arXiv subjects

Akisato Suzuki

Publications and source records attributed to Akisato Suzuki.

4 recordsLinked to original sources

arXiv2024

Which Type of Statistical Uncertainty Helps Evidence-Based Policymaking? An Insight from a Survey Experiment in Ireland

Which type of statistical uncertainty -- statistical (in)significance with a p-value, or a Bayesian probability -- enables people to see the continuous nature of uncertainty more clearly in a policymaking context? An original survey experiment used a hypothetical scenario, where participants from Ireland were asked whether to introduce a new bus line to reduce traffic jams, given a research report estimating its effectiveness. The treatments were uncertainty information: statistical significance with a p-value of 2%, statistical insignificance with a p-value of 25%, the 95% probability that the estimate is correct, and the 68% probability that the estimate is correct. In the case of lower uncertainty, both significance and Bayesian frameworks resulted in a large proportion of participants adopting the policy (0.82 and 0.91 respectively). In the case of higher uncertainty, the significance framework led a much smaller proportion of participants to adopt the policy (0.39 against 0.83). The findings suggest participants saw the continuous nature of uncertainty more clearly in the Bayesian framework than in the significance framework.

stat.OT↗

arXiv2022

Presenting the Probabilities of Different Effect Sizes: Towards a Better Understanding and Communication of Statistical Uncertainty

How should social scientists understand and communicate the uncertainty of statistically estimated causal effects? I propose we utilize the posterior distribution of a causal effect and present the probability of the effect being greater (in absolute terms) than different minimum effect sizes. Probability is an intuitive measure of uncertainty for understanding and communication. In addition, the proposed approach needs no decision threshold for an uncertainty measure or an effect size, unlike the conventional approaches. I apply the proposed approach to a previous social scientific study, showing it enables richer inference than the significance-vs.-insignificance approach taken by the original study. The accompanying R package makes my approach easy to implement.

stat.AP↗

arXiv2022

Policy Implications of Statistical Estimates: A General Bayesian Decision-Theoretic Model for Binary Outcomes

How should we evaluate the effect of a policy on the likelihood of an undesirable event, such as conflict? The significance test has three limitations. First, relying on statistical significance misses the fact that uncertainty is a continuous scale. Second, focusing on a standard point estimate overlooks the variation in plausible effect sizes. Third, the criterion of substantive significance is rarely explained or justified. A new Bayesian decision-theoretic model, "causal binary loss function model," overcomes these issues. It compares the expected loss under a policy intervention with the one under no intervention. These losses are computed based on a particular range of the effect sizes of a policy, the probability mass of this effect size range, the cost of the policy, and the cost of the undesirable event the policy intends to address. The model is more applicable than common statistical decision-theoretic models using the standard loss functions or capturing costs in terms of false positives and false negatives. I exemplify the model's use through three applications and provide an R package.

stat.ME↗

arXiv2022

Uncertainty in Grid Data: A Theory and Comprehensive Robustness Test

This article makes two novel contributions to spatial political and conflict research using grid data. First, it develops a theory of how uncertainty specific to grid data affects inference. Second, it introduces a comprehensive robustness test on sensitivity to this uncertainty, implemented in R. The uncertainty stems from (1) what is the correct size of grid cells, (2) what is the correct locations on which to draw dividing lines between these grid cells, and (3) a greater effect of measurement errors due to finer grid cells. My test aggregates grid cells into a larger size of choice as the multiple of the original grid cells. It also enables different starting points of grid cell aggregation (e.g., whether to start the aggregation from the corner of the entire map or one grid cell of the original size away from the corner) to shift the diving lines. I apply my test to Tollefsen, Strand, and Buhaug (2012) to substantiate its use.

stat.ME↗