Cherry Picking Data Is the Pits
by James J. S. Johnson, J.D., Th.D. *
Picking cherries sounds like good, clean fun, but it isn’t always a good thing. In origins science the practice of “cherry picking” refers to an analytical fallacy—using a few selected illustrations to demonstrate a point, as if those examples fairly exemplify a generalized trend when they actually don’t.
Illustrations can’t prove universals. So, are all illustrations misleading? No. Illustrative examples by themselves are not misleading—unless and until those examples are suggested as representing “all” or “most” or “generally” when those qualifiers don’t fit the facts.
Two illustrations follow that exemplify this fallacy. One quotes a federal court decision, the other involves comparing the genomes of humans and chimpanzees.
Rejecting Courtroom Testimony
A trial judge should reject an expert witness’ testimony if it demonstrates untrustworthy cherry picking.1 In a lawsuit involving discrimination-in-employment allegations, a government witness named Kevin Murphy presented a statistical analysis of a dataset to portray alleged trends in personnel hiring practices. But Murphy’s data summaries painted a distorted picture because inconvenient data were ignored—a classic case of cherry picking.
Murphy cherry-picked the data that he included from outside the time period covered by the Excel spreadsheets….Murphy deliberately ignored the data that was available from the relevant time period in favor of data from outside the time period.2
The federal judge noticed that the relevant database included 58,892 total examples, yet Murphy selected only about 2,000 examples for his trend analysis. “In an egregious example of scientific dishonesty, Murphy cherry-picked certain individuals…in an attempt to pump up the number of ‘fails’ in his database…conveniently increas[ing] the fail percentage by over twenty percent, rendering it a meaningless, skewed statistic.”2
Was Murphy’s testimony accepted at trial? No. Murphy’s report was rejected as incompetent and thus unreliable.
The judge concluded, “The story of the present [litigation] has been that of a theory in search of facts to support it.”3
But fighting the facts appears out of court, too. It also happens in the arena of origins science because Darwin’s natural selection theory is likewise “a theory in search of facts to support it.” (See Romans 1:18, 28.)
Rejecting Genome Comparisons
Do you remember hearing or reading that supposedly 98% of the human genome matches that of chimpanzees, due to humans and chimpanzees allegedly sharing common ancestors?
Those cherry-picked genome comparison reports suppressed huge categories of contrary evidence. As ICR’s Dr. Jeffrey Tomkins has reported, “All analyzed cases of reported high human-chimp DNA sequence similarity are based on biased data selection and exclusion techniques.”4 Using “highly selective data-discarding techniques,”4 such “statistical” results failed to account for major disparities between the respective genomes.
For the primary chimp chromosomes (autosomes), the amount of optimally aligned DNA sequence provided similarities between 66 and 76 percent, depending on the chromosome. In general, the smaller and more gene-dense chromosomes showed higher DNA sequence similarity—although there were several notable exceptions. Only 69 percent of the chimpanzee X chromosome (female sex chromosome) and only 43 percent of the Y chromosome was similar to human DNA. Genome-wide, only 70 percent of the chimpanzee genome assembly was similar to human DNA under the most optimal sequence-slice conditions.5
Not even close to the evolutionists’ trumpeted 98%! That’s because cherry-picked analyses are unreliable misrepresentations of the true facts—in this case, the real truth is that man was uniquely created in God’s image and is not just a chimpanzee’s cousin.
- EEOC v. Kaplan Higher Learning Education Corp., 2013 WL 322116, at *10-*11 (N.D. Ohio 2013). See also Federal Evidence Rules 104, 107, 702.
- EEOC v. Freeman, 961 F.Supp2d 783, 795 (D. Md. 2013), affirmed, 778 F.3d 463, 471 (4th Cir. 2015). (“Despite Murphy’s record of slipshod work, faulty analysis, and statistical sleight of hand, the EEOC continues on appeal to defend his testimony.”)
- EEOC v. Freemen, 961 F.Supp.2d at 803.
- Tomkins, J. 2012. Journal Reports Bias in Human-Chimp Studies. Acts & Facts. 41 (6): 6. From a forensic perspective (in light of Federal Evidence Rule 901(b)(3)), it was even worse that those so-called “comparisons” excluded quantitatively meaningful analysis of the large portions of human DNA to which there is no chimpanzee alignable genome counterpart, and vice versa.
- Tomkins, J. 2013. Chromosome Comparison Shows More Chimp-Human Differences. Acts & Facts. 42 (2): 9.
* Dr. Johnson is Associate Professor of Apologetics and Chief Academic Officer at the Institute for Creation Research.
Cite this article: James J. S. Johnson, J.D., Th.D. 2015. Cherry Picking Data Is the Pits. Acts & Facts. 44 (7).