In this experiment, suggested by 13 year old David Roffman in the closing months of 2005,
we employ a methodology by which a proper test of the Null hypothesis
of no Torah effect can be done against a complex of
alternative hypotheses: that almost all the events
in one or two of the their key word sets have ELSs that are
in a more compact relation than expected by chance.
We apply this test to the list of American presidents.
The American president experiments is driven by a ruled based spelling
transliteration scheme of English names into multiple alternative Hebrew spellings.
This methodology is more immune to claims of
non a priori data selection because it is rule based
and because the rules generate many spelling alternatives all of which are used
in the experiment. The rule based scheme itself becomes incorporated into
the alternative hypothesis.
With our new
methodology, the Null Hypothesis is tested against the alternative that
nearly all presidents have one or two spellings that have ELSs in a more compact
meeting than expected by chance. The Null Hypothesis had to be rejected with a
p-level below 1/100,000.
American Presidents English Hebrew Transliteration Rules
American Presidents Input Data
Post Experiment Analysis coming