Anecdotes, Bias, and the Bigger Picture

19 Jul 2025 Christoff Truter Spousal caregiver

0

One of the most common and valid criticisms of a website like this is that it heavily relies on anecdotal evidence, meaning personal and subjective lived experiences. That is fair. This is, after all, a reflective journey centred around my wife’s health struggles.

I only have one data point to defend and can only speak for that data point, and I do not have access to vast pools of clinical data from which to conduct a proper study.

However, it would be a critical mistake to reject all anecdotal evidence simply because it is anecdotal. If a single data point or patient case contradicts the larger dataset it supposedly belongs to, that contradiction deserves investigation. It may point to blind spots in our understanding or indicate a possible misdiagnosis. (unexpected responses to medicine, symptoms that do not fit the variability narrative, progression)

When presented with a study, we should ask:

• What was the framework (parameters, bounds, etc.) and context used?
  Was the study designed to examine bacteria, but now being used to draw conclusions about viruses?
  
  
• How was the data collected?
  When studying the popularity of coffee vs tea, did they collect all their data at local coffee shops?
  
  
• What happened to outliers?
  Were outliers excluded because they contradicted the study’s conclusions?
  
  
• Was there an inherent bias in the study?
  Did they start with a conclusion and mine data to fit the narrative, or did the data guide the study?
  
  

  For example, a friend once claimed that 90% of men worldwide are circumcised. In reality, the rate is closer to 30-38%. Among Xhosa men in South Africa, the rate rises to 90–98%, and in Jewish or Islamic communities, it approaches 100%. While her statement was not false (in her context), her sample was biased, and her personal context skewed her perception.

  Research on circumcision has faced substantial bias, both for and against, e.g. cultural, The American Academy of Pediatrics 2012 report supported circumcision, while European medical bodies opposed it. This likely reflects cultural influence in the United States
  PubMed religious, Some members of the AAP task force were Jewish and circumcised themselves, which may have influenced their conclusions
  Matthew Tontonoz publication Studies showing positive results such as HIV prevention are more likely to be published than those showing neutral or negative outcomes
  arXiv funding, Many African circumcision studies are funded by organisations like the Gates Foundation, creating potential conflicts of interest
  JAIDS Journal ethical, Debates often focus on consent and bodily autonomy in infants, which can overshadow data-driven discussions
  Wikipedia data framing, African trials reported a 60 percent relative reduction in HIV risk, but the absolute reduction was closer to 1.3 percent, which changes the perception of benefit
  ResearchGate selection, Participants in African randomised trials were often volunteers motivated by free healthcare or social pressure, not randomly selected
  PMC confirmation biases. Advocates often emphasise supportive findings and ignore conflicting data or cultural harm
  Bond University

This line of questioning is especially relevant in light of the replication crisis Open Science Collaboration (2015). Estimating the reproducibility of psychological science. Science 349(6251): aac4716. Ioannidis, JPA (2005). Why most published research findings are false. PLoS Medicine 2(8): e124. Baker, M. (2016). 1,500 scientists lift the lid on reproducibility. Nature 533, 452–454. Munafò MR, Nosek BA, Bishop DVM, et al. (2017). A manifesto for reproducible science. Nature Human Behaviour 1, 0021. The Reproducibility Project: Cancer Biology (2017). Investigating reproducibility in preclinical cancer biology. eLife 6:e23383. , where many studies fail to reproduce when repeated.

FND studies (starting with the assumption that we are dealing with a neural communication disruption issue), often lack clarity because diagnostic criteria vary, case selection relies on subjective clinical judgment, and study designs focus on different symptom types Contemporary positive signs of functional limb weakness in post-acute sequelae of SARS-CoV-2: an exploratory analysis of their utility in diagnosis and follow-up. Read more. . This inconsistency and limited methodological detail make many FND research formulations somewhat obscure and hard to compare. I would love to have access to more raw data to analyse.

At the end of the day, my wife's outlier data might not even be an outlier at all, but rather a result of misdiagnosis and therefore not relevant to the functional dataset.