3

I was searching the correlation between drinking soft drinks and IQ, because I wanted to verify if countries that drank a lot of coke/soft drinks were more likely to have more intelligent populations (e.g. the basic rationale was that sugar in carbonated drinks would speedup brain faculties, thus having more synapses, and overall better brain performance)

I found in that 2012 NCBI article citing the following:

A total of 529 children participated in this study, (77.7%) of the children have high IQ level. Almost (60%) of the children drink carbonated drink daily, while (74%) eat chips every day. The association between drinking carbonated drinks, eating chips and intelligence level was significant (P= 0.043, 0.001) and prevalence odds ratio of 1.5 and 2.4 respectively.

And the study concludes with:

There were significant associations between carbonated drinks, chips intake and IQ score. More regulations on what to sell inside the schools canteen are needed. Increasing nutritional knowledge of the parents, especially the mothers is very important and more health promotions should be given regarding children nutrition in the early years of school.

Am I reading this right, that junk food and drinking soft drinks is correlated with higher intelligence quotient ? Does all that sugar beside its negative side effects (tooth decay, insulin highs) have this positive effect on intelligence ?

  • While I can't comment on the validity of the correlation, I would be suspect of any study that claims nearly 78% of the children have IQ's in the 75 percentile or higher. That is a very abnormal distribution. – JohnP Feb 6 '18 at 14:42
6

Am I reading this right, that junk food and drinking soft drinks is correlated with higher intelligence quotient ? Does all that sugar beside its negative side effects (tooth decay, insulin highs) have this positive effect on intelligence ?

No you are not reading it right, but your mistake is very a common misunderstanding of statistics.

The key point to remember is: correlation does not imply causation.

That means that just because two things appear to coincide, you can't conclude that one causes the other - or even that they are related at all.

This is a cross-sectional study. They are looking at these variables at a single point in time. This kind of study allows you to say "interesting, many people who own red cars have cats." Correlation. But what it doesn't do is allow any inference of which causes the other - you can't say that people who have cats are likely to get red cars, or the other way around.

To attempt to establish causation, a study design that manipulates or strictly observes factors across time is required. Of these designs, the Randomized Control Trial design is generally the strongest for that purpose - but that is a discussion in and of itself.

This is a humorous website that I share to emphasize the key message to take from this - just because the prevalence or incidence of two variables coincides, you can't say anything about whether there is causation or relationship at all.

To answer your question, in review of the study, there is absolutely no evidence to support that one causes the other, or that they are in any way related. Any theorizing about why they correlate would have to be tested in another study design in order to hold weight.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.