Figure 218 Oct 2026From “What is publication bias?”

A test register

Figure 2A test register

Source: Author's decision worksheet. The register as a known population follows Franco et al. (2014), pp. 1 and 2 of the Science Express version; the planned size field follows their note that TESS requires adequate statistical power (p. 1), and the effect sought beside the size is this guide's addition; the write-up and report fields follow their finding that most null results were never written up (pp. 1 and 2, Table 3 on p. 6) and Munafò et al. (2017), p. 3; the reason field follows the authors' explanations Franco et al. report (p. 2). The fields are reader inputs; no company data is supplied.

PNG4:3, 1600 × 1200, with the title, the source line and the attribution set in the image isoglu-2026-what-is-publication-bias-fig02.png

Source and evidence

Each line names a claim related to this exhibit, its source, and how it was checked.

B PUB-OWN-C3 In a pricing or sales analytics team, a review that shows only tests with significant results is a dashboard of winners and carries the selection Franco et al. found at the write-up, a step the analyst controls and which, as in TESS, can anticipate what the review is expected to want; Sterling's chain can run inside a team when a test with a null result leaves no note and is rerun until a chance success is reported; a register of every test made before it starts makes the team's file drawer countable, so Rosenthal's question, how many hidden studies it would take to overturn a result, need not be asked, because the number of tests left out is known (Rosenthal computed a threshold, not an estimate of any drawer); Figure 2 records the question and registration date, the planned size with the effect sought (power is defined for a stated effect, so a size alone cannot show whether a null could have detected the effect; recording the effect is this guide's addition), the result as strong, mixed or null against the plan, the write-up, the review that shows it and the reason it was left out; illustrated with a hypothetical 24 price tests, each with its question, sample, run length and smallest effect worth finding fixed before the start, 7 significant, all shown by result with notes on the nulls; a null from a test too small for the effect sought says little either way; the mapping is untested.Author mapping grounded in STE59-C2, FMS14-C2, FMS14-C4, ROS79-C2 and MUN17PB-C1; untested

Grades: A, verified against the printed page of the primary source · B, primary source, text layer only · C, authoritative secondary · D, reported.

From the essay

From the research bench

What is publication bias?

Publication bias is when a study's result decides whether it reaches print. In one known set of studies, most null results were never even written up.

18 Oct 2026

Read the essay

This exhibit belongs to an essay: the numbers live here, the argument lives there.

Cite or embed

APA 7

Isoglu, S. (2026). A test register [Figure]. isoglu.com. https://isoglu.com/insights/exhibits/isoglu-2026-what-is-publication-bias-fig02/

BibTeX

@misc{isoglu-2026-what-is-publication-bias-fig02,
  author = {Isoglu, Sinan},
  title = {{A test register}},
  year = {2026},
  howpublished = {Figure 2, isoglu.com},
  url = {https://isoglu.com/insights/exhibits/isoglu-2026-what-is-publication-bias-fig02/},
  note = {From the essay {What is publication bias?}}
}

Embed this exhibit

Copy this HTML to use the PNG exhibit on your site with its source line and a link back to the exhibit page.

<figure>
  <a href="https://isoglu.com/insights/exhibits/isoglu-2026-what-is-publication-bias-fig02/">
    <img src="https://isoglu.com/social/isoglu-2026-what-is-publication-bias-fig02.png" alt="Figure 2: A test register" width="3200" height="2400">
  </a>
  <figcaption>Figure 2: A test register. Source: Author&#39;s decision worksheet. The register as a known population follows Franco et al. (2014), pp. 1 and 2 of the Science Express version; the planned size field follows their note that TESS requires adequate statistical power (p. 1), and the effect sought beside the size is this guide&#39;s addition; the write-up and report fields follow their finding that most null results were never written up (pp. 1 and 2, Table 3 on p. 6) and Munafò et al. (2017), p. 3; the reason field follows the authors&#39; explanations Franco et al. report (p. 2). The fields are reader inputs; no company data is supplied. <a href="https://isoglu.com/insights/exhibits/isoglu-2026-what-is-publication-bias-fig02/">View exhibit</a>.</figcaption>
</figure>