Psychology

Human beings constantly form hypotheses.

• —“He is losing interest.”
• —“My boss doesn't respect me.”
• —“This project is going to work.”
• —“Something is wrong with me.”

A hypothesis is not a cognitive failure.

It is one of the ways the mind makes uncertain information manageable.

The difficulty appears when the distinction between:

“This is one possible explanation” and “This is what happened”

begins to disappear.

Confirmation bias does not require consciously deciding to ignore reality.

It can arise more subtly after an initial belief, interpretation or choice begins influencing the way later information is processed.

Before information can influence a judgment, we first have to encounter it.

Research on selective exposure suggests that people often show a preference for information that is consistent with their existing attitudes, beliefs or behaviours.

This does not mean people always avoid disagreement.

Accuracy motives matter too.

The tendency is probabilistic rather than absolute.

Accessible interpretation

You don't necessarily hide from the truth.

You may simply keep looking in the places where your current story is easiest to find.

Connection to Perspectives

Remember Alex.

If the interpretation is: “He's losing interest,” which moments from the previous week are easiest to retrieve and inspect?

• —The shorter reply?
• —The delayed message?
• —The distracted conversation?

The psychological question is not: “Is the interpretation wrong?”

It is: “Is the search for evidence balanced?”

Confirmation processes do not have to involve ignoring contradictory information.

Sometimes the information is noticed.

The asymmetry appears in how much influence it receives.

Experimental research has found that after making a choice, subsequent evidence consistent with that choice can receive greater weight than inconsistent evidence.

This suggests that an existing conclusion can influence how later information is integrated.

He cancelled our date. “I knew it.” He suggested another date himself. “Maybe he just feels guilty.”

• —Same person.
• —Same reasoning system.
• —Different standard of evidence.

This does not prove that the negative interpretation is wrong.

It demonstrates how supporting and contradicting evidence can be treated asymmetrically.

Confirmation bias has also been studied in relation to memory.

Recent research has examined confirmation-consistent processing across information search, evidence weighting and memory recall.

At group level, confirmation-consistent effects were observed across these domains.

However, the strength and reliability of effects differed across tasks.

What comes back to mind may participate in confirmatory processing, but the relationship between memory and confirmation bias is more nuanced than a simple rule that contradictory evidence disappears.

So the popular slogan — that people remember only what supports them — states more than the evidence does.

Prior beliefs are not inherently irrational.

Existing knowledge should influence how new information is interpreted.

A mind that abandoned every previous conclusion whenever new information appeared would not necessarily reason better.

The deeper issue is updating.

• —When should a prior belief remain stable?
• —When should confidence decrease?
• —When should the model change?

The critical-thinking question is therefore not: “Am I influenced by what I already believe?”

Almost certainly, yes.

The better question is: “Can my belief still change when reality stops supporting it?”

Not necessarily.

Rollwage and Fleming explored whether selective confirmatory processing can sometimes be adaptive when paired with effective metacognition.

Their modelling suggests that the consequences of confirmation bias depend partly on whether a person can accurately monitor confidence and recognize when additional information is needed.

This complicates the simplistic idea:

Confirmation bias = bad thinking

A prior belief can sometimes help organize information efficiently.

The danger increases when confidence becomes disconnected from the quality of the evidence.