<< Just say no to marijuana, kids | Main | The Racial Bias Lie >>

Deceit (or Reckless Disregard of the Truth) Continues in Nebraska Cost Paper

Thursday I noted a paper by Ernest Goss, et al. with a quotation that was purportedly from me but was completely false. I did not write those words, and the words report my study as finding the opposite of what I actually found.

If I received notice that a paper with my name on it contained a gross error, I would make the correction with scrupulous care and go over the correction with a fine tooth comb -- myself, not delegated -- to be very certain that the corrected paper was unimpeachably correct.  Evidently, Professor Goss does not share this view.  The "corrected" version remains a serious misrepresentation, either deliberately deceptive or with reckless disregard of the truth, which are morally about the same thing.
Here is the "corrected" version:

According to Alarcón, et. al., [Footnote 48: Arthur L. Alarcón and Paula M. Mitchell, Costs of Capital Punishment in California: Will Voters Choose Reform this November?, 46 Loy. L.A. L. Rev. S1 (2012). Accessed at http://digitalcommons.lmu.edu/cgi/viewcontent.cgi?article=2803&context=llr.]  "There is no credible evidence that replacing the DP with LWOP will result in significant added trial costs to the state due to defendants refusing to plead guilty and forcing prosecutors to meet their burdens at trial. The few studies that have been completed support the proposition that the threat of the DP does not increase plea bargain rates."
Although the mischaracterization of my study is now correctly identified as to its source, it remains a mischaracterization. In the Alarcón and Mitchell article, the first sentence in the above quote is followed by a reference to page 10 of my study, which, as explained in the previous post, says exactly the opposite.

My study found a plea bargain effect, but the effect did not reach the statistical rule of thumb known as "statistical significance." The fact that a study's result does not qualify for this somewhat arbitrary criterion most definitely does not provide evidence that the effect does not exist or is not significant in the usual sense of that word.  The only thing that "not statistically significant" means is that given the sample size and methods of the particular study we cannot say with 95% confidence that the result observed is a real one and not a random happenstance.

Translating statistics jargon into legal jargon, an effect that is observed but not "statistically significant" is analogous to a fact that has been proved by a preponderance of the evidence but not beyond a reasonable doubt. To cite such a result as disproving the fact in dispute is preposterous.  The fact that the notorious Paula Mitchell* says it despite the explanation to the contrary in the very paper she cites does not make it any less preposterous.

Professor Goss surely knows the difference between a study that observes an effect but does not qualify for "statistical significance" and a study that refutes the existence of the effect in question.**  For him to allow this egregious error to continue in the "corrected" version of his paper is gross negligence and academic malpractice at best. At worst, it is intentional deception.

In the preface to their paper, Goss et al. say, "The Goss & Associates research team thanks the staff of Retain A Just Nebraska for their assistance in facilitating the completion of this study. However, any conclusions, findings, errors or misstatements contained in this study are solely the responsibility of Goss & Associates, Economic Solutions, LLC."  I suspect it was those political advocates who initially made that egregious misquote, and I suspect it was the same advocates who made the grossly inadequate correction. But as the preface says, Goss and his team are responsible. Relying on political advocates for your research without checking it yourself is reckless disregard of the truth. That is particularly so with anti-death-penalty advocates, who are exceptionally notorious for intentionally misleading the public.

If he really accepts responsibility, Professor Goss should personally make sure that my work is characterized correctly in his paper and not delegate the correction to the political hacks who hired him.  Also, he told the Omaha World-Herald that he had apologized. I'm still waiting for it.

* The article in question is the last of a series, the first by Judge Alarcón and the later ones bearing the byline "Arthur L. Alarcón and Paula M. Mitchell."  Anyone who reads the series cannot help but be struck at the shift in tone across the articles, beginning quite reasonable and measured and ending very shrill and partisan. I strongly suspect that the articles become less Alarcón and more Mitchell as the series progresses.

** Indeed, Goss's own paper under discussion here does not give the outcome of any standard statistical tests regarding the result of his own multiple regression analysis. That is a red flag.  I have been told that Goss has admitted in a private email that his own results are not "statistically significant."  There is nothing wrong with reporting results that fall short of that arbitrary criterion, but some explanation is in order.

Leave a comment

Monthly Archives