“DO YOU THINK I’M STU-PID?”

                              If you want my ‘ttention
                             And you think I’m stupid
                             Just reach out and quiz my brain

Perhaps this marketing  professor would be better off sticking to what he knows best rather than trying to improve (and sanitize) aging rock singers’ work.  Nevertheless, it is tempting to speculate that if Rod Stewart had been writing one of his songs in these days of Amazon.com, perhaps the lyrics in one of them might have been a bit different. 

Recently, shoppers at Amazon.com have been invited to “win a nickel” that will be credited to their accounts by supplying answers to very simple questions.  Does Jeff Bezos—who prides himself on hiring only extremely bright people for his management team—think we are stupid?  He probably doesn’t.  Still, the strategy makes sense for a number of reasons.

Let’s look at what is actually being asked.  One question, for example, asks whether it is “true or false” that Amazon carries “at least four times” as many models of HDTVs as most electronics stores do.  A “hint” is then given that a typical store “only” carries about twelve different models.

These questions serve two purposes.  One is to educate or to persuade.  By getting consumers to select the answers that are clearly intended to be right, Amazon is able to “plant” an advertising claim in the consumer’s mind.  In psychological jargon, we refer to this as “attitude change by adding a belief.”

The strategy actually goes a bit beyond just communicating this one claim.  It is typically very difficult to get consumers to give much of their attention to an advertising claim.  Just getting a consumer to stay at one’s web site or getting someone to read an advertisement is a challenge.  This is why direct mailers such as Publishers’ Clearinghouse will allow a consumer to send in a sweepstakes entry without making a purchase—but only if he or she goes through a lengthy procedure that might involve finding a “non-purchase” stamp on a sheet full of magazine article choices.   The marketer has now “captured” a large part of the consumer’s attention for a few precious moments.

Let’s look at what is going on.  In the HDTV example, note that the question hint actually “temps” the consumer to do a quick calculation and make an inference:  If Amazon has four times as many HDTVs as typical retailers, which have about twelve—that must be—hum hum—about forty-eight or fifty.  This simple arithmetic problem was able to hold the consumer’s attention for a fleeting moment—but long enough to make the consumer think that maybe it would be fun to go shopping on Amazon for HDTVs.  If an HDTV is not in the picture at the moment, the consumer may be tempted to go shopping for something else.  Come to think about it, I could use a DVD player, couldn’t I?

As we have seen, focusing on one aspect of Amazon—its HDTV selection—is likely to “trigger” other thoughts about Amazon.  In a process psychologists call “elaboration,” processing the information about Amazon is likely to “activate” a “link” to some other belief about Amazon—e.g., that this seller carries a large number of CDs.

What else might happen to consumers who participate in these contests?  Research has shown that people who are given a reward—even if trivially small—often have the moods boosted.  Happy people are more likely to buy, and also tend to be less critical in evaluating product claims and offers.   And, the next time the consumer buy something from Amazon, he or she will get a reinforcement—at the checkout point, he or she is reminded of the earned discount as this subtraction is shown from the total cost.

Maybe Rod Stewart isn’t “with it” enough anymore to modernize his song.  Odds of having a hit rap song about Amazon’s promotional objectives are probably far greater.

NOW, WHAT DOES THIS RESEARCH REALLY SHOW?  Some cynics occasionally like to suggest that “You can prove anything with research” and that research results therefore cannot be taken seriously.  One should, of course, not be too discouraged by cynics—after all, they’re a dime a dozen—but the concern about misleading research results should be taken seriously.

One of the biggest problems in research is what scientists call a “confound.”  That is, when two variables appear to go together, it is tempting to conclude that one causes.  Yet, in many cases, both are actually caused by something else.  Unfortunately, even when one thing doesn’t cause another, an explanation that suggests a causal relationship can seem to make a lot of sense—at least until we look more closely.

Consider, for example, the clear research results that children who have more toys seem to be more intelligent than those children who have fewer.  This makes sense—playing with a lot of different types of toys might very well stimulate more areas in the children’s brains, with the result that they grow up to be smarter.  Before we conclude that this is the case, however, we need to rule out the possibility that the results were caused by something else.  Having more toys, for example, could be caused by greater parental affluence, which would allow these same parents to provide better nutrition and access to higher quality schools.

To rule out that the results are “driven” by other variables, social scientists will often “adjust” the analysis to “control” for other potential causes. For years, one argument against the hypothesis that smoking increases a person’s risk for cancer is that smokers may differ from non-smokers in other important ways—e.g., by having less healthy eating habits or being more sedentary.  In the case of children’s toys, if the researcher has access to both family income and the number of toys owned by the child, the researcher would be more confident if the number of toys still had an independent effect even when income was entered into the equation.  There are, of course, other potential causes that might need to be ruled out.  Perhaps it is not so much family income per se that makes the difference, but the priority that the family puts on the children.  Parents who buy more toys for their children might also spend more time with them.  Thus, no one research effort is likely to settle the question.

Confounds can also be introduced by the realities of the way the world works.  For example, direct marketers, who want to maintain their low “bulk mail” postal rates, are highly motivated to show public policy makers that consumers enjoy receiving mail.  To make this point, they have done research that shows that one of the first—if not the first—things consumers do when arriving home is to check the mail.  Yes, of course it is!  Chances are that the mail box is located on the way into one’s home, so the easiest time to pick it up is immediately upon one’s return even if nothing special is expected.  Nowadays, chances are that that important love letter—and even more important, the Consumer Psychologist Newsletter--have been sent by e-mail anyway.  Now, what would be a “fairer” test—especially for the attractiveness of direct (“junk”) mail?  One more realistic measure might be when—or even whether—people actually look at or read the commercial mail pieces that arrive.

The reality that “correlation does not mean cause” can often be overlooked.  Over the years, there have, for example, been a number of reports that people on certain antidepressive medications are many times more likely to commit suicide than those not taking the medication.  Statistics clearly show this to be the case.  However, it is important to keep in mind who takes these medications in the first place—usually seriously depressed individuals.  Other statistics show that among depressed people, those who do not take medications are more likely to commit suicide are than those on medication.  The latter is a much more relevant comparison.

One of my favorite empirical questions—but perhaps one that only a marketing professor would bother to ask—is whether tall women are more or less likely than average height or short women to wear high heels.  I don’t know the answer one way or the other.  One might speculate that tall women either would not feel the need for the extra height or would prefer not to “stand out” as much by being even taller.  On the other hand, one might also hypothesize that many tall women are proud to be tall and would like to emphasize their stature.  Would comparing the shoes of a random sample of tall and short women help answer this question?

There are several problems with this kind of test.  One problem is factors other than height are at work.  Some women, for example, might choose their shoes based on expectations in a managerial or professional work setting.  This, however, would not necessarily lead us to incorrect conclusions (unless height is related to one’s career choice).  We also know that within each “height group,” segments with different tastes are likely to exist. Throwing more variation into the picture does, however, make it more difficult to detect an actual relationship that might exist.  (You may remember from statistics the notion of “residual variance.”)

A much more serious problem, instead, is the reality that height is highly correlated with ethnicity.  Since there are so many other factors at work, it is unlikely that height alone would account for most of the variation in women’s shoe choices.  Differences in style among ethnic groups—which may vary over time—could be much more significant than height.  Thus, in this study, we should at the very least compare tall and short women within different ethnic groups rather than comparing others.