Statistically untestable hypotheses cannot be proven or disproved with the available evidence. This can happen for several reasons, such as the hypothesis needing to be more specific or complex or the evidence needing to be more comprehensive or conclusive. Reversing the burden of proof is a common rhetorical technique used to derail hypotheses and theses.
Hypothesis testing and burden of proof:
Several changes can be made to the hypotheses presented in the article to make them statistically untestable. These changes are the following (Montañez-Juan et al., 2019):
- Reverse the burden of proof: Instead of requiring the researcher to prove a relationship between work design, organizational fairness, and job satisfaction, the researcher will ask the skeptic to prove that there is no relationship.
- Making Hypotheses Untestable: Instead of saying that work design is positively associated with job satisfaction, the researcher can say that work design is positively associated with job satisfaction for employees with a high level of organizational commitment.
- Making the hypotheses more complex: Instead of saying that work design is positively associated with job satisfaction, the researcher can say that work design is positively associated with job satisfaction for employees with a high level of organizational commitment, but only if the organization has a strong culture of innovation.
This would make testing the hypothesis more difficult. The researcher would need to determine how to measure organizational commitment and control for other factors that could influence job satisfaction.
Implications of the burden of proof:
A potential relationship exists between statistically untestable hypotheses and the burden of proof implications, but it depends on the specific context and situation. A statistically untestable hypothesis cannot be tested or evaluated using conventional statistical methods (Moosa, 2017). This may be because the theory is not refutable or because the data required to test the theory are not available or are not observable.
In some cases, proponents of such hypotheses may argue that the burden of proof is on others to disprove the hypothesis rather than provide evidence to support it. Reversal of the burden of proof is a logical fallacy in which the burden of proof is shifted from the person claiming to the person who disputes or doubts the claim. This is an unfair tactic because the burden of proof should always fall on the person making a claim, not the person questioning it (Epstein, 2020).
Real-world example:
This often happens to me when discussing someone who believes in conspiracy theories. When asked about providing evidence of any conspiracy from him, I find him saying that it is clear without evidence, and I have to search to find the evidence that I am reassured of. As for him, he does want to refrain from explaining his claims or discussing them objectively. Thus, he reflects the refutation of the evidence on me, as if I am the owner of the theory that he believes in.
Conclusion
It is important to note that reversing the burden of proof does not mean winning an argument. It simply means that you have shifted the burden onto the other person. The other person may still be able to provide evidence to support their claim; if they do, you will still need to process that evidence. When a hypothesis is not statistically testable, it can be challenging to determine whether it is true or false. This can lead to people reversing the burden of proof. In other words, they might argue that the person who proposed the hypothesis must prove it authentic rather than the person who doubts the hypothesis and must prove it false (Walton, 1988).
References
Epstein, Michael (2020). Defining Science / General Taxonomy / Burden of Proof / Null Hypothesis. Retrieved from: http://www.amplaboratory.org/classes/statistics/statistics-unit-1-defining-science-general-taxonomy/ (pass: teaching stats)
Montañez-Juan, M. I., García-Buades, M. E., Sora-Miana, B., Ortiz-Bonnín, S., & Caballer-Hernández, A. (2019). Work design and job satisfaction: The moderating role of organizational justice. Revista Psicologia Organizações e Trabalho, 19(4), 853–858. Retrieved from EBSCO multi-search
Moosa, I. (2017). Covered interest parity: The untestable hypothesis. Journal of Post Keynesian Economics, 40(4), 470-486.
Walton, D. N. (1988). The burden of proof. Argumentation, 2, 233-254. https://www.academia.edu/10978804/Burden_of_proof