Statistically Speaking
Evan Zabawski | TLT From the Editor January 2015
Use caution when searching for statistics and data.
‘You can come up with statistics to prove anything. Forty percent of all people know that.’
--Homer Simpson
USING STATISTICS TO SUPPORT A CONCEPT OR AN OPINION is a fairly common practice. What sometimes makes a reader wary is that all too often a statistic may be strengthened by an omission or a false attribution to a reputable name or even simply fabricated.
Many of us have heard this phrase often attributed to Mark Twain: “There are three kinds of lies: lies, damned lies, and statistics.” It is from his book,
Chapters From My Autobiography, where he wrote: “Figures often beguile me, particularly when I have the arranging of them myself; in which case the remark attributed to Disraeli would often apply with justice and force: ‘There are three kinds of lies: lies, damned lies, and statistics.’” The common usage of this phrase is an example of omission since it neglects any mention of Disraeli.
This usage also is an example of a false attribution, as there is no recorded proof that British Prime Minister Benjamin Disraeli ever used it, and the earliest known usage appeared a decade after his death. My favorite example of a fabrication is from an episode of The Simpsons where Homer is being interviewed by news anchor Kent Brockman. When faced with some condemning statistics regarding his behavior as a vigilante, Homer replies: “Aw, you can come up with statistics to prove anything, Kent. Forty percent of all people know that.”
In this digital age, the prevalence of information does not seem to have curbed any of the aforementioned practices, so one must be cautious even when a Web search of a statistic has multiple hits. Once information has been published on the Internet, it is available to the masses instantaneously and is easily copied to other sites. Each transcription has the potential to alter the original message, like the proverbial telephone game.
The benefit of being able to access so much information is it can be easier to vet information by searching the citation, assuming it was correct the first time it was published on the Internet. Such a search for authentication was recently posited by STLE member Brian Schmidt of Chevron Global Lubricants. He asked STLE to validate the claim: “STLE and the National Research Council of Canada (NRCC) estimated that 82 percent of machine wear is particle induced.”
The easy part of the answer is that STLE does not fund research projects, but if one attribution is incorrect there is a chance that the second one is incorrect as well. To answer this question, I searched for other occurrences of the statistic and did find a citation that gave a publication, but the publication did not cite the original source. Another occurrence provided a table of supporting data, lending the statistic more credibility.
The next obvious path was to search the NRCC archives, which yielded no hits for any portion of the quote or any of the tabular data. My suspicion was growing that the entire claim could be incorrect, but the tabular data seemed too elaborate to be fiction.
Luckily, dogged determination from one of the NRCC librarians did locate the original paper publication, NRCC No. 26556, titled, “A Strategy For Tribology In Canada: Enhancing Reliability And Efficiency Through The Reduction Of Wear And Friction.” The study was published in 1986 and was sponsored by the National Research Council Canada/Associate Committee on Tribology. The data originally appeared in Table 2.18.
So for those seeking statistics to support their arguments, I urge caution to ensure the data is cited all the way back to the original document and not simply to the first Web hit. It is due diligence to ensure that nothing has been lost in translation, nor has a falsehood been propagated. Besides, let’s give credit where credit is due.
Evan Zabawski, CLS, is the senior reliability specialist for Fluid Life in Edmonton, Alberta, Canada. You can reach him at evan.zabawski@fluidlife.com.