Why You Only Believe the Evidence You Already Agreed With (Confirmation Bias)

You open your phone to prove someone wrong. You find four pieces of evidence that agree with you, and a fifth one that doesn't, which you don't read. Twenty minutes later you feel more certain than when you started, and you call it research. This video breaks down confirmation bias, the heaviest bias in this series, because it attacks the one thing every person believes they still have: the ability to reason fairly. This is Episode 4 of The Hidden Blueprint, closing out Phase 1, the phase about your own mind. We start with Peter Wason's 1960 rule discovery experiment, where participants given the sequence 2, 4, 6 almost never tried a sequence designed to prove their own theory wrong, and Wason's four-card selection task, one of the cleanest demonstrations of why people search for confirming evidence instead of the evidence that could break their theory. We cover Charles Lord, Lee Ross, and Mark Lepper's landmark study showing identical evidence handed to two opposing groups made both sides more polarized, not less, each side rating the same research as strong or flawed depending only on which direction it pointed. We break down a 2020 Nature Neuroscience study by Kappes and colleagues, showing that disagreeing opinions register with reduced sensitivity in the brain's posterior medial prefrontal cortex, meaning a strongly held disagreement and a weak one land almost identically. We explain why intelligence doesn't fix this, based on Stanovich and West's research testing fifteen separate measures of my-side bias. We cover Bakshy, Messing, and Adamic's 2015 Facebook study showing personal choices, not algorithms, primarily drive information bubbles, and Taber and Lodge's research showing people who read more opposing arguments often come away more confident in their original view. Finally, we correct a popular myth: that correcting false beliefs always backfires. Wood and Porter's 2019 research found it largely doesn't. Corrections usually work, they just produce