Lesson 17 of 25 · 6 min

Confirmation, and the research read after the draw

After money is committed, the search for information quietly becomes a search for reasons, and the two produce different reading lists.

Confirmation bias is usually described as believing what one wants to believe. The more accurate description is narrower and more mechanical: it is an asymmetry in search. Supporting material is looked for harder, found faster, read further and remembered longer than the opposing kind.

A draw sharpens the asymmetry because it converts an opinion into a position. Before the draw the question is open. After it, disconfirming material carries an additional cost — it implies that a decision already taken and already funded was the wrong one.

The asymmetry shows up as a ratio. Suppose that in the six months after a $340,000 draw, 23 items about the holding are read, of which 19 make the case for it. That reading is 19 ÷ 23 = 82.6 percent favourable. If the available material were evenly split — an illustrative assumption, since no one counts — the sample overstates the favourable side by 32.6 points. The gap was produced by the selection, not by the evidence.

What the reading supports is a second draw. Collateral is $1,430,000 and the threshold is 65 percent. At a $340,000 balance the trigger value is $340,000 ÷ 0.65 = $523,077 and headroom is ($1,430,000 − $523,077) ÷ $1,430,000 = 63.4 percent. A further $215,000 takes the balance to $555,000, the trigger value to $853,846, and headroom to ($1,430,000 − $853,846) ÷ $1,430,000 = 40.3 percent.

Twenty-three points of cushion moved on the strength of a reading list that was 82.6 percent one-sided. The arithmetic of the second draw is exact; the input that authorised it was a sample assembled by the same person who wanted the answer.

Key terms

Confirmation bias: Searching for, favouring and recalling material that supports an existing view, while treating opposing material more sceptically.
Search asymmetry: The mechanical part of confirmation: supporting material is easier to look for, so more of it is found, independently of how much exists.
Disconfirming evidence: Material that would lower confidence if accepted. Its scarcity in a reading list is usually a fact about the search.

Worked example

A one-sided sample, and what it authorised

23 items read in six months, 19 of them favourable. Collateral $1,430,000, threshold 65% of market value.

Share of the reading that was favourable: 82.6%
Trigger value at a $340,000 balance: $523,077
Headroom before the second draw: 63.4%
Trigger value at a $555,000 balance: $853,846
Headroom after the second draw: 40.3%
Cushion moved: 23.1 points

The reading list has no error bars and the headroom figure has no opinions. They are connected by a single decision.

Check yourself

A $215,000 second draw takes a balance from $340,000 to $555,000 against $1,430,000 of collateral at a 65 percent threshold. How much headroom does it consume?

Saved in this browser only. No account, nothing sent anywhere.

Educational only. Nothing in this lesson is investment, legal or tax advice, or a recommendation to enter into any transaction. Figures are illustrative and computed from the inputs stated in the lesson.