Bad Science

[rat­ings]
Posted with per­mis­sion of the author.

BAD SCIENCE
by D. H. Garrison, Jr.
Forensic Services Unit
Grand Rapids Police Department
Grand Rapids, Michigan
KEYWORDS: Crime Laboratory, Ethics, Forensic Science, Police Ethics, Prosecutorial Misconduct

——————————————————————————–

Forensic sci­ence is the prod­uct of an uneasy and unholy mat­ing of Science, the objec­tive seeker of truth and knowl­edge, and Forensics, the argu­men­ta­tive per­suader of court­room advo­cacy. It is not called Justice Science, Law Science, or Truth Science, as many of us would like to imag­ine. We are a bas­tard child, an orphan, but still the sub­ject of an intense child cus­tody bat­tle between our estranged par­ents, the truth seeker and the advo­cate. The tug-of-war goes on daily for our loy­al­ties and con­fi­dences, each side offer­ing candy and warm hugs. These sep­a­rated par­ents have vis­i­ta­tion rights. Sometimes they take our broth­ers and sis­ters away. Sometimes they don’t come back.

We in foren­sic sci­ence like to think of our­selves as our mother’s child–Mother Science, pure and incorruptible–and most of us start out this way. Some of us remain pure. Some grow up to be delin­quents. The advo­cacy half of foren­sic sci­ence will not go away; it has week­day vis­i­ta­tion rights and the power-of-subpoena. It has advo­cate friends called pros­e­cu­tors, attor­neys, cops, the press, and the Government. The advo­cates rarely under­stand the appeal of Mother Science, can­not fathom a search for truth in a game plan which calls for scores and tro­phies. They are con­stantly try­ing to per­suade us to see it their way, to com­pro­mise, to bend just a lit­tle. They don’t real­ize it, but what the advo­cates are ask­ing for is Bad Science.

The pres­sure to be a Bad Scientist, to fit in and go along, is great, and it doesn’t go away unless you put your foot down and say Enough Is Enough! And keep say­ing it to each super­vi­sor, each detec­tive, and each fair-haired boy from the prosecutor’s office. Bad Science is what foren­sic sci­ence becomes when an attor­ney or pros­e­cu­tor, who often dis­play all the ethics of a full-grown ham­ster, get a foren­sic sci­en­tist to play ball, to get with their pro­gram and see their big pic­ture. There is an old Bad Science joke about a sci­en­tist who was work­ing with an ant. The sci­en­tist would cut off one of the ant’s legs and shout, “Jump!” And the ant would jump. The sci­en­tist cut off a sec­ond leg, told the ant to jump, and again the ant jumped. And so it went, until the sci­en­tist had cut off all six of the ant’s legs. This time, when told to jump, the ant did not jump. This proves it, the sci­en­tist con­cluded: when you chop all the legs off an ant, the ant goes deaf!

You may rec­og­nize some scenes from the fol­low­ing exam­ples of Bad Science at Work. Some are laugh­able, oth­ers dis­turb­ing. Some sim­ply haven’t hap­pened to you yet. I have not per­son­ally encoun­tered all of these sit­u­a­tions, but I know that each is true. If you haven’t wit­nessed at least some of them, you will. If this helps you steel your­self against the onslaught of the Advocates, so be it. Finally, not all Advocates are mali­cious. Many, in fact, are sim­ply not versed in the ways of good sci­en­tific method. When they ask for Bad Science, you can pity them as help­less peo­ple doing the wrong thing for the right rea­son. This type of Advocate needs to be taught … and watched.

Misinterpretation of Test Results
In a rob­bery case the vic­tim, a bar­tender, tes­ti­fies that the defen­dant had come into the tav­ern ear­lier in the night for a glass of beer. Three unwashed glasses were found at the scene and were processed for latent prints. Two of the glasses yielded prints, but these were of per­sons unknown, not the defen­dant. The pros­e­cu­tor sug­gests that the print exam­iner tes­tify that the third beer glass must have been used and then wiped clean by the defen­dant, because the other two glasses were obvi­ously not his. The print exam­iner sug­gests that the pros­e­cu­tor look else­where for this kind of tes­ti­mony. The pros­e­cu­tor looks surprised.

Manipulation of Raw Data
An acci­dent recon­struc­tion expert with a com­puter is hired by a plaintiff’s attor­ney to deter­mine the speed of the defendant’s vehi­cle in a two-car col­li­sion. The expert enters into his com­puter pro­gram the road sur­face drag fac­tor, skid and yaw mark lengths, and the loca­tion and sever­ity val­ues of the vehi­cle dam­age. The first run of his com­puter pro­gram pro­duces an unre­al­is­ti­cally high speed for the defendant’s strik­ing vehi­cle. The expert changes his drag fac­tor esti­mate and tries again. The fig­ures are still out­ra­geous. Three pro­gram runs and sev­eral crush data changes later, the speed deter­mi­na­tion begins to look more believ­able. The defendant’s attor­ney begins his attack with a sub­poena for all five of the expert’s com­puter printouts.

Jeopardy
As in the tele­vi­sion game show where con­tes­tants reply in the form of a ques­tion, cer­tain man­agers give their sub­or­di­nates a desired answer and demand that they come up with the appro­pri­ate research ques­tions to sup­port it. During one police department’s trial period of a 9mm pis­tol, a police offi­cer wounds an assault sus­pect. Because the sus­pect was not instantly inca­pac­i­tated, the police chief scraps the entire 9mm changeover pro­gram. He hears of the FBI’s 10mm pis­tol pro­gram. One of the the­o­ries he returns with states that, by virtue of its “larger size,” the 10mm is much bet­ter at strik­ing blood ves­sels than the smaller 9mm bul­let. The department’s shoot­ing instruc­tor points out that en extra half-millimeter along on each side of the 10mm bullet’s diam­e­ter would not really make much dif­fer­ence, unless you missed a blood ves­sel by half a mil­lime­ter with a 9mm bul­let. Then the instruc­tor begins his litany about the train­ing bud­get, that train­ing is at least as impor­tant as hard­ware, but the admin­is­tra­tor doesn’t hear him, because it’s time to play Double Jeopardy with the police chief.

Comparing Apples and Orangutans
In a prod­uct lia­bil­ity suit, the plaintiff’s attor­ney finds an expert wit­ness who will tes­tify that, if the shot­gun involved in the shoot­ing had as safe a fir­ing mech­a­nism as a rivet gun, the inci­dent may not have hap­pened at all.

Manipulation of Test Results
During a bur­glary trial, the pros­e­cu­tion pro­duces seven latent prints recov­ered from inside the victim’s house. The fin­ger­print exam­iner tes­ti­fies that he has iden­ti­fied these prints as belong­ing to the defen­dant. The pros­e­cu­tor sug­gests that the fin­ger­prints are like seven lit­tle pho­tographs of the bur­glar inside the house. Because he does not want a repeat of an ear­lier case lost to the defense attor­ney, the pros­e­cu­tor calls a sec­ond exam­iner to the stand to ver­ify the com­par­i­son per­formed by the first. The pros­e­cu­tor then states that the seven latent prints, times two print exam­in­ers, make for four­teen lit­tle pho­tographs of the defen­dant inside the crime scene. Later, when jok­ingly asked why he didn’t call a third exam­iner to up the score to twenty-one fin­ger­prints, the pros­e­cu­tor replies that he had sim­ply neglected to sub­poena another print examiner.

Compulsive Computing
A .223 Remington bul­let is found lodged in a house sev­eral hun­dred feet to the rear of a rifle prac­tice range at which .223 weapons are fre­quently fired. The inves­ti­ga­tors want to know if it is pos­si­ble for a .223 bul­let to fly the sev­eral hun­dred feet nec­es­sary to reach the house, so they ask a firearms exam­iner. The exam­iner, who had recently invested in a bal­lis­tics pro­gram for his home com­puter, took down the range, wind speed, bul­let shape, tem­per­a­ture, baro­met­ric pres­sure, and sev­eral other pieces of data. His com­puter charted the results. Finally, his answer to the inves­ti­ga­tors was, “Yes, it’s pos­si­ble.” As a qual­i­fied firearms exam­iner, he had already known that the house was well within the range of the .223 car­tridge and could have given the same answer when first asked the ques­tion … with­out computation.

Denial
In many major crim­i­nal inves­ti­ga­tions it is the prac­tice of a detec­tive unit to offer poly­graph exam­i­na­tions to the sus­pects and, in cases of ques­tion­able accu­sa­tions, to the vic­tims. While they are not admis­si­ble in court, the poly­graph results are relied upon as a valid inves­tiga­tive tool. One day a young police offi­cer shoots and wounds a juve­nile who he claims fired at him first, although no weapon is found. The offi­cer claims he was also struck sev­eral times about the head and shoul­ders with a board prior to the shoot­ing, although he exhibits no bruises, head injuries, or defense injuries to his hands or arms. When asked about this lack of con­sis­tent injuries, a detec­tive reports that the young offi­cer was wear­ing a bullet-resistant vest. The detec­tives do not offer the sus­pect or the offi­cer a poly­graph exam­i­na­tion in this par­tic­u­lar case.

Ethical Bankruptcy
In a homi­cide case the pros­e­cu­tion demon­strates a laser recon­struc­tion of a bullet’s path through a woman which indi­cates her hus­band fired a rifle from his shoul­der height. The husband’s story is that he was clean­ing the weapon while it lay on a table­top. The defense attor­ney finds a firearms expert who will claim that, while the weapon was not mal­func­tion­ing before the inci­dent, was not mal­func­tion­ing when col­lected from the crime scene, and is not mal­func­tion­ing now at the time of trial, it may have sud­denly mal­func­tioned and fired all by itself as a result of a buildup of dirt and pow­der within the weapon’s mech­a­nism on the day of the shoot­ing. The expert does not address the issue of the shoot­ing recon­struc­tion, but the jury does and returns a guilty verdict.

No Scientific Methodology
A city truck dri­ver runs a stop sign and causes an acci­dent with seri­ous injuries. Instead of rely­ing on the skid­marks, crush dam­age, and scene evi­dence, the city author­i­ties order a traf­fic inves­ti­ga­tor to con­duct accel­er­a­tion tests to deter­mine the max­i­mum pos­si­ble speed the truck could have achieved in the one-block dis­tance lead­ing up to the crash. Because the truck involved was dis­abled in the acci­dent, the traf­fic inves­ti­ga­tor uses a motor­cy­cle to run the one-block accel­er­a­tion test and reports back a peak speed of 35 miles-per-hour for the city truck.

Too Many Cooks Spoil The Broth
A city bus rear-ends and crushes a car­load of teenagers, killing four. The first traf­fic inves­ti­ga­tors at the scene mea­sure the skid­marks of the bus and deter­mine that the bus dri­ver was speed­ing. A national civil rights leader says the bus dri­ver is being made a scape­goat by the city solely because he is a racial minor­ity. The follow-up inves­ti­ga­tion by city author­i­ties reports that the orig­i­nal traf­fic inves­ti­ga­tors, who have been abruptly removed from the case, must have been mea­sur­ing tire marks tracked through melted road­way tar and that, on sec­ond thought, the city bus dri­ver was not really speed­ing. A local tele­vi­sion sta­tion gets a radar gun and reports that most dri­vers, includ­ing all city bus dri­vers, reg­u­larly exceed the speed limit on this sec­tion of road. Tire tracks in tar look noth­ing like skid­marks to the trained eye of the traf­fic inves­ti­ga­tor. Excessive speed aside, it is unlaw­ful to fol­low another vehi­cle at an unsafe dis­tance in that state.

Pursuit of the Inconsequential
In the faked rob­bery of a fast food restau­rant, the night man­ager shoots to death an employee in a walk-in cooler, hides the “stolen” money and a .357 Magnum revolver, and calls the police. The crime scene per­son­nel notice fallen dust on a restroom floor and dis­cover the money hid­den in a ceil­ing panel. The revolver is found among the night manager’s pos­ses­sions. In prepa­ra­tion for trial, the pros­e­cu­tor asks for a shoot­ing sound test to be done inside the restaurant’s cooler. This, he says, will deter­mine whether or not the fatal shots could have been heard by a teenage girl who was hav­ing sex with a man (not her boyfriend) in her boyfriend’s van parked across the street from the restau­rant. The girl, who inci­den­tally have a full-length cast on her leg at the time (another mys­tery alto­gether), did not recall hear­ing much of any­thing, least of all gun­fire. Her part­ner that night also some­how missed the sounds. The crime scene inves­ti­ga­tor refused to par­tic­i­pate in such an exper­i­ment, argu­ing that it was invalid, irrel­e­vant, and silly … and what would it prove any­way? The pros­e­cu­tor sug­gested that the defense might use the fact that the girl had not heard the shots to argue that the time of the mur­der was some­how dif­fer­ent. “Then let the defense make a sound test,” the inves­ti­ga­tor says, leav­ing. The pros­e­cu­tor is insis­tent. After being turned down by the police firearms trainer and the state regional lab­o­ra­tory exam­in­ers, the pros­e­cu­tor gets three detec­tives to fire the shots for the sound test. To dupli­cate the sounds of a .357 Magnum, they load the weapon with light .38 Special tar­get loads; they fire the qui­eter ammu­ni­tion into a sand­bagged pipe inside the walk-in cooler so as not to make holes in the walls. It is sev­eral months later, and the air tem­per­a­ture is sixty degrees lower than the night of the mur­der. By the time the test begins, the noisy morn­ing rush hour traf­fic has clogged the street in front of the restau­rant. To dupli­cate the hear­ing of the busy girl with the cast on her leg and other things on her mind, they use the prosecutor’s ears as he stands across the street. (Later there were sev­eral pro­fane alle­ga­tions about what the pros­e­cu­tor had to endure to fully recre­ate the event.) The results of the test? “It sounded like a hand clap,” said one of the detec­tives sta­tioned in the restaurant’s din­ing room. So, appar­ently, one can induce deaf­ness by mak­ing love to a girl in a full-length leg cast, the same as one can by cut­ting all six legs off an ant.

Examples of truly Bad Science are every­where. So, what can one do to avoid ambush by the Bad Scientists? Three small philo­soph­i­cal exer­cises come to mind. The first is a method­olog­i­cal bat­tle plan called “Ockham’s Razor,” named after the 14^th cen­tury philoso­pher William of Ockham. In phi­los­o­phy, it says that a prob­lem should be stated in its basic and sim­plest terms. In sci­ence, accord­ing to Ockham’s Razor, the the­ory that fits the facts of a prob­lem with the fewest num­ber of assump­tions is the one that should be selected. This is the great-grandfather of the K.I.S.S. (Keep It Simple, Stupid) the­ory, and it works well against Bad Scientists.

The sec­ond tac­tic is termed “reduc­tio ad absur­dum,” which is the dis­proof of a propo­si­tion (or stu­pid exper­i­ment) by show­ing the absur­dity to which it leads when car­ried out to its log­i­cal con­clu­sion. A good exam­ple of such a sit­u­a­tion is the afore­men­tioned case of the pros­e­cu­tor who argued that seven fin­ger­prints iden­ti­fied by two print exam­in­ers make a total of four­teen lit­tle traces of the bur­glar defen­dant. The reduc­tion ad absur­dum of that case is the notion that a third print exam­iner would up the ante to twenty-one clues, or that a dozen exam­in­ers iden­ti­fy­ing a sin­gle fin­ger­print would make for 12 traces of the sus­pect. The clues mul­ti­ply like bunny rab­bits. The mind bog­gles. Think of where the Bad Scientist is try­ing to lead you and look to the dark at the end of the tunnel.

The final fall­back is to com­mon sense, the bane of Bad Scientists the world over. It was Thomas Huxley who said, “Science is sim­ply com­mon sense at its best–that is, rigidly accu­rate in obser­va­tion and mer­ci­less to fal­lacy in logic.” This is where juries trod on the best laid plans of elo­quent attor­neys. They step back for a moment and resort to instinct, to com­mon sense. Lawyers, espe­cially those True Believers who do the pros­e­cut­ing, are noto­ri­ously bad at feign­ing com­mon sense. They are bet­ter at reduc­tion ad absur­dum. Cops, on the other hand, are excel­lent at instinct and com­mon sense, but poor on see­ing the absur­dity of a proposition’s log­i­cal conclusion.

Lastly, one needs to stand one’s ground. And this means more than just Do Not Testify To Methods Beyond Your Expertise or Do Not Selectively Ignore Evidence To The Contrary or Do Not Overstate Your Qualifications. Standing your ground means you have to get in the face of any­one who even hints at being a Bad Scientist. You’ll need to gen­tly redi­rect the novice Bad Scientist at times, show­ing him the light and let­ting him know where you stand. With the more sea­soned advo­cates (pros­e­cu­tion OR defense), you may need a chain­saw to carve out your turf in the Bad Scientist’s office, be it a med­ical examiner’s office, a lawyer’s office, or a supervisor’s office. Draw the line. Let them know that Enough Is Enough. After all, you’re the bas­tard child of both Science and Forensics. They’ll expect you to be incor­ri­gi­ble. J. Robert Oppenheimer said it best when he wrote: “The sci­en­tist is free, and must be free to ask any ques­tion, to doubt any asser­tion, to seek any evi­dence, to cor­rect any errors.”

——————————————————————————–

“Bad Science” has pre­vi­ously appeared in the Midwestern Association of Forensic Scientists Newsletter, Oct. 1991; the California Association of Criminalists Newsletter, July 1992; the Canadian Society of Forensic Sciences FORUM, Dec. 1991; the Mid-Atlantic Association of Forensic Scientists Newsletter, April 1993; the Southern Association of Forensic Scientists Newsletter, March 1992; and the Southwestern Association of Forensic Document Examiners Newsletter, Winter 1991.

——————————————————————————–

Send ques­tions or com­ments about this essay to:

Dean H. Garrison, Jr., Gunhand@aol.com