Subscribe to DSC Newsletter

Official statistics and bounded rationality research questions

During a session of the 46th session of the UN Statistical Commission on the the post-2015 development agenda, UN Deputy Secretary-General Jan Eliasson said data will be the “lifeblood of decision-making and the raw material for accountability” in the new agenda and called for a statistical framework that would meet such expectations ( Statistics has always been presented as a support to decision making, whether is is official statistics or statistics collected for monitoring and evaluation purposes. There is here the implicit assumption that, if we provide the right statistical information to decision-makers they will make full use of it in order to make rational decisions. That's exactly what theories of bounded rationality are contesting: for various reasons people do not use fully rational problem-solving methodologies even in the presence of full information. This has been shown by many authors, among which the work of Herbert Simon is considered as seminal. As shown by further research, reasons for not using fully rational-problem solving methodology with full information include limited time or the high computational cost of fully rational methodologies.

Can a discipline that's objective is to provide information for better decision-making ignore well-known flaws in the use of that information for rational decisions? Of course no. Furthermore, the advent of big data is likely to increase one of the causes of heuristics biases: the overflow of information. If computer power does not follow the increase in the flow of information, people will never be able to analyze the massive amount of data and will necessarily rely on various heuristics using only part of the available information. For these reasons, integrating knowledge on bounded rationality can certainly help official statistics. If the ultimate objective of official statistics is better decision-making then providing bias-correcting information may be even more effective than producing big amount of data that will not be used in a fully rational manner. 

Here are some questions related to bounded rationality and which answers can help improving official statistics, in particular in the era of big data:

  • Is providing correct data enough as mission of official statistics for improving decision-making when these data are used with flawed methodologies?
  • Should official statistics consider correcting heuristic biases in decision-making one of it's central objectives? How does it relate to the issue of statistical literacy?
  • Research has shown that simply making people aware of the biases in their decision-making heuristics does not eliminate these biases. Should official statistics be also concerned with the way the data are presented to user? 
  • Should official statistics be concerned with how the available information impacts heuristic biases and define its priorities in data production based on that knowledge?
  • Should official statistics be concerned with systematically comparing predictions and outcomes in order to identify possible biases and see whether the availability of data or the way the data are presented can correct some of these biases?
  • Should official statistics systematically try to identify biases in previous decision-making every time new data are available?
  • Will the advent of big data reduce heuristic biases or increase them by making the computational cost even higher given the amount of available data?
  • Does the existence of heuristic biases have an implication on the usefulness of SDGs statistics? In general, how do heuristic biases impact the usefulness of synthetic statistical indices? Are we some times increasing decision-making biases instead of promoting evidence-based decision-making?
  • Are we creating heuristics biases by trying to build statistical capacity by sectors and by collecting and publishing data by sector rather than by using an integrated approach?
  • Are organizations creating heuristic biases by insisting on the usefulness of their data only? Are they doing it sometimes deliberately in order to give more weight to the data in their areas in decision-making? What are the implications in terms of coordination of statistical activities?
  • What are the implications of the answers to all above questions on statistical development in general and statistical capacity building in particular? Do we have to readjust some tools in the existing toolbox to take into account the existence of heuristic biases?

There is no doubt that the key to some difficult issues in statistical development may be found in the study of the heuristics used by decisions-makers under 'bounded rationality'. It is even possible that some intractable problems in statistical development are just the result of the bounded rationality in decision-making and that the approaches to correcting these issues by providing objective information have failed just because they have always assumed full rationality in decision-making. In any case, with the advent of big data, more than ever,  official statistics should take into account knowledge on bounded rationality in order to fulfill it's mission of improving decision-making for pursuing sustainable development goals.

That's why study of bounded rationality and the heuristics used in decision-making should be a central concern for official statistics, as well as for monitoring and evaluation research.

Special case: Data visualization, bounded rationality and heuristics

Is data visualization linked to heuristics? Does a good visualization present information in a way that it fits an heuristic model and reduces the cost of processing? In this case, a good understanding of the heuristics used by decision-makers can help improve data visualization.

Views: 610


You need to be a member of Data Science Central to add comments!

Join Data Science Central

Comment by Sione Palu on June 19, 2016 at 1:43pm

From economics & physics on this topic of bounded rationality:

"Inductive Reasoning and Bounded Rationality"



  • Add Videos
  • View All

© 2020   Data Science Central ®   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service