One scientist cited a 2008 pre-election study in which American voters unconsciously rated Tony Blair as "more American" than Barack Obama. (How they revealed the alleged unconscious choice wasn't clear from the interview.) The researcher blamed race. Maybe, but there are other possibilities. Here's one: Tony Blair was George Bush's ally; perhaps the results merely revealed loyalty to Bush. The other scientist cited a study in which diners tipped more generously when servers repeated orders back to them verbatim. He concluded that mimicking was causal. Maybe, but again, there are other possibilities. Here's one: perhaps repeating the order increases diners' confidence that the server won't screw up the order, and that leads them to tip better.
Responsible scientists acknowledge their own fallibility and guard against leaping to conclusions. That's why double and triple blind tests, replication, peer review, and tests that eliminate other possibilities are hallmarks of the scientific process.
Marketers who discipline themselves in the same way are more likely to learn what works. But even marketers with the best intentions are subject to the guile of self-serving conclusions. When sales go up during an ad campaign, the account manager, media planner, creative director, writer and art director will be prone to give the credit to, respectively, strategy, targeting, concept, copy and design. Who knows. Maybe an interior designer will say that the wallpaper he or she picked out for the conference room provided the inspiration for the winning campaign.
So here's an exercise for you. Gas mask sales rose in the months following September, 2001. Outside of strategy, targeting, concept, copy, design and wallpaper ... can you think of any other possible causes?