Short answer
Statistical reasoning helps people choose methods whose assumptions fit the question, quantify uncertainty, and avoid conclusions the available evidence cannot support.
About Statistical reasoning
Statistical reasoning uses statistical methods and assumptions appropriately to quantify patterns and uncertainty. It distinguishes measured effects from noise and communicates the limits of inference.
Use this competency for
- Roles that estimate effects, compare populations, forecast outcomes, or interpret uncertain evidence.
- Work where sampling, variation, bias, or model assumptions materially affect a decision.
Do not use this competency for
- Use data analysis when descriptive evidence and decision framing matter more than selecting or validating statistical methods.
Important distinctions
Data analysis
Data analysis answers a defined question from data, while statistical reasoning focuses on assumptions, variation, and valid inference.
Experiment analysis
Experiment analysis applies causal methods to controlled tests, while statistical reasoning also covers observational and predictive settings.
Expectations by level
IC1
IC1: Applied methods
Applies an established statistical method to a well-defined question with guidance. Checks named assumptions and reports estimates with their uncertainty.
Observable behaviors
- Identifies the population, measure, and comparison before running a method.
- Checks the documented assumptions required by the selected method.
- Reports effect size and uncertainty instead of relying on a threshold alone.
Examples
- Compares two process times with the approved method and notes that the smaller group produces a wider interval.
- Builds a forecast from a standard model and reports the expected range alongside the point estimate.
IC2
IC2: Method selection
Independently selects and validates methods for ambiguous team-level questions. Examines bias, sensitivity, and alternative explanations before drawing an inference.
Observable behaviors
- Chooses a method based on the data-generating process and decision question.
- Runs sensitivity checks when conclusions depend on exclusions or assumptions.
- Explains sources of bias and how they affect the direction of a conclusion.
Examples
- For a survey result, evaluates nonresponse and weighting before comparing departments.
- When estimating a program effect from observational data, documents confounders and tests whether results change under alternate specifications.
IC3
IC3: Statistical direction
Defines statistical approaches for complex work spanning teams and reviews whether methods support the claims made from them. Establishes reusable checks for uncertainty and bias.
Observable behaviors
- Designs analysis strategies when data sources have different sampling or measurement limits.
- Reviews high-impact models for assumption failures, leakage, and unsupported inference.
- Sets reporting standards that separate practical importance from statistical uncertainty.
Examples
- Creates a common uncertainty approach for forecasts consumed by operations and finance.
- Reviews a multi-market impact study, identifies selection bias, and narrows the claim to what the design can support.