Never Trust the Science - On the need to identify bias & interpret data yourself
6 hours ago
- #science-bias
- #critical-thinking
- #media-misinformation
- Science begins with questions, but funding bias can limit which questions are explored, favoring politically acceptable inquiries over others.
- Methodological choices in research can introduce bias, such as selecting less precise measurements to favor null hypotheses or adjusting for confounders based on desired outcomes.
- Statistical analysis introduces further bias through the choice of tests and adjustments, which can subtly influence results based on preferences.
- Interpretation and publication bias occur when data is selectively presented or rejected based on alignment with reviewers' or editors' biases.
- Media reporting often misrepresents scientific findings, stripping context, exaggerating claims, or ignoring inconvenient results, leading to widespread misinformation.
- A case study on vegetarian diets and cancer risk highlights how media and researchers can misrepresent observational data as causal, ignoring significant limitations and inconvenient findings.
- The study's only robust finding, after adjustments and sensitivity analyses, was that vegetarians had a higher risk of oesophageal cancer, contrary to the media's narrative.
- Critical reading of scientific papers, understanding statistical methods, and recognizing biases are essential to navigate and trust scientific information.