Peer effects – the impact of friends, colleagues and family members – are clearly central to social life. But quantifying the causal effect of peers is difficult. Peers operate in the same social environment so are affected by the same common outside influences and the impact of behaviour between two (or more) individuals is generally two way – if two people are friends it is likely that each affects the others behaviour. This makes isolating the effect of one person on another very difficult.
Social scientists have spent their time devising ingenious ways of trying to measure peer effects. A favorite amongst American social scientists is to exploit the fact that individuals at University in the USA generally share rooms and that allocation of roommates is often random within gender. This has allowed them to examine peer effects in smoking, underage drinking, religious beliefs and mental health.
But these studies are limited by the fact that American college students are a special group. In addition, the college roommate design does not allow examination of the effect of a very important group of peers: siblings. The time siblings spend with each other is far larger than the time even close friends spend together.
In recent research undertaken between CMPO and the University of Bergen, we have tried to estimate the effect of siblings on each other’s behaviour. Specifically, we were interested in whether having an older sister who has a teen birth increases the chance that her younger sister has a teen birth too. Teen birth is an important issue and it is clear that these run in families. So isolating the effect of peers from other shared influences, such as family income, attitudes and the more general social environment is also important.
To quantify the effect of an older sister, we exploit an educational reform. In Norway in the 1960s the minimum school leaving age was raised from 14 to 16.But in contrast to many other European countries this did not happen all at once. Instead it was rolled out, pretty randomly, across local areas (municipalities) over a 13 year period. This meant that at any one point during this period, there were some children who could leave school at 14 while others, who had very similar backgrounds and lived in similar areas, had to stay on till 16.
Exploiting this ‘natural experiment’, earlier research found that the extra years of education reduced the chance of a girl having a teen birth. So essentially this gave us a ‘natural experiment’ in teen births. Using this we looked at the effect of the reforms on the probability that an older sister would have a teen birth and then of this teen birth on the chances that her younger sister would also have a teen birth. We found large effects: having an older sister who had a teen birth raised the chance of her younger sister having one too from around one in five to two in five. This effect was larger when the sisters were close in age and where families had less resources. This all makes sense: closer age sisters are more likely to spend time together and girls in families with lower resources have more to gain from sharing the costs of having a child (for example, child care).
Finally, the positive peer effect dwarfed the negative effect of an extra two years at school on teen births. This suggests that if policy makers want to reduce teen births, they need to be able to influence what happens within the family.
The research can be found at http://www.bristol.ac.uk/cmpo/publications/papers/2011/abstract262.html
Last week’s Green Paper set out the government’s strategy for encouraging people to give time and money – part of its vision of a Big Society. There were few concrete ideas – beyond the suggestion that ATMs provide prompts to donate – but instead a set of guiding principles: Great opportunities, Information, Visibility, Exchange and reciprocity and Support (GIVES). In a nutshell – people need new and exciting ways to give (such as at ATMs) and they need to know about them; their giving needs to be visible and it needs to be valued. This is a potentially exciting time for the sector – even as many are worried about the effects of government spending cutbacks – it provides fertile ground for experimentation to see what works and what does not. In designing potential pilots, there is a growing body of evidence to build on.
A number of field experiments with individual charities have found successful triggers that can encourage people to give – these include announcing lead donations, providing a match, rewarding donors with small gifts, making donations public and telling people how much others have given. The findings have led people to generalize about what motivates people to give – signals of quality for individual charities, the desire for prestige, reciprocity etc – yet single-charity studies can only tell us about what motivates people to give to specific charities. None of these studies has looked at whether the triggers simply cause people to give more to charity A at the expense of charity B. There is a real risk that all the shiny new opportunities simply cause people to change the way that they give and a need to show that new schemes increase total giving, not just shuffle it around. To achieve that, there needs to be more understanding of what the real barriers are to people giving – and what can be done to eliminate them.
One of the big ideas in behavioural economics is that defaults can have a powerful effect on people’s behavior in overcoming inertia; they have been shown to work in relation to employer pensions with auto-enrolment leading to big increases in participation. Payroll giving is an obvious extension that we hope to test. But one important lesson from past research is that the detail of the default matters – a low default, while increasing participation, could lead to some people reducing the amount that they give. There also needs to be evidence that a scheme that helps someone to help others can work as well as a scheme that helps someone to help their future selves.
The emphasis on visibility accords with the findings from research which finds positive peer effects. For example, Frey and Meier (2004) found that when students were told that a higher proportion of past students had donated to a good cause, 64 per cent compared to 46 per cent, this had a positive effect on the proportion who gave – but the increase was small at around 2 percentage points. But, as with defaults, providing so-called “social information” can have a negative effect if the amounts that others have given are low. Alpizar et al (2008) found that informing people about a low modal donation increased participation but reduced the average donation (compared to no social information). More interestingly for the ATM proposal, suggestions to give particular amounts that are imposed from above have been found to have a negative effect. Alpizar & Martinsson (2010), show that compared with a social reference that comes from peers, a suggestion from the charity reduced both the probability of giving and the conditional amount given.
Evidence from ultimatum games in the lab, and from elsewhere, suggests that individuals have a preference for fairness, and that this preference is a driving force behind their charitable donations. Based on this, and the increasing prevalence of ideas such as a “Robin Hood tax” and “UK Uncut”, suggesting a belief that corporations, and in particular banks, are bearing too little of the burden of economic hardship, would seem to suggest that the encouraging charitable giving through boxes emblazoned with the logos of banks may not have the desired effect. One response to the ATM suggestion on BBC’s Have Your Say website was that “If bank cards started nagging me to donate I think I’d give LESS not more”.
 Frey, B. & Meier, S. (2004) “Social Comparisons and Pro-social Behavior: Testing “Conditional Cooperation” in a Field Experiment” American Economic Review Vol 94 No 5 pp 1717-1722
 Alpizar, F., Carlsson, F. & Johansson-Stenman, O. (2008) “Anonymity, reciprocity, and conformity: Evidence from voluntary contributions to a national park in Costa Rica” Journal of Public Economics Vol 92 issues 5-6 pp1047-1060
 Alpizar & Martinsson (2010) “Don’t tell me what to do, tell me who to
follow! – Field experiment evidence on voluntary donations” Working Papers in Economics No. 452
 Barr & Zeitlin (2010) “Dictator Games in the lab and in nature: External validity tested and investigated in Ugandan Primary Schools” CSAE WPS/2010-11