Thoughts on Psychology

Just another site

Learning About Research From a Cereal Packet

leave a comment »

Actually, from a cereal advert, but that makes a less eyecatching heading. Anyone seen that Special K advert, that makes the following claim?
“Research shows that people who eat a healthy breakfast are more likely to be slimmer than people who eat no breakfast at all”
There’s an interesting lesson to learn from this claim. It seems counter-intuitive, surprising even. The implication, as I read it, is that eating a “healthy breakfast”, presumably Special K, leads to you being slimmer than if you ate no breakfast. However, there’s nothing in the claim that actually supports that implication.

Imagining how this research must have gone, then I’d guess that they took a group of people who claimed to eat a healthy breakfast; and a group of people who claimed to eat no breakfast; and compared the weights or BMIs of the two groups; finding that the healthy group scored lower on average. Fine, but what we have here is a classic quasi-experiment: the groups already exist. As such, you can’t ascribe causality, and there are three possible interpretations of the finding:
* eating a healthy breakfast makes you slimmer (what they’d like you to believe)
* being slim makes you eat a healthy breakfast rather than no breakfast (unlikely)
* people who care about healthy eating are both more likely to be slim and more likely to eat a healthy breakfast (my preferred option)
Quasi-experiments are essentially correlational designs, even if you use a difference test to analyse the results. All the Special K results mean is that there’s a correlation between healthy eating and being slim. Big surprise!

Of course, Special K could have done experimental research, asking one group of people to eat a healthy breakfast for a month; and one group of people to eat no breakfast for a month; controlling for equivalent eating behaviour at other times of day; and then comparing change in weight or BMI over time. If people in the healthy group show a larger drop, then you can claim a causal relationship. The reason I don’t think they did this is because the advert is actually very careful not to claim a causal relationship – they don’t say “eating a healthy breakfast helps weight loss compared to eating no breakfast”. The Advertising Standards Authority wouldn’t let them make such a claim unless they’d done experimental research; the fact that they don’t make the claim suggests they’ve done different research, of the type I described at the start. However, they make their claim in a very clever way, to give the impression that if you eat Special K for breakfast, you’ll lose weight – the first of the three possible interpretations.

Now, I hesitate to criticise anything that encourages people to eat healthily, and particularly anything that discourages people from avoiding eating. However, there’s a cautionary tale here. People routinely over-state the findings of correlational research, and outside the control of the ASA routinely ascribe causality when there’s no basis. Be careful when reading about research results like this. In terms of psychology, the research described is exactly equivalent, in logical terms, to race or gender difference research. We could rewrite the claim like so:
“Research shows that people who come from a white ethnic group are more likely to score highly on IQ tests than people who come from afro-caribbean ethnic group.”
This has shown to be the case in various research projects, but as those of you who’ve suffered my rants in lectures will know, the reasons for this are highly debatable, and to my mind most likely to be because of differing social opportunities and experiences for the two groups.


Written by daijones

September 11, 2010 at 6:55 pm

Posted in Full post, Research methods

Tagged with

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: