Smart Methods for Complex Policy Evaluation

I would like to thank Wayne Wakeland and his colleagues in the systems science program at Portland State University for the chance to present at their noon systems science and complexity seminar.

I presented on a new avenue of methods development that my colleague Corey Schimpf (Department of Engineering Education, University at Buffalo) first identified, which he and I are calling smart and approachable methods or AM-Smart for short.

My presentation was a quick summary of a paper on the topic we have in review and its application to policy evaluation. In particular, I focused on our computational modelling software, COMPLEX-IT.


TITLE: Smart Methods for Complex Policy Evaluation
ABSTRACT:  Advances in the integration of smart technology, computational modelling and statistical platforms has created a new methods environment, approachable modelling and smart methods – AM-Smart for short. The AM-Smart environment consists of bespoke tools that facilitate user-driven learning of a topic, creating an intuitive, supported but open-ended environment designed to solve specific tasks. Unlike most statistical platforms, AM-Smart methods focus on a single technique or small network of closely interrelated methods (mostly computational in focus), which help users to simultaneously use and learn new methods. They do so by providing scaffolding while allowing for user-exploration, rapid and formative feedback and by requiring modest technical skill, while still being rigorous, authentic, and reliable. A major focus of AM-Smart methods is policy evaluation, as demonstrated in participatory systems mapping tools, complex evaluation toolkits, and R shiny programmes and fast ABM modelling. For this talk, I will introduce the world of AM-Smart methods and their value for policy evaluation. I will specifically introduce a package we have developed called COMPLEX-IT.


Examples of AM-Smart methods are as follows:

COMPLEX-IT, a computational modelling environment for case-based modelling and policy evaluation for non-experts.


SOMbrero for artificial intelligence


PRSM for participatory systems mapping


SAGEMODELER for learning systems dynamics through designing models


FactoShiny for statistical analysis and visualization


Occam, a Discrete Multivariate Modeling (DMM) tool based on the methodology of Reconstructability Analysis (RA).


COMPASS for comparative methods software


NetLogo Web for agent-based models


MAIA for designing multiagent models for institutional analysis


Cytoscape for modeling complex networks


Several platforms that meet many of the AM-smart attributes but may be less representative on a few attributes including


IQAir, the world's largest real-time, air-quality information platform, including a suite of tools – IQAir Earth, Map, App and AirVisual


UK CDRC Mapmaker, including indices of multiple deprivation


Gapminder for exploring global trends statistically and visually