Happiness and income comparisons: Biased methods consecrate a worldview

Posted on May 1, 2011 by


File:2010 - A year plenty of Hopes.jpgA paper came out (May 2010) in the Economic Journal on the impact of income comparisons on subjective wellbeing. The researchers, Andrew E. Clark and Claudia Senik from the Paris School of Economics and the Institute for the Study of Labour at Bonn (IZA), analysed data on income comparisons from the European Social Survey, covering 18 European countries. Though I was unable to access this paper (God bless the marketisation of knowledge), I found a discussion paper by the project team publicly available.[1]

The authors claim that, unlike other previous studies (which I haven’t read so couldn’t comment), their study acknowledges variation in reference groups. This means they paid attention to who people compare their income to, how important it is for them, and other socio-demographic aspects.

The study illustrates a trend which has been troubling me for years every time I read about a ‘groundbreaking’ new social study. Firstly, its ‘groundbreaking’ findings were entirely obvious through common sense. What is the function of such research, proving what we already know rather than pushing new boundaries? It seems that it aims to convert common sense to ‘knowledge’ – scientific evidence that you can quote or cite. What such studies often fail to do is to properly examine the biases within the research, which frequently reinforce existing cultural ideas and stereotypes. This does exactly that.

Newsflash: scientific study proves the obvious

It never ceases to amaze me that expensive, large-scale statistical studies often present as ‘groundbreaking’ revelations things that are really rather obvious.

The data revealed the shocking and bizarre phenomenon that people tend to compare themselves more with groups that they interact with more, and are more relevant to their lives. Who would have expected that.

A distinction was made between comparisons with colleagues, family, friends and ‘others’. Unsurprisingly, they found the following:…

Comparisons with friends and family members decrease over the age of 25, while comparisons with colleagues become more common.

From my own experience and observing others around me, this seems rather obvious. Over the age of 25 you are more likely to have a ‘proper job’ which means you spend a lot more time working than socialising with friends.

Married people and people with children compare themselves more to family members.

Again, rather obvious that those who are consciously building a family are more likely to find family groups important and interact with them more.

Men compare themselves less to family members than women.

The most obvious interpretation for this is that women tend to do most of the domestic work and have a stronger role in bringing up children, therefore family is a bigger part of their lives.

The more people who were exposed to more information media on other people’s lifestyles, the more people were likely to compare.

Again, this is totally obvious. The more you know about other people, the more likely you are to compare yourself with them.

On top of this, people in poorer countries compared themselves more with others, as did people with lower income; the study also shows that people who rate comparing their income with others as more important to them are less happy. No shit, Sherlock.

Differences between reference groups

The most interesting finding of this study was that there was a big difference in happiness implications for those who compared themselves with colleagues as opposed to those who compared themselves with friends and family members. Comparing yourself with colleagues is considerably less detrimental to your happiness. The authors suggest this relates to differing impact of ‘status’ and ‘signal’ types of comparison. ‘Status’ comparison suggests you see a linkage between income and social status, therefore comparisons make you less happy. ‘Signal’ comparisons are noticing the success of people above you on a hierarchical ladder – somewhere you could get too. The inspiring effect of ‘signal’ comparisons to some degree cancels out the negative status envy effect.

Professor Professor Cary Cooper, an expert in organisational psychology and health at Lancaster University Management School, was interviewed by BBC News on the subject. He gave the following interpretation:

“With work colleagues it’s a fairness issue, but with school friends who had the same opportunities as you, you might think ‘They’ve done much better than me so I must be less competent’.”

This is also an interesting recognition of how perceptions of sameness affect people’s comparisons. Work colleagues tend to be more different to you in terms of personality etc. It makes sense you have greater affinity with those you choose to hang out with, i.e. your friends, or share genetic material, life history or upbringing with, i.e. your family. A workplace is also hierarchical – those higher up are often more qualified, more experienced and older – leading you to perceive them more as equivalent to yourself at a later stage of life than as equivalent to your life now. It also raises the question of transparency over pay. People are probably more likely to know what their friends and family are paid than to have detailed knowledge of what their work colleagues are paid, unless they work in a very transparent organisation. Unfortunately the study doesn’t talk about this.

The framing of the debate

The authors, like good statisticians, are careful to avoid imputing causality from these correlations. However, the possibilities they suggest indicate a particular perception concerning the nature of income inequality and its relation to happiness.

In interpreting the relationship between the intensity of comparison and unhappiness, the authors provide two explanations. One option is comparing yourself with others makes you unhappy; the other option given is that unhappy people may compare themselves more with others in order to ‘explain and justify their lower wellbeing.’ Professor Cooper formulates a similar set of hypotheses:

“We need to know what comes first – is it those who are glass half empty types who do the comparisons as a consequence of that, or is it the comparison that makes them unhappy?”

The focus is all on the individual here: either it’s your fault for comparing, or the comparison itself is just an excuse for your intrinsic unhappiness, due to you being a ‘glass half empty type’. There is no suggestion that people with low income could perceive their status as unfair and illegitimate, yet beyond their control due to an unfair socioeconomic system.

Professor Cary’s advice is based on this individualised perception of the problem:

“I would advise people to not compare themselves and be happy with who they are and the situation they’re in.”

This is echoed by the discussion on beliefs regarding the governmental role in redistributing resources – again unsurprisingly, it is those who compare more, as well as those who are poorer, who are more in favour; this is also the case with those who compare more to peers and family. The
authors conclude from this that:

‘Although it is obviously difficult to generalize, a number of these findings are consistent with a greater intensity of income comparisons being associated with a greater demand for redistribution when the latter is partly self-regarding, in the sense that the individuals themselves would likely benefit from it.’

Though they are careful to avoid defining ‘self-regarding’ in a negative and normative sense here, the assumption is that the reason why this might be the case is that the individuals in question would benefit from it. This is of course a fair assumption but it is not the only one: it is also likely that those who are losing in a social systemare more likely to see their situation as unfair, and therefore have a greater belief in societal redistribution of resources, with the corollary of a reduced faith in the market to produce deserving,
meritocratic outcomes.

By individualising the problem, any critique of the social structure is avoided. Instead, the problem is conceptualised as one of envy, desire for personal benefit, or as a reason to justify your own intrinsic unhappiness. None of these explanations present this unhappiness as valid (i.e. that people have a good reason for being unhappy, which perhaps should be addressed). The underpinning assumption is that happiness is the individual’s problem – a hypothesis which has in recent years led to the proliferation of wellbeing self-help books, Cognitive Behavioural Therapy, positive thinking courses and the like.

What is also not discussed is the fact that those who compare more might well be more inclined to see their own social worth in terms of their relative income – and therefore, the comparison makes them unhappy.

Unfortunately, there is also no examination (by direct questioning) of why people consider income comparison important and why they consider income redistribution is important. Why is it left up to the researchers to impute possible reasons for the facts, rather than the people they have asked themselves? It might help to understand the nature of people’s choices if one tries to understand their rationale.

Biases in sociological methods and their social implications

So why are these hypotheses presented? Why are these methods used? Why did no one ask why until after the data collection, from which point it is only possible to speculate? What is the point of statistical studies that prove the obvious?

The first of these questions entails a number of possibilities. The bias towards assumptions of the individual responsibility for personal happiness fits with current intellectual trends that have stimulated the current growth of an industry that provides individual remedies for individual unhappiness.

This industry may exist for a number of reasons:

1) It is attractive to individuals to suppose that they can fix their unhappiness individually;

2) There is a lot of money to be made and this creates an interest in promoting the idea of an individually fixable ‘unhappiness’ – thereby making phenomena to some degree self-reinforcing;

3) Notions of individual responsibility for your fate are an intrinsic assumption of our society – both in an (ostensibly) meritocratic economic system and in our legal system, which places sanctions as a punishment for individual action which individuals bear responsibility for;

4) Such individualised notions are extremely convenient to promote for those who wish to maintain and support the current social and economic system, as advises changing yourself rather than society.

The method of hypothesis testing by statistical data collection which is sometimes presented as more ‘objective’ than qualitative research, is in fact built on a number of normative assumptions which influence the choice in hypotheses, what questions are asked and what conclusions examined.

Some of the reasons behind the use of these methods could simply be practical. It is easier to measure choices or preferences that can be rated on scales along an axis, than it is to measure the reasons why. Although, if reasons were given as multiple choice (with the option of rating their importance also), this would not be much harder to statistically measure than rating the importance of comparing your income with others. So why is this not adopted? Again, practical considerations must be taken into account – the survey data came from the European Social Survey, which collects a whole range of social, demographic and perception data. The data was not collected for the purposes of this specific research question, and likely the authors did not have the resources to conduct such a survey themselves, particularly not across eighteen European countries. These practical reasons may well be the most important.

There is another possible explanation, which relates to notions of objectivity in conducting social research. If you cannot use interviews or other qualitative data for reasons of, why not ask why in multiple choice which allows for ranking of importance? Multiple choice options could be considered as leading questions, biasing response towards assumptions of the survey architect, but here bare statistics of revealed behaviour are without justification considered more objective. The validity of the individual’s perception is distrusted and deemed subjective. This view is echoed by the question asked by both the authors and Professor Carey, as to whether people choose to believe what they believe in order to justify their intrinsic unhappiness.

When it comes to the presentation of possible interpretations, the researchers’ conceptualizations, suggestions and conclusions reflect a specific way of viewing the social world. The danger of this is that this transformation into an ‘objective’ scientific analysis gives their study a validity which will both inform future research and future students. However, this ‘objectivity’, while trying to avoid one kind of bias, projects others instead – biases towards what is easier to measure and what accords with the worldview of the researchers. Their subjective assumptions frame the debate of interpretation and this is what is projected more widely in the media. The method itself gives the people in question, the study subjects, no opportunity to give their own interpretations.

We should be wary of any claim to objectivity. Though the authors may not call themselves objective, their failure to examine their own potential biases does undermine their authority to speak as objective commentators. Their conceptualisation serves to bolster an individualised notion of happiness, where the solutions lie in oneself rather than cultural or social change. I am not here solely to argue the right or wrong of this (though it is probably clear by now where my view lies), but wish to demonstrate what ideas are closed off to the general public by this biased interpretation of findings and the damaging potentialities of having this consecrated as formal knowledge. Science holds great social responsibility, far too infrequently considered even by those who hope to work for the public good.

Annex – list of countries included

Austria Ireland Slovakia
Belgium Netherlands Slovenia
Bulgaria Norway Spain
Finland Poland Sweden
France Portugal Switzerland
Germany Russia United Kingdom


Clark, A. and C. Senik, 05.2010, “Who Compares Who to Whom? The Anatomy of Income Comparisons in Europe,” The Economic Journal, Vol. 120: pp. 573 – 594, available at http://www.res.org.uk/economic/freearticles/2010/may2010.pdf (last accessed 20.04.2011)

Clark, A. and C. Senik, 09.2009, “Who Compares Who to Whom? The Anatomy of Income Comparisons in Europe,” IZA Discussion Paper, No. 4144, available at http://www.politiquessociales.net/sites/politiquessociales.net/IMG/pdf/dp4414.pdf (last accessed 20.04.2011)

Wilkinson, Emma, 29.05.2010, “Comparing Income With Peers Causes Unhappiness,” BBC News, available at http://www.bbc.co.uk/news/10182993 (last accessed 20.04.2011)

[1] Please note that when this article was written the original paper in the journal of economics was not publicly available – it has since been made available. All discussion here is based upon the 2009 discussion paper on the same study, and its reporting on BBC News. The original paper referred to in the BBC article varies in its presentation though not on the major points