Two scientists can examine the same data and reach very different conclusions. Is that proof that science is broken or simply evidence that assumptions matter? A recent paper in Science Advances claims that a researcher’s personal views can shape scientific results.1 The authors suggest this reveals a serious weakness in science itself. However, their study does not indicate that science is broken. Instead, it shows that scientific conclusions depend on starting assumptions and method choices, which should not surprise us. All people bring bias into their work, and Christians have recognized this truth for centuries.
In the study, 158 researchers in 71 teams were given the same dataset and asked one question: does immigration affect public support for social welfare programs?1 Each team analyzed the data individually and got varied results. Teams whose members supported immigration often reported positive effects, but teams that opposed immigration often reported negative effects. Based on this pattern, the authors concluded that ideology influenced the production of the findings, reducing confidence in the results.1 But their own conclusion rests on a fragile assumption.
The authors’ key assumption is that systematic alignment between researchers’ ideology and their analytic conclusions demonstrates ideological bias. But this does not necessarily follow. In research, scientists make methodological decisions at every step: which variables to include, how to define key terms, how to treat missing data, and which statistical models best address the research question. These choices are not signs of dishonesty or hidden ideology but are normal features of complex data analysis and are widely acknowledged in research methodology.2 When data are multifaceted, different reasonable decisions can legitimately produce different results. Correlation between ideology and outcomes does not by itself demonstrate that ideological bias caused those differences; it may simply reflect the inherent flexibility within standard scientific practice.
Again, correlation should not be confused with causation. Showing that two things appear together does not prove that one caused the other. The study did not explain how personal beliefs changed specific analysis steps, nor did it evince careless or flawed models. It also did not test whether the differences would remain if the same teams repeated the work. Without tracing these decision paths, claims of ideological influence remain speculative.1 This highlights the broader issue about how assumptions guide interpretation.
And this issue does appear throughout science. For example, DNA stores information using a precise coding system, complete with error-checking and repair mechanisms that protect genetic integrity. Conventional scientists often assume these systems arose through unguided processes. Creation scientists begin with the expectation of purposeful design. Both groups study the same molecular structures but interpret their origin differently. The contrast lies not in the data but in the assumptions brought to the data—just as the Science Advances study unintentionally demonstrates.3
Jeremiah 17:9 explains why this happens: “The heart is deceitful above all things.” All people are subject to God’s diagnosis, including conventional scientists.4 Yet Scripture does not say truth is unreachable. It calls us to humility, honesty, and careful testing. These traits support good science.
By treating normal method differences as ideological bias, the paper goes beyond what its data support. From a creation perspective, the study points to a deeper truth—science works best when assumptions are clearly identified and tested against observable reality. True objectivity does not come from denying one’s worldview but from submitting human reason to the God who designed both the mind and the orderly world it seeks to understand.
References
- Borjas, G. J. and Breznau, N. 2026. Ideological Bias in the Production of Research Findings. Science Advances. 12 (1).
- Gelman, A. and Hill, J. 2006. Data Analysis Using Regression and Multilevel/Hierarchical Models. Cambridge, UK: Cambridge University Press.
- Lohrey, A. and Boreham, B. 2023. The Two Principles That Shape Scientific Research. Communicative & Integrative Biology. 16 (1), article 2203625.
- Cupps, V. R. Hijacking the Scientific Method. Creation Science Update. Posted on ICR.org July 31, 2014.
* Dr. Corrado earned a Ph.D. in systems engineering from Colorado State University and a Th.M. from Liberty University. He is a freelance contributor to ICR’s Creation Science Update, works in the nuclear industry, and is a Captain in the U.S. Naval Reserve.
