Two papers in the same statistics journal, five years apart: the 1994 claim and the 1999 rerun
- Origin
- A statistics journal
- Dated
- August 1994; May 1999
- Extent
- 1994 paper 14 with appendix; 1999 paper 24 with supplementary experiments pages
- Status
- accepted
- Redactions
- None.
Chain of custody
Neither was obtained from anybody. Both have been publicly available since publication and both are free to read. This newspaper is filing them because the argument about them has been conducted for thirty years by people quoting the first one's headline probability, and the thing that actually decides it is a list of name-forms in the back of the first paper and a rerun in the second.
Held by: They Buried, document room; both mirrored publicly
Excerpt
[1994] …the randomization analysis showed the effect to be significant at a level of 0.00002. [1994, appendix] The name-form used for each rabbi. Thirty-odd entries. Every one defensible; none of them fixed before the analysis began. [1999] The result reflects the choices made in designing the experiment and collecting the data for it. Reruns with independently chosen name-forms detect nothing. The same procedure, tuned the same way, finds the same names in a Hebrew translation of a Russian novel.
Transcribed by the data desk. Line breaks follow the original. Bracketed matter is ours.
Editor's note
Filed as a matched pair because separately each is a paper and together they are a demonstration. The 1999 authors do not allege dishonesty and neither does this newspaper: the mechanism they describe works on people trying their best. The document desk notes that this file is the clearest argument in the vault for the thing this newspaper does with its own studies — write the list down first, seal it, and hand it to somebody else.
Cited in
- 296 Papers Accepted Before Anyone Saw A Result. Positives Fell To 39 Per Cent.Suppressed Science · Aug 22
- They Found The Same Hidden Names In War And Peace.Culture · Aug 10