Data Core · Performance Analysis · Module

Reading the result

What the finding does and does not say, including the alternative explanations — the step where confidence is either calibrated or inflated.

The idea

How it works

The most common failure is not a wrong calculation but an over-confident reading. A correlation is presented, a mechanism is imagined, and the mechanism becomes the finding.

Caveats present in the analysis routinely disappear in the summary, and the summary is what people act on. Keeping them in the first paragraph is the only reliable defence.

Working with it

In practice

  1. 01

    List the alternative explanations

    Three, before concluding. If none can be excluded, the finding is a hypothesis.

  2. 02

    Report effect size, not just direction

    A real effect too small to matter is a common and expensive finding to act on.

  3. 03

    Keep the caveat in the summary

    A limitation on page four is a limitation nobody reads.

  4. 04

    State your confidence

    Explicitly, so the reader can weight it rather than assuming certainty.

One level in

The components of reading the result

A component is something that exists afterwards which did not exist before — a deliverable or a mechanism, not an intention.

  1. The finding

    What the analysis actually shows, with effect size.

    Learn
  2. The alternatives

    Other explanations for the same result, and which can be excluded.

    Learn
  3. The confidence

    How much weight the finding can bear, stated rather than implied.

    Learn

A caveat on page four is a caveat nobody reads. Put it in the first paragraph or accept it will be lost.