You must. You've gone to all of the trouble to care enough about a topic to risk public humiliation by getting up and talking about it. You've thought long and hard about how your experience might be employed to change the world of the audience member in some tangible way. You have even gone to the trouble to consider how best to organize your thoughts into a message that will be both readily comprehensible to and recallable by the rest of us.
And yet we ask . . . Who Cares?
Okay, mom cares! But even mom doesn't care as much as
you think she does. No one does care--and no one should
care as long as your claims are just suppositions floating in
space unconnected to any source of gravity save the weak
attraction of wishful thinking. To make us care, you must
give your claims gravity, and you do so by attaching them to the
substance of the world, of being and experience, of consistency
and likelihood. You give your claims gravity by attaching them to
Evidence!
In Virtual Insistence, I made the claim
that a good claim is
demonstrable--that the speaker should be able to point
to a place in the world where the truth of his or her claim is
bourne out by experience. The provision of of evidence is the act
of doing just that.
Following I will elaborate on the uses strengths and weaknesses
of the three basic kinds of evidence:
A recent edition of the New York Daily News offered advice on how to avoid being taken advantage of by spiritualists. Among the nuggets of wisdom offered was [paraphrase] "Be suspicious if they claim someone has put a curse on you, and it will take several visits to lift it." Well, since to my limited mind the best way to avoid being taken advantage of by spiritualists is not to believe in them, I wondered if there was anyone in my immediate circle of acquaintance who had ever, in fact, been cursed. My investigation led to one suspicious anecdote about a teacher putting a voodoo hex on her student (Nothing in the handbook specifically prohibits this!) and several other contributions which began approximately "Cursed? No, but . . . " and ended with a foray into the ethereal plain.
In each case, the individual was convinced that something super-normal had occurred, simply because it had happened to them or to a close acquaintance, in most cases no more than once.
Such is the nature of EXAMPLE. An example is simply A case in which the truth of the claim has been demonstrated by experience. Want proof that someone can be cursed? Find someone who has been cursed. Want proof that a given action can lead to dire consequences? Find someone who has engaged in that action and suffered same.
Examples are called by many names: anecdote, illustration, case study, etc. But they all work on the same principle: Support the existence of a phenomenon or relationship by pointing to it in the world.
Examples are a profoundly popular form--probably the most often employed form--of evidence, because they are Visual! They provide the listener with an image on which to hang the abstract case. By engendering specific images (usually from the individual listener's own experience, since "my friend Bill" will at least initially signal some visual recall of your friend Bill), they engender imagination and emotion. Examples, because they are framed in the human tradition of narrative, have the capacity to call forth the total range of human emotions.
Since the primary value of an example is its capacity to visualize and intensify the claim, it should be used as soon in the development of the claim as possible. If your example illustrates the nature of a specific idea, it should be employed as soon as possible in the development of that idea. If an example illustrates the issue underlying the entire speech, you should consider using it as an introductory device.
The example is able to visualize the case only to the extent that the speaker take the time to build into it properly concrete images. If you say:
"I had a friend who got conned by a phony psychic, so I know it can happen." Your friend lost his money in vain.
If on the other hand, you say:
"My friend Bufford, in New Orleans, went to a psychic because he was unable to deal with the untimely death of his wife in a car accident. The psychic convinced him that his wife wanted to communicate with him from beyond the grave--that in fact she had vital information for the well-being of their children--but spiritual barriers had been put in place by Bufford's sins against others who had departed the mortal realm. By the time Bufford found out he was being defrauded, he had lost over $17,000."
you are more likely to achieve the desired effect. The second telling of the example has names and laces and motives, and consequences--the stuff of good story-telling. The audience member will become engaged in the latter version, whereas he or she will merely assent or dissent to your unsubstantiated claim of experience in the first.
Be careful to avoid what is called Argument From the Exception. Even the best related case is only one case! It serves no better function than any other. For an example to "prove" your claim, it must be Generalizable--that is it must represent the Rule as opposed to the Exception. Argument from the exception (obviously) is an argument that does the latter. Witness
as proponents of somewhat exceptional means of overcoming difficulties. Even assuming that all of the above met with some level of success, who would want to accept their approaches as policy? Argument from the exception ignores where it does not exploit the weakness of anecdotal evidence by inferring that the example does reflect the rule unless evidence surfaces to the contrary. Faced with the choice of employing argument by the exception or eschewing examples altogether, the speaker is both ethically and strategically advised to steer toward the former.
Beware of Apocryphal Stories.
Also known as Urban Folklore, these stories combine credible proximity ("It happened to a friend of mine.") with a can't-miss surprise ending (" . . .and that man was baseball legend Reggie Jackson.").
Any good cultural anthropologist will tell us that apocryphal stories also share one other important feature--a barely tangential relationship to reality. They may be loosely based on actual occurrances but are more likely fictional depictions of common social fears (e.g. A.I.D.S., foreigners, other races, etc.). The problem for most of us is that by the time such stories have filtered to us, they have filtered to others in the audience as well. Be aware that no matter how good a piece of evidence might be, nothing is worth the risk of your credibility. If anyone in the audience hears you claim "ownership" of an apocryphal event, your credibility will suffer a serious of not fatal blow. Don't risk it.
If you are going to use a story, verify the story!
"There are three kinds of lies--lies,
damned lies and statistics."
--Mark Twain
With tongue firmly in cheek, Twain was relating the commonly held notion that statistical evidence is suspect because a statistic can be found to support almost any position one desires to espouse. Advocates of this view, though, need to consider two essential facts about statistics. First, statistics describe, they do not explain. Statistics are numerical descriptions of phenomena. They may tell us what percentage of midwestern, soccer-playing, twenty-five to forty year-old women shop lift, but they are bereft of motive. They may tell us who you would vote for were the election held tomorrow, but they neither have nor claim the authority to tell us why (nor for that matter how you feel about their having rescheduled the election). True, they can tell us why you say you will vote, but even that is merely a description, not an accounting. Second, where statistics get their primary power to disrupt is--like sugar and mower blades--in the way we spin them. Most of us derive explanatory power (at least the illusion of explanatory power) from statistical descriptions by combining them into models and trends:
A model is a combination of statistics which purport to describe the same thing. Our aforementioned midwestern soccer player is a model. So are "market basket" descriptors, the simulations of the universe constructed in super computers, and Lara Croft's caves.
Trends employ repetitions of the same numerical description to chart a forecast or extension of the statistic beyond the present.
Each can be of significant value as long as its employer remembers to distinguish between the statistic itself and its interpretation.
That said, statistics lend an air of generalized authority to claims that no other type of evidence can muster. They provide an atmosphere of scientific expertise magnified by the sense of categorical stability. To be good evidence, however, they must conform to four criteria:
Statistics have the capacity to describe complicated phenomena with astounding accuracy, but their capacity to do so is directly related to the expertise with which they are derived. The point is
There are no "pretty good" statistics!
Statistics are either professionally derived or pure Ouija Board. There is no middle ground. Those television call-in polls where you dial 1-900-rip offs to vote yes and 1-900-con jobs to vote no . . . pointless exercises in primal rage. If the poll taker can't control the population of respondents, it is unlikely that the poll taker can derive an accurate result. Can the researcher tell how harmful are the effects of nicotine intake on humans by pumping it into white rats? Probably. Can the researcher gauge the effect of the rats' yellowing nails and foul breath on their courtship rituals, almost certainly not (Just to be on the safe side, Ratso, skip the Luckies and just saw Gnaw!). We are like white rats in some ways, unlike them in many more. The point is that the statistician has to know and be able to prove the ways in which we are alike before he or she can even begin to design the test.
But let's face it. If you are the potential speaker hunting down statistics to use in your speech, or worse yet the innocent audience member being asked to ingest them at conversational pace, how are you to know which numbers are professionally derived and which are not? If the phrase statistical correlation and regression leaves you hungering for a life . . . if ANOVA is that '79 Chevy you've been saving for . . . if scatterplot sounds like that new trivia game . . . welcome to the world--and it is densely populated--of people who know nothing about how to calculate statistics (If on the other hand you do recognize all of those terms, look to your left; Mr. Wonderful wants to borrow your notes.). We the uninitiated need a way to tell which numbers we should embrace and which we should spurn.
Luckily the test is simple:
Morris's First Statistically Unsupportable Law of Evidentiary Cynicism.
If the counter had anything to gain from the count, count the count out!
Numbers are too easy to cook, tests too simple to rig. Guess what? McDonald's has conclusive proof that Americans prefer it's fries to those of Burger King. Guess what else . . . come on . . you know!
Thanks to our friends at Ken White's Coin Flipping Page, I just flipped one hundred pennies and got 55 heads - 45 tails. The outcome was completely a matter of chance. Hold on a second, I'm going to go do it again .. .. ..57 heads - 43 tails, 43 heads - 57 tails .. .. .. Now substitute the word Bubblicity for the word heads and Bucket o' Caffeine for the phrase tails. Employing no more than three random trials, I have derived actual advertser worthy proof that in scientific tests, people overwhelmingly chose Bubbilicity over Bucket o' Caffeine . . . or was that Bucket o' Caffeine over Bubblicity? Don't like the mere 14 point spread? Head back to Ken's and start flipping. Sooner or later random chance will provide you with even more startling proof. Paging Mr. Twain . . . phone call for Mr. Twain!
Given this reality, your best bet is simply to discount any statistical evidence not distinctly presented as objective. If you are the speaker, it is to your strategic advantage to build a case for your number's objectivity into its presentation.
Instead of saying "According to published polls . . . ." say "According to the independent polling firm, Dinnerwreckers Inc. . . ." Don't say "Statistics show . . ." say "Figures from the Bureau of Labor Statistics show . . .." It only take a few seconds longer and adds manifestly to the credibility of the proof.
A statistic is a numerical description of a phenomenon at the moment that the phenomenon is described. The farther away from the count you get, the greater the likelihood that something about the phenomenon has shifted. The more complicated the phenomenon being counted, the greater the rate of decay. So use recent numbers. President Roland Reagan, for instance, had an affinity for statistics and a penchant for nostalgia. As a result he often employed nostalgic numbers without reference to their historical freshness dates. He had a knack for getting away with it. You probably don't! The good news is that there are tons of statistical resources at your fingertips. The bad news is that everybody these days knows that and expects you to have fresh numbers.
This one gets a little tricky. Does the statistic offered as evidence actually support the case? In many circumstances--modern advertising being among the most egregious--statistics are offered which appear to support the case, but in fact do not.
Consider the following paraphrases of popular advertising appeals (Hey, who wants to get sued?):
In laboratory tests, no toothpaste was proven better that Gumsplash.
3 out of 4 dentists who recommend sucking on charcoal embers recommend letting them cool first.
7 out of 10 users of Chub-No-More report significant weight loss.
In the first instance, one can assume that Gumsplash is superior to its competition, but one is probably safer in assuming that no discernible difference existed on the tested criterion. If they could say they were better, they would.
In the second instance, one needs to know what percentage of dentists fall into the category who think sucking on charcoal embers is unwise.
In the third instance, the makers of Chub-No-More likely do not know how many people use their product. They probably only know how many people buy it. In addition, we need to know whether they compare number of people who report success to those who report failure or whether they compare people who report success to sales.
A prominent Right Wing radio talk show host once used Household Income statistics to show that the middle class had improved its financial situation in the thirty years between 1965 and 1995. He forgot to mention that, in 1965, a household income was earned by one person, whereas in the majority of middle class households, in 1995, it took two. On needs to wonder whether almost doubling productivity to create a 3% gain constitutes progress.
Another place where dissonance exists between what statistics say and what they only appear to say is in visual presentation.
Would you prefer putting your money in the Buchsucher Account or in the Blindseid Fund? Me? I'm going with Buchsucher. They have big pointy lines that seem to go up more often than Blindseid's which look pretty flat by comparison. And where money is concerned, big pointy lines that go up are good!
If you look at the vertical indices, though, you will notice that Buchsucher's is demarcated (divided) five times, whereas Blindseid's is demarcated into only four. SO it is possible that were these real numbers (which they are not) they could represent exactly the same level of performance.
This type of presentation scheme is common. So are such practices as using icons (e.g. missiles, syringes, soldiers, etc.) to represent the lines in bar charts, using different colors to exaggerate the contrast in pie charts, etc.
To be sure your statistical evidence does what it purports to do, you have to go beyond the presentation and see the data presented.
Additional examples of statistical mistakes are discussed in the chapter Virtual Disaster.
Of course, no statistic is any better than the audience member's capacity to receive it. The employer of statistics should first have a firm grasp of what they actually say, then think about different ways to present them for maximum effect.
| Eight times as likely | and | 800 percent more likely |
say the same thing. Which would have better impact for your purpose? To say 75% is to say 3 out of 4. Again, which most effectively conveys the point?
Also, be wary of assaulting the audience with numbers. Remember that eight out of ten people report that over two thirds of the information they receive within the first 40 percent of a message loses over half of its retainability as compared to the 85 percent retainability rate of anecdotal support which occurs within the last quartile of the same event.
Got that? There will be a quiz!
"Don't quote what I say; quote what I mean." |
|
| --from the 365 Dumbest Things Ever Said Calendar |
Before we get started, run out and rent the classic Kurosawa film Rashomon. I'll wait here.
Back? Great. In a moment, we can dispense with the worst possible kind of evidence--Eyewitness Testimony. But first let us define our terms.
Testimony is the co-optation of someone else's credibility. In essence, you say something is true. We don't care, because you are just you, and what do you know? You point out, though, that Ms. So-and-So, big shot expert in the field you are discussing is on record as saying the same thing you just said. We are more likely to agree with you, because someone who's credibility we trust says it too.
It's as simple as that. Testimony is an attempt to bolster your own credibility by aligning yourself people people who already have some.
The criteria for employing testimony do not differ greatly from those for using statistics.
It is necessary, when holding someone accountable for what he or she as said to make she that he or she has in fact said what you claim. Whenever possible, check the primary source of a quote. If an expert is cited in a newspaper article (a secondary source), try to go to the original circumstance and find the quote. Was it a book or magazine article written by the source? Was it a press conference? You've played Telephone. You know that every next level of distance between you and the point of origin further obscures the original transmission of the message.
What if I can't?
If you get a quote from a credible secondary source, include the source in the delivery of the quote.
"Newsweek Magazine quotes Congressman Tsunami as saying . . . ."
If you get the quote from a less-than-credible source and can't find its point of origin, get another quote!
e.g. Don't pass along Kurt Vonnegut's commencement advice to "Wear Sunscreen." He never said it.
There is often a substantial difference between accuracy and truth. It possible--even easy--to follow all of the prescriptions above and still utterly misrepresent what someone says. The quote with which I began this section illustrates the point (if somewhat ironically) that what a person means is at least as important as the verbatim effluvium that might escape. Why not head over to the Student Speech Quote Hall of Fame and peruse some of the utterances that people just like you hope they will not be held accountable for.
e.g. Don't cite Vice President Dan Quayle as actually meaning he had trouble communicating in Latin America, because he didn't speak Latin. He said it, but he did so in private, off the record as a self-parody (actually not a bad one for Dan). Besides there is plenty of other ammunition in the breach.
Opinions don't degenerate as quickly as statistics. Still, people do change their minds,and the farther from the origin of the quote one gets, the less likely the source is to still espouse the view.
Hillary Clinton on the tests of grounds for impeachment, while she was working to remove Richard Nixon from office?
Robert Kennedy's view of McCarthyism when he worked on McCarthy's staff?
Ronald Reagan when he said, in 1980, no President who can't balance the budget in four years deserves a second term?
There are less obvious but equally instructive cases. If you assumed George Wallace still supported "segregation now, segregation forever" in his later years, you would be wrong. If you assumed that Ivan Boesky still thought "Greed is good." after he got out of prison, you would be wrong. If you assume the people at Universal Studios still consider Clint Eastwood's adams apple to big a hurdle to stardom, you would probably be wrong (You know how movie people are.).
People change their minds. Ethically and strategically, it is in your best interests to respect that.
For the credibility of the expert to have any effect on the speaker's two conditions ought to exist:
Answer: Wrong! Wrong! Wrong! Wrong! Education, specifically secondary education, which is of course a worthwhile degree, the holders of which, including Dr. Ruth, are worthy of admiration. The point is:
- did it prepare her to be a credible sex therapist, and
- did it ever occur to you to find out?
It is not my point to disparage any of the people or institutions mentioned above (Yeah, right Pinocchio!), simply to emphasize that titles do not necessarily translate into knowledge and/or expertise. If you are a speaker, argue the credibility of your source, do not just assume it. If you are a listener to whom the speaker does not offer such information, disregard the evidence. If no further evidence is forthcoming, disregard the claim.
The current wisdom on evidence comes down to these simple points:
[Return to Virtual Clarity] [Go Ahead to Virtual Consensus] [Return to Table of Contents]