In Choice-based economics, data types specify the structural forms in which individual decision behavior is recorded prior to any analysis or modeling. These forms determine what variation can be observed, what patterns are detectable, and what information is irretrievably lost at capture. Data types therefore fix the evidentiary limits of choice-based inquiry.
1. Time Series
Choice evidence commonly appears as time-indexed sequences of decisions.
These include repeated choices by an individual across trials, response times recorded over successive tasks, or longitudinal panels tracking decisions across periods. Time series structure enables observation of learning, adaptation, persistence, and temporal dependence, but constrains inference to behavior observable at the chosen temporal resolution.
2. Spectra and Distributions
Choice data often take the form of distributions over options, responses, or outcomes.
Examples include choice frequency distributions, response-time distributions, payoff distributions, or empirical choice probabilities aggregated across trials or individuals. These formats capture variability and stochastic structure while discarding information about sequential order and individual trajectories.
3. Images, Volumes, and Spatial Fields
Spatial data are not primary evidence forms in Choice-based economics.
When present—such as eye-tracking heatmaps or spatial choice maps—they are auxiliary encodings layered on top of decision records. Spatial formats typically represent derived visualizations rather than raw choice data and must be explicitly justified when treated as evidence.
4. Tables, Counts, and Matrices
The dominant raw data form in Choice is structured tabular data.
Examples include individual-by-option choice matrices, payoff tables, response count tables, and panel datasets indexed by individual, option, and condition. This format preserves categorical structure and supports systematic comparison, while abstracting away narrative or contextual detail.
5. Curves and Derived Plots
Choice data are frequently summarized as curves constructed from underlying records.
Examples include demand curves, indifference curves, learning curves, and response functions. These are derived representations, not raw evidence. Their validity depends entirely on the structure, resolution, and completeness of the underlying choice data.
6. Symbolic or Structural Data
Choice evidence may include formal or encoded representations of decision structure.
Examples include preference orderings, encoded decision rules, constraint representations, or symbolic choice trees. These formats preserve logical structure but require explicit mappings to observed decision events to remain evidentially grounded.
Structural Limits (Choice)
Across all data types, Choice-based evidence is fundamentally:
- Individual-level
- Event- or trial-based
- Categorical or numeric
- Bounded by task-defined resolution
Information not captured at this stage—such as unobserved intention or system-level effects—cannot be recovered downstream.