TY - GEN

T1 - Cross-method validation as a solution to the problem of excessive simplification of measurement in quantitative IR research

AU - Beach, Derek

PY - 2007

Y1 - 2007

N2 - The purpose of this article is to make IR scholars more aware of the costs of choosing quantitative methods. The article first shows that quantification can have analytical ‘costs’ when the measures created are too simple to capture the essence of the systematized concept that was supposed to be measured. If we create measures that do not accurately map the structure of the systematized concept, we are left with inaccurate and misleading measures that result in biased inferences. This is illustrated by replicating a recent article published in International Organization, and then discussed in more detail based upon a review of the democratic peace literature. I then offer two positive suggestions for a way forward. First, I argue that quantitative scholars should spend more time validating their measures, and in particular should engage in multi-method partnerships with qualitative scholars that have a deep understanding of particular cases in order to exploit the comparative advantages of qualitative methodology, using the more accurate qualitative measures to validate their own quantitative measures. Secondly, quantitative scholars should lower their level of ambition given the often poor quality of the data that they are working with, engaging instead in multi-method partnerships where statistical tools are used to detect possible interesting correlations that are then more systematically investigated using qualitative case study methods.

AB - The purpose of this article is to make IR scholars more aware of the costs of choosing quantitative methods. The article first shows that quantification can have analytical ‘costs’ when the measures created are too simple to capture the essence of the systematized concept that was supposed to be measured. If we create measures that do not accurately map the structure of the systematized concept, we are left with inaccurate and misleading measures that result in biased inferences. This is illustrated by replicating a recent article published in International Organization, and then discussed in more detail based upon a review of the democratic peace literature. I then offer two positive suggestions for a way forward. First, I argue that quantitative scholars should spend more time validating their measures, and in particular should engage in multi-method partnerships with qualitative scholars that have a deep understanding of particular cases in order to exploit the comparative advantages of qualitative methodology, using the more accurate qualitative measures to validate their own quantitative measures. Secondly, quantitative scholars should lower their level of ambition given the often poor quality of the data that they are working with, engaging instead in multi-method partnerships where statistical tools are used to detect possible interesting correlations that are then more systematically investigated using qualitative case study methods.

M3 - Article in proceedings

BT - Ikke angivet

PB - Department of Political Science, University of Aarhus

Y2 - 28 February 2007 through 3 March 2007

ER -