Effectiveness in Advertising - An Answer

This is a follow-up on my earlier post Effectiveness in Advertising - A Question. The basic premise of that post was that massive amounts of money is spent on advertising, but that there is little public or common knowledge about how advertising works - or rather there is public knowledge, but it is highly socially constructed knowledge, and is not exactly empirical.

I mentioned a meta-analysis of 389 real world experiments in advertising (Lodish et al, 1995), where they measured the effect of advertising on consumer purchases. Although the results are frustratingly complicated, this is their primary conclusion:

There is no simple correspondence between increased T.V. advertising weight and increased sales, regardless of whether the increased spending is compared to competition or not. ... changing the spending level had little or no impact on sales for established brands.

In other words, there is little relation between the time and money spent on t.v. advertising and how much the advertised product actually sells. That is not to say there is no relation - it is complicated and subtle - new products benefit from advertising, and brands that are reinventing themselves benefit from advertising. But in general, for established brands, the effects of advertising are not detectable in terms of sales.

Persuasion is not easy.

Yet we are all too ready to accept the myth that commericals actualy sell products and that educational campaigns really do alter behaviour. Most of the time, it does not even occur to us that these phenomena are questionable in their function. It is a social construction that even the persuaders are unaware of.

Some more examples:

Drug Abuse Resistance Education (DARE) has been tought to 36 million children (including me) in 54 countries. The program is now 24 years old. There is just one problem with it: it doesn't work. Ok, I lied; there are tons of problems with it, but I will focus on the most salient one. I will quote from three research studies:

suburban students who participated in DARE reported significantly higher rates of drug use … than suburban students who did not participate in the program.

- from a longitudinal study of 1800 students over six years (Rosenbaum, 1998)

DARE didn’t keep children from using drugs. In fact, it found that suburban kids who took DARE were more likely than others to drink, smoke and take drugs.

- from a 1999 study by the California Legislative Analyst's office

Our results are consistent in documenting the absence of beneficial effects associated with the DARE program. This was true whether the outcome consisted of actual drug use or merely attitudes toward drug use. ... The only difference was that those who received D.A.R.E. reported slightly lower levels of self-esteem at age 20.

- from a 10-year longitudinal study, N = 1000 (Lynam et al, 1999)

There is more from West and O'Neal (2004) in a meta-analysis, but it is all the same stuff: DARE doesn't work and sometimes it makes things worse.

DARE has many problems though - such as being rather dishonest in their methods. What about other persuasive campaigns?

Let's take healthy-eating campaigns in schools. The US federal government spends about $1 billion on these every year. But a recent meta-anaylsis of 57 programs found that only 4 showed even a modest increase and the rest had no effect in behavioural change. Uh oh.

What about abstinence-only sex education? This is something the US government has spent over $1 billion on in schools (again, I sat through these in school.) Not only that, Bush (and others) have linked African aid programs to abstinence-only education programs. Teach people how condoms work, and bye-bye funding. But, as you can now guess, these programs do not work. Not only that, teens who take virginity pledges have sex at the same rate as non-pledgers but are less likely to use condoms and more likely to contract STDs. Now that's progress! Let's take a program shown to increase STD rates in the US, and implement it in Africa in the name of reducing HIV/AIDS! Hooray for morality!

On to a topic that is closer to my research efforts - education about the environment has little to no effect in influencing rates of environmental behaviour (recycling, saving energy). Because there are not national-level programs on this, most of these studies are not as high-profile or as publicly documented, but I assure you, they are there. See Clayton and Opotow (2003) for examples of many. Yet environmental NGOs, like the Sierra Club, and local governments still take the education --> behavioural change approach. It does not work.


The take home point: Standard persuasion attempts rarely work, whether in selling products, behavioural changes, or world views. Effects that do occur tend to be blunt and rather uncontrolled. Sometimes they even backfire. This is well documented in hundreds of studies. That's kind of interesting, but the more interesting point is that the persuaders seem to ignore this so doggedly that they almost seem to be consciously suppressing the results, even to the detriment of both their pursestrings and their cause. You would think the advertisers and campaigners would want to know this. The socially negotiated perception of persuasion is so strongly cemented in place that most who march for a cause do so in endless circles in the desert, accomplishing little, victims of a creeping attrition.

Mass influence is possible - but we need to escape our artifical realities of how influence works, to understand how and why.

How exactly that is done is a discussion for another day.