Confirmation Bias in Decision Making: How to Test Your Own Logic

A few years ago, I spent three weeks researching a business idea. I read case studies, talked to potential customers, studied competitors, ran the numbers. By the end of week three, I had a tidy document that proved the idea was worth pursuing.

The problem was that I had wanted it to work before I started researching. And every piece of evidence I gathered had been quietly filtered through that wanting.

The customer interviews? I remembered the enthusiastic ones and downplayed the skeptical ones. The competitor analysis? I focused on their weaknesses, not their advantages. The financial model? I chose optimistic assumptions for growth and conservative assumptions for cost. Each step felt like analysis. In total, it was a performance.

This is confirmation bias: the tendency to seek, interpret, and remember information in ways that confirm what you already believe. It is one of the most studied effects in behavioral psychology, and one of the hardest to catch in yourself, because it does not feel like bias. It feels like thinking.

That is what makes confirmation bias especially dangerous in decision making. Unlike impulsive choices, where you skip the analysis entirely, confirmation bias sits inside the analysis. You do the research. You weigh the evidence. You follow a process. But the process has a hidden filter, and the filter only lets through what supports the answer you already prefer.

Here are four examples of how it works in practice.

Hiring. A manager reads a resume and forms a first impression within seconds. The interview that follows is not really an evaluation. It is a confirmation exercise. Questions are chosen to validate the impression, not to challenge it. The candidate who "felt right" gets softballs. The one who did not gets stress tests.

A product idea. A founder falls in love with a concept and launches a survey. The survey is designed, without deliberate intent, to produce encouraging results. Leading questions. A sample biased toward early supporters. When a few responses are negative, they get labeled as "not the target audience." The data says yes because the data was asked to say yes.

Investing. You buy a stock and then start reading articles about it. Positive analysis feels credible. Negative analysis feels uninformed. You are no longer researching the investment. You are defending it. The deeper the position, the stronger the filter.

A workplace disagreement. Two teams interpret the same quarterly data differently. Each side finds patterns that support their strategy and dismisses the patterns that do not. The data has not changed. The lens has. Both teams believe they are being objective, and both teams are wrong in the same way.

The deeper problem is not that we are biased. It is that confirmation bias in the workplace and in personal decisions looks identical to careful thinking. You cannot feel it happening, because it hijacks the feeling of being rational.

So what actually works?

Inversion. Instead of asking "What evidence supports my preferred option?" ask "What evidence would prove me wrong?" This single question changes the direction of the search. You stop building a case and start testing one.

The difference matters. Building a case is advocacy. Testing a case is thinking.

Here is a short checklist you can use before committing to a decision:

1. Write down the option you currently prefer.
2. List three pieces of evidence that support it.
3. Now list three pieces of evidence that would change your mind.
4. Ask: did I search as hard for the disconfirming evidence as I did for the confirming evidence?
5. If you cannot find disconfirming evidence, that is a warning, not a green light.

This is closely related to the way confirmation bias distorts how we read data. But where that problem sits in spreadsheets and research, this one sits in everyday decisions: which job to take, which strategy to fund, which person to trust. The filter is the same. The stakes are personal.

It also connects to hindsight bias. Once a decision works out, confirmation bias rewrites the story so the evidence looks cleaner than it was. And when you are overthinking a decision, the loop often is not more thinking. It is the same biased thinking, repeated.

Second-order thinking helps here too. Even if the confirming evidence is real, what happens after the first result? A decision that looks right on the surface may create consequences that the confirming evidence never addressed.

The uncomfortable truth is that wanting to be right and wanting the truth are not the same thing. They feel identical from the inside. The only reliable difference is what you do when the evidence disagrees with you.

If you dismiss it, you are advocating. If you sit with it, you are thinking.