 at identifying that issue compared to an issue Y; and vice versa [2-4]. Therefore, it is easy to find content C (e.g., an academic article) that, according to the analysis of different researchers, contains different degrees and types of dis-misinformation. Specifically, C may be categorized as dis-misinformation by S1 but not by S2 (this is frequent in web discussions) [2-4].

This creates a critical issue in achieving the infodemiological objective established above since all actors are separate entities observing the same phenomenon (within the same domain of reality) but perceiving different aspects of it through different perceptual modes. At this point, communication comes into play, a fundamental component for all social groups and systems (including science itself) [4-6]. In particular, decision-making must be based on the degree of collective agreement regarding the validity of such a process, which depends on epistemic evaluations (e.g., whether the phenomenon exists or not) and contextual evaluations (costs, risks, benefits) [5-8].

Now, let's all regain our personalities and abandon the alien point of view. This means that everything I say from here on will be more conditional on my personal knowledge. The key point is that every human being - like every animal - fails to fully recognize and understand the ecology in which they are. To date, there is no reliable infodemiological model that can establish the overall degree of dis-misinformation, especially in a constantly changing psycho-social reality (e.g., the impact can vary greatly over time, a characteristic of all phenomena marked by amorphous and situational regularities) [9, 10]. This is evident within the scientific community itself, as shown by the low average level of agreement on the management of the COVID-19 pandemic and the related infodemic [2-4].

The factual impossibility of reaching such an agreement so far is plausibly linked to the high margin of subjectivity in assessing episodes that are difficult to categorize due to relevant scientific uncertainties, as well as to more insidious dynamics like those described regarding S1 and S2. Among the widely overlooked (subordinate) factors is the role of conflicts of interest, many of which are hidden in the fog of unfounded but largely shared methods such as those related to the mis/abuse of statistical significance (a remarkable, decades-long, and still seriously underestimated threat for public health worldwide) [9-16]. Alongside this, I too often encounter the tendency to create false value-driven dichotomies (e.g., 'denouncing a possible conflict of interest regarding a drug = asserting its ineffectiveness' or 'biased = useless') and epistemic dichotomies (e.g., 'absence of evidence = evidence of absence') [17]. In such a scenario, I am personally glad that freedom of speech exists, even if it means tolerating everything with which I disagree (provided that fundamental human rights are not violated). Indeed, our inability to grasp situations in their complexity becomes dangerous when we take on the role of censors, as we risk eliminating information that may be useful in ways we do not recognize [1-4, 6, 18]. As can be inferred, accepting such a level of uncertainty is psychologically complex, and usually, addressing it can lead to violent emotional reactions or mental closure (e.g., finding comfort in one's own prejudices) [4, 6].

Nonetheless, it is fair to acknowledge that I have also encountered many extremely rational defense lawyers (i.e., capable of expressing themselves in ways consistent with the shared goals of their social group) who are good at evading criticism with the rhetoric of 'science improving lives,' which is absolutely useless for identifying and adequately addressing those aspects in which it struggles to achieve such an honorable goal (I was myself one of them). Of course, this is unsurprising [6]. Thus, in my honest opinion, the best we can do is to adopt an unconditional approach to investigations, which does not mean never taking a stance but rather striving to give equal cognitive and methodological weight to all contextually relevant hypotheses (including those we consider inconvenient for ourselves or others) [19]. We must try to describe what we see instead of what we want to see while taking into account that our vision of the universe is partial and mediated by external sources (e.g., journals, data, others' evaluations) or internal confounding factors (e.g., unconscious beliefs, limited skills and tools). This includes accepting that within every piece of information (even those we dislike), there may be possible nuances of 'truth.'

REFERENCES

1. Spitzberg, B.H. (2021). Comprehending Covidiocy Communication. In Communicating Science in Times of Crisis (eds H.D. O'Hair and M.J. O'Hair). https://doi.org/10.1002/9781119751809.ch2

2. O’Hair HD, O’Hair MJ, editors. Communicating Science in Times of Crisis: COVID-19 Pandemic. Hoboken, NJ: Wiley-Blackwell; 2021. ISBN: 978-1-119-75179-3. Available at: https://www.wiley.com/en-us/Communicating+Science+in+Times+of+Crisis%3A+COVID-19+Pandemic-p-9781119751793

3. Rovetta, A., & Castaldo, L. (2022). Are We Sure We Fully Understand What an Infodemic Is? A Global Perspective on Infodemiological Problems. JMIRx med, 3(3), e36510. https://doi.org/10.2196/36510

4. Rovetta, A., & Castaldo, L. (2024). Empathy, Kindness, and Moderation are not Just Formalities in Science. Information & Media, 96, 153-160. https://doi.org/10.15388/Im.2023.96.71 (Original work published 2023). Full text available at: https://www.journals.vu.lt/IM/article/view/32291/32560

5. Hennig, C. (2010). Mathematical models and reality: A constructivist perspective. Foundations of Science, 15(1), 29–48. https://doi.org/10.1007/s10699-009-9167-x

6. Greenland, S. (2023, April 19). There’s not much science in science: Addressing the psychosocial gap in methodology [Lecture]. The Department of Epidemiology Seminar Series. UCLA School of Public Health. Available at: https://www.youtube.com/watch?v=N7-yn5dd7Hg

7. Greenland S. (2021). Analysis goals, error-cost sensitivity, and analysis hacking: Essential considerations in hypothesis testing and multiple comparisons. Paediatric and perinatal epidemiology, 35(1), 8–23. https://doi.org/10.1111/ppe.12711

8. Rovetta, A., Mansournia, M. A., & Vitale, A. (2024). For a proper use of frequentist inferential statistics in public health. Global epidemiology, 8, 100151. https://doi.org/10.1016/j.gloepi.2024.100151

9. Ting, C., & Greenland, S. (2024). Forcing a Deterministic Frame on Probabilistic Phenomena: A Communication Blind Spot in Media Coverage of the “Replication Crisis”. Science Communication, 46(5), 672-684. https://doi.org/10.1177/10755470241239947

10. Amrhein, V., Trafimow, D., & Greenland, S. (2019). Inferential Statistics as Descriptive Statistics: There Is No Replication Crisis if We Don’t Expect Replication. The American Statistician, 73(sup1), 262–270. https://doi.org/10.1080/00031305.2018.1543137

11. Amrhein, V., Greenland, S., & McShane, B. (2019). Scientists rise up against statistical significance. Nature, 567(7748), 305–307. https://doi.org/10.1038/d41586-019-00857-9

12. Gelman A. (2018). The Failure of Null Hypothesis Significance Testing When Studying Incremental Changes, and What to Do About It. Personality & social psychology bulletin, 44(1), 16–23. https://doi.org/10.1177/0146167217729162

13. McShane, B. B., Gal, D., Gelman, A., Robert, C., & Tackett, J. L. (2019). Abandon Statistical Significance. The American Statistician, 73(sup1), 235–245. https://doi.org/10.1080/00031305.2018.1527253

14. Rafi, Z., Greenland, S. Semantic and cognitive tools to aid statistical science: replace confidence and significance by compatibility and surprise. BMC Med Res Methodol 20, 244 (2020). https://doi.org/10.1186/s12874-020-01105-9

15. Juni, Robin, When the Math Matters: Use of P-Values in Pharmaceutical Litigation (2023). 48:153 Vermont L. Rev. (2023 or 2024)., GWU Legal Studies Research Paper No. 2023-36, GWU Law School Public Law Research Paper No. 2023-36, Available at SSRN: https://ssrn.com/abstract=4501656

16. Barnett, A. G., & Wren, J. D. (2019). Examination of CIs in health and medical journals from 1976 to 2019: an observational study. BMJ open, 9(11), e032506. https://doi.org/10.1136/bmjopen-2019-032506

17. Altman, D. G., & Bland, J. M. (1995). Absence of evidence is not evidence of absence. BMJ (Clinical research ed.), 311(7003), 485. https://doi.org/10.1136/bmj.311.7003.485

18. Sayers, F. (2024, May 8). How ‘fighting disinformation’ turns into political censorship: Self-appointed monitors can financially harm publications as they choose. The Washington Post. Available at: https://www.washingtonpost.com/opinions/2024/05/08/disinformation-political-censorship-unherd-gdi/

19. Greenland, S., Rafi, Z., Matthews, R., & Higgs, M. (2022). To aid scientific inference, emphasize unconditional compatibility descriptions of statistics. arXiv. https://arxiv.org/abs/1909.08583v7