People find smokers less physically attractive, study reports

Wednesday December 13 2017

"It's been scientifically proven that people are more likely to find smokers less attractive," The Independent reports. But the "proof" isn't as conclusive as reported.

Researchers asked people to rate sets of photographs of identical twins where one twin smoked and the other didn't.

They found people could sometimes tell whether people smoked, and tended to prefer the faces of non-smokers.

But there are problems with the study that make it hard to rely on its results. It was a small study involving only 23 sets of mostly white twins, and all of them, apart from 3 sets, were men.

Researchers wanted to see whether people could identify smokers from non-smokers by face alone, and whether they preferred non-smoking faces.

The study had inconclusive results. Analysed one way, results suggested people could tell smokers from non-smokers by face alone.

But analysed another way, the results suggested people's chances of identifying smokers was no better than flipping a coin and they had no preference for non-smokers over smokers.

Participants were also shown computer-generated images using the photos of all the twins to see if they could spot the typical smoker and the non-smoker. Here too they also found the non-smoker more attractive.

Aside from the obvious advantages of reducing your risk of cancer, heart disease and stroke, quitting smoking can help delay skin ageing, keep your teeth from becoming stained, and improve the smell of your breath.

Get more advice on stopping smoking.

Where did the story come from?

The study was carried out by researchers from the University of Bristol and the University of Oxford.

It was published in the peer-reviewed journal Royal Society Open Science on an open access basis, so it's free to read online.

The research was funded by the Medical Research Council and the University of Bristol.

The Independent refers to "proof", but based on the inconclusive results of the study, such proof would probably not stand up in court.

Similarly, Mail Online's headline is wrong – it says the study involved a "survey of 500 twins", but only 23 sets of twins took part. The photographs of the twins were viewed by 500 people.

Their article said people could "easily identify" the smoker out of a pair of smoking and non-smoking twins – which is correct on one analysis, but not the other.

The article didn't explain that the results were described by the researchers themselves as "inconclusive".

What kind of research was this?

This twin study recruited people online who were then asked questions about sets of photographs.

Researchers wanted to see whether people could identify smokers from non-smokers by face alone, and whether they preferred non-smoking faces.

Studies using twins are often used when researchers want to separate out genetic differences from environmental factors, such as smoking. This is because identical twins have the same genes.

This type of study can provide information about average preferences and perceptions, but doesn't really tell us how people will perceive any one individual.

What did the research involve?

Researchers used photographs taken of 23 sets of identical twins, 20 female and 3 male.

They recruited 590 volunteers to view the images and say whether they thought they were smokers or non-smokers.

They then recruited 580 volunteers to view the same images and say which image they thought was more attractive.

In addition to the 23 sets of twins, people viewed 4 "prototype" images, created as averages of the male and female smoking and non-smoking faces.

The researchers looked at whether people could correctly identify smokers and which faces they preferred.

The photographs were taken in the US for a study of identical twins between 2007 and 2010. Researchers picked sets of twins where one smoked and the other didn't.

The prototype images were made by using averages of shape, texture and the colour of the faces in the smoking and non-smoking male and female groups.

This was intended to remove "idiosyncratic variations" in facial appearance and variations in lighting, pose and expression.

People were recruited to take part online via a crowdsourcing online research platform and were paid a small amount (50p) for taking part.

The analysis of photographs was done in 2 ways:

  • by participant, in which each participant's responses were averaged for their performance in response to all the photographs
  • by photograph, which reports the average response to each photograph from all the participants

Researchers also used questionnaires from the twins to see if alcohol use, sun exposure, moisturiser use or weight affected their results.

What were the basic results?

Using the first type of analysis:

  • photographs of smokers were slightly more likely to be judged to be smokers than non-smokers by male participants (mean response to smokers 0.53, where 1 is smoker and 0 is non-smoker, 95% confidence interval [CI] 0.52 to 0.54) and female participants (0.55, 95% CI 0.54 to 0.56)
  • photographs of non-smokers were slightly more likely to be thought more attractive than smokers by male participants (mean response 0.44 where 0 is a preference for non-smokers and 1 is a preference for smokers, 95% CI 0.43 to 0.45) and female participants (0.44, 95% CI 0.43 to 0.45)

But when analysed by photograph rather than by participant, there was no evidence that people could identify smokers more than half the time, or that people thought non-smokers were more attractive than smokers more than half the time.

There was a lot of variation in how each participant responded to each photograph and how likely they were to spot the smoker.

When using the prototype images, however, 70% of men and 68% of women correctly identified the smoking male prototype and 70% of men and 73% of women correctly identified the smoking female prototype.

Men (66%) were more likely to prefer the non-smoking female prototype and women (68%) were more likely to prefer the non-smoking male prototype.

The researchers said taking account of alcohol use, sun exposure, moisturiser use or weight didn't affect their results.

How did the researchers interpret the results?

The researchers said findings for the twin photographs were "inconclusive", but results from their prototype images "provide evidence that smoking may negatively impact facial appearance".

They suggest the results "have the potential to be of utility in developing and improving smoking behaviour change interventions".

Conclusion

Everyone knows that smoking is bad for you – and its effects on your physical appearance are the least of your worries. Smoking remains a leading cause of illness and death.

Quitting smoking can reduce your chances of dying from heart disease, stroke, cancer or lung disease, helps you breathe more easily, gives you more energy, improves your mental health, and may even improve your sex life.

In addition, we know non-smokers have better dental health, whiter teeth and fresher breath, and smoking can damage the elasticity of the skin. There are many, many reasons to stop smoking.

This study provides some evidence that people may be able to tell whether someone is a smoker by looking at their face, and may prefer the faces of non-smokers.

But the research itself isn't very reliable, so we should probably avoid reading too much in to the results.

Limitations include:

  • the small number of photographs included
  • the weakness and inconclusive nature of the results
  • the self-selecting (and possibly cross-contaminating) nature of the participants
  • the late inclusion of "prototype photographs" in the study
  • the monocultural nature of the study, which seems to be based entirely on white twins (ethnic mix isn't stated)

But the study is just one small drop in a sea of evidence that says that stopping smoking is one of the best things you can do for your health.

Analysis by Bazian
Edited by NHS Website