Progressive Clarity

Methods and Technology

Progressive Clarity leverages traditional and cutting-edge methods and technology to provide unparalleled detail and accuracy to our clients' micro-targeting, fundraising, and advocacy efforts. We combine statistical analysis, GIS mapping, and agent-based simulations to generate rich models of individual behavior, identifying and motivating prospective supporters and contributors. A brief overview of the methods that we employ, include:

microtargeting through predictive market segmentation, to harness highly detailed consumer, demographic, and behavioral information about individual voters to identify potential supporters;
GIS mapping and 3D visualization, to match individuals with school-districts and legislative districts, examine neighborhood effects and spatial clustering, and provide pinpoint targeting and mobilization for policy change;
econometrics (linear and logistic regression, maximum likelihood estimation discriminant analysis, ANOVA, spatial and network autocorrelation analysis, etc.), to:

develop statistical models of individual behavior,
make inferences across multiple diverse data sources,
analyze the uncertainty of model predictions, and
determine the impact of client activities on the behavior of the system;

advanced simulation modeling, from Monte Carlo simulation to agent-based simulations where each targeted citizen and his/her decision making process is modeled, capturing the dynamics of the target population in an intuitive, readily extensible, framework;
computational laboratory analysis, for intelligent sensitivity analyses and model optimization; and
game theory, for alternative specifications of models, integration of existing research, and incorporation of strategic considerations within the agent-based simulations.

Progressive Clarity leverages the strong foundation of tools and techniques already in place and adds our cutting-edge data analysis, simulation and optimization tools.

For the first step, we employ traditional econometric and microtargeting analyses on the target population to generate statistical models of individual behavior. We draw together diverse data sources including voter file and Catalist records, IRS contribution records, the nationwide Current Population Survey's Volunteer, and School supplements and the Indiana Center on Philanthropy's survey of charitable giving and volunteering over time, as well as more traditional and US Census and FEC data. We then use GIS tools to understand the spatio-demographic terrain of the client's project, mapping the target population and relevant legal and electoral jurisdictions in a series of intuitive images.

Next, we deploy our detailed analysis of individual behavior in an agent-based model of the entire target population. Unlike existing service providers, we model the individual decision-making processes of citizens and place these millions of individuals in a rich virtual environment. Progressive Clarity has pioneered the use of agent based modeling and simulation to improve target efficiency for political and issue advocacy campaigns. The use of this nascent technology significantly differentiates Progressive Clarity from our competitors. The process is extremely computationally "expensive" in that it requires significant computing resources, but allows Progressive Clarity to provide clients with far greater target efficiency, enabling them to wisely spend their limited resources to achieve greater impact.

For example, in electoral campaigns the voters interact with the candidates' campaigns, third-party organizations and issue advocates, media companies, and local, state, and national political parties. We model the project's progression over time and space, for example, as voters learn about the candidate, reach their contribution limits for direct contributions, and media companies adjust their prices to demand. We rigorously calibrate this model to diverse streams of historical data, in a process similar to, but more flexible than, traditional econometric estimation.

Finally, we place the agent-based model in a computational laboratory framework to optimize the client's outreach and fundraising strategy, and estimate the uncertainty surrounding each action. We provide a selection of targeting, mobilization, and fundraising options at each stage of the project to provide our clients with the analysis they need to make well informed decisions.