Hyperbolic discounting as an embodiment of a critical realist approach to social research

I've argued previously about the importance of a critical realist approach to social enquiry and the advantages it has over positivist or constructivist perspectives. For context, positivist approaches broadly focus on an objective reality and assume behaviour can be measured in a relatively unproblematic way. In contrast, constructionist approaches emphasise individual perspectives and interpretations and harbour reservations about the possibility of unbiased accounts of the world or even (in the most extreme) whether a social reality exists as such. 

So, a positivist approach to addiction would see it as something that could be measured and explained largely by material and biological factors. A constructionist approach would be much more interested in the historical and cultural meaning of the term 'addiction' and want to understand addicts' experiences from their perspective without necessarily accepting that addiction is related to any real 'thing' in the world. A critical realist would want to look at both behaviour and individuals' experiences and perceptions and how the two interact both with each other and the cultural and historical context. They would assume that addiction could be a real thing though it might be a very complex construct. 

So where does hyperbolic discounting come in? Hyperbolic discounting is a concept that comes from behavioural psychology and is a way of modeling how benefits and costs today are weighed against those in the future. The traditional model (exponential discounting) assumes that a discount rate is applied each day multiplicatively, so any value today is 'discounted' by a rate 'delta' raised to the power of the number of days into the future the cost or reward is delayed. Hyperbolic discounting builds on this model by adding another discount parameter, 'beta' , which is the same for every day in the future. So in terms of delta, the further away in time the benefit or cost is, the less I care about it. In terms of beta, tomorrow counts much less than today, but if it's not today, I don't care if it's tomorrow or any day after that. 

The benefit of this new model is that it reflects the reality that someone may decide to stop taking drugs next week, but when next week arrives, they don’t do it. This is something that exponential discounting can't explain. So far, so positivist. The critical realist part comes in because as well as an actual beta (which researchers can estimate), people have a more or less accurate belief about their own beta. Some people think their beta is 1 when it isn't, so they think they'll be able to give up drugs next week, but when it comes to it they don't. Others have a very accurate belief and know they're going to find it very hard. A third category of people know their delta is less than 1 but don't think it is quite as low as it is. So they know it will be hard, but not as hard as it is in reality. 

This isn't just abstract theory. In an experiment on workers in India, people were given the option of being paid to turn up for a breathalyser test irrespective of the outcome of the test or to only get paid if the outcome of the test was negative. About 30% opted to only get paid if the test was negative. In other words, they had a good idea about their delta and used their belief to inform their behaviour by creating what is known as a 'commitment device'. This is behaviour couldn't be understood within a positivist account of addiction, which focuses only on material and biological factors, nor within a constructionist one, which only focuses on people's subjective experiences. Both material reality and people's perspectives and beliefs need to be explored to understand the complexity of people's lives and behaviour, which is exactly what a critical realist framework facilitates.