Data Science · Advanced level

Advanced level · Module 1 of 6 · 22 guided hours

Causal Inference and Programme Evaluation

Move from association to cause with experiments and quasi-experimental designs, and understand their assumptions.

Learning objectives

  • Represent causal assumptions with diagrams
  • Analyse randomised evaluations and A/B tests
  • Apply difference-in-differences and recognise other quasi-experimental designs

Topics

  • counterfactuals
  • causal diagrams (DAGs)
  • randomised controlled trials and A/B tests
  • difference-in-differences
  • matching and regression discontinuity (overview)
  • assumptions and sensitivity

Lessons

  1. Lesson 1 Thinking Causally

    Frame causal questions with counterfactuals and draw causal diagrams to reveal confounders.

    counterfactualcausal effectdirected acyclic graphconfounder vs mediator

  2. Lesson 2 Randomised Evaluations and A/B Tests

    Analyse randomised experiments correctly and understand why randomisation gives credible comparisons.

    random assignmentbalance checksstatistical powerintention to treat

  3. Lesson 3 Difference-in-Differences

    Estimate effects from before-and-after data in treated and comparison groups, and check the parallel-trends assumption.

    before–after comparisoncomparison groupparallel trendsregression form of DiD

  4. Lesson 4 Other Designs and Sensitivity

    Recognise matching and regression discontinuity designs and test how sensitive conclusions are to assumptions.

    propensity score matchingregression discontinuitysensitivity analysisexternal validity

Practical activity

Estimate the effect of a school feeding programme on attendance with difference-in-differences, check the parallel-trends assumption and write the limits.

Expected outcomes

  • Sound causal reasoning
  • Correctly applied evaluation designs

Assessment

Evaluation report marked on design choice, assumption checks and careful conclusions.

Taught live by an instructor — online or in person. Free course lessons are for self-study and do not lead to a certificate — see registration options.