According to Michael Dahlstrom, Ph.D., who teaches science communication at Iowa State University, “Narratives follow a particular structure that describes the cause-and-effect relationships between events that take place over a particular time period that impact particular characters. This inevitability, combined with the lack of a need for justification, supports the many normative elements with a story — what is good, what is bad — without ever needing to clearly articulate or defend them. Because narratives are able to provide values to real-world objects without argument, it is difficult to counter their claims.”
This is one reason why anecdotes are more appealing than logical, scientific facts. While quantitative scientific measures are more accurate and less wrong than personal stories that account for a single event, we are likely to believe and trust stories than abstract, dry data.
For example, some people argue that smoking does not cause early death because their grandparents had smoked several packs of cigarettes a day and lived to be in their eighties or nineties. They believe that their personal experience and example are more reliable than decades of data that show a strong correlation between long-term smoking with a higher risk of early mortality.
When we put these into a narrative, it makes much more sense: I'm hungry so my girlfriend and I are going to eat at a nice restaurant at 6pm.
See how narrative is automatic and connects to a very basic level of the brain?
Anecdotal evidence (e.g. “but it worked for me!”) is fine for an individual or personal decision-making level. But when issuing recommendations, selling or promoting services, or treating the general population, the alleged effects and benefits have to be objectively shown to be generalisable and repeatable over large samples at a level significantly above chance effects and placebo effects — if the stakes are higher than relaxation and self-care (which are perfectly good goals for healthy people!).
If a person is trying to solve a health or medical/physiological problem, the goal is to converge on an effective solution as efficiently as possible, which means starting with the most probable and/or proven methods first. Otherwise, therapists and trainers can waste a lot of time chasing around after n=1 (but it worked for me!) dead ends.
Anecdotal evidence can help us make low-risk decisions, such as buying a different brand of shampoo or watching a new foreign movie. But if you have cancer, diabetes, or other disease, relying exclusively on anecdotal evidence can be fatal. Most situations fall somewhere in between, and an ethical practitioner should be be mindful of the gray areas.