But what constantly has me puzzled is how even people who are 100% convinced all three were guilty can read that 'confession' and not see that something isn't right with the techniques used.
Well I've yet to come across such a person, let alone could I speak for them, but I'm far from impressed with how Gitchell and Ridge handled the interview. Of course reality never reaches perfection though, and in a podunk town like West Memphis few things come anywhere close. I mean you couldn't rightly expect the plods in Dubbo to handle such a situation notably better, could you?
Regardless, your bringing up McClish while I was vacationed until this particular date for my lack of interest in mincing words is rather serendipitous. I stumbled across his page on Misskelley's initial confession back when I was first starting to take interest in this case, and since at that time I wasn't nearly familiar enough with the body of evidence to judge his conclusions regarding it, I checked his list of famous cases for a topic I had previously researched extensively and found
he'd touched on the murder of JFK which occurred half a century ago on this day. I was rather perplexed to see Lee Harvey Oswald on the list though, being well aware of the fact there's scant actual documented statements from him between the time of JFK's murder and his own to work with, mostly just notes, reports, and testimony from those who were around him along with a few short comments recorded by the press. So I clicked the link with great curiosity, and was shocked to find McClish using Mae Brussell's The Last Words of Lee Harvey Oswald, which is essentially a work of historical fiction. McClish does note Brussell's warnings in that regard, but promptly proceeds to throw that caution to the wind and nitpick at variations in terminology such as gun/pistol and movie/picture show absent documentation of Oswald actually having used any of those particular words. Of course Oswald had plenty of secrets, but McClish's attempt to divine them from a compilation of what Brussell took him to have said based on the accounts of others falls flat on its face. So at that point I quickly lost interest in anything McClish had to say, which concluded my research into statement analysis until last night.
Since the topic of statement analysis came up here though I figured I might as well look into the matter more, and in that regard I recommend the last section of chapter 5 in
Language of Confession, Interrogation, and Deception and chapter 10 of
Detecting Lies and Deceit. Also, Richard Leo, a colleague of Richard Ofshe at Berkeley who many are quite fond of here, wrote more recently on the topic in
Police Interrogation and American Justice, referencing the aforementioned works among others and harshly concluding:
But statement analysis is just another form of junk science. Despite its name, there is nothing scientific about so-called Scientific Content Analysis. Statement analysis in general and SCAN in particular are theoretically vague, if not vacuous (miller and Stiff, 1993; Shey, 1988; Shearer, 1999) The unstated theory of statement analysis appears to be that deceptive and and truthful people use different types of memory (i.e. real and nonexistent). Even if that is true, however, there is no reason to believe that this alone would cause a deceptive and truthful individuals to write and speak different, that they would do so in the specific ways that SCAN suggest, or that third parities could infer truth and deception from the speech patterns or language changes in a person's prewritten narrative independent of any case facts or knowledge of the individual's background or history.
If the theory underlying statement analysis is based on little more than speculation, the empirical evidence for its claims is no better. Simply put, there is no empirical evidence for SCAN (Porter and Yuille, 1996; Shuy, 1998), just endless post hock illustrations and testimonials by Sapir and his former students (
www.lsiscan.com, 2004; Scientific Content Analysis Handbook, 1990). Nor is there good reason to believe that there will be any validation for SCAN or statement analysis any time soon (Miller and Stiff, 1993; Shuy 1993; Shearer, 1999). As Ruger Shuy. (1998; 75) has pointed out "the accuracy of the detection of Deceitful language is... at about the level of chance." Moreover, SCAN completely ignores linguistic research on how people talk (Shuy, 1998). As with other Behavioral Methods of Lie Detection analyzed above, the value of SCAN and statement analysis lies simply in its utility as an interrogation technique.
Granted, I've been unimpressed by Leo and Ofhse's claims of being able to divine lies from truth as well, not just in this case but also the other eight examples discussed throughout
The Guilty and the "Innocent": An Examination of Alleged Cases of Wrongful Conviction from False Confessions, with
The Facade of Scientific Documentation: A Case Study of Richard Ofshe's Analysis of the Paul Ingram Case providing further detail on another of those cases. Also, Reedus recently directed me to
this appeal ruling which adresses Ofshe's attempts to explain away a confession of Larry Dewayne Hall, an
admitted serial killer who has remained behind bars thanks to evidence used to convict him for one count of kidnapping which resulted in the death of 15 year old Jessica Roach, the charge he'd confessed to but which Ofshe attempted to explain away. So yeah, while I consider techniques used by law enforcement in this case far from ideal, they come out looking like Sherlock Holmes when contrasted with what I've seen from proponents of statement analysis and other techniques for evaluating the veracity of a suspect's words with little to no regard for the body of evidence surrounding them.