← All workshopsRead Reweighting Records Without Inventing Missing Evidence

Causal Inference and Experiments

Calculate an inverse-probability contrast

BOTIE

Select treated outcomes through the binary assignment and divide by their supplied propensity.
Ready to test
0 passed · 0 failed · 2 not run

An empty floor. Start with the information your machine needs.

Received result

Your answer will arrive here.

Examples, experiments and workspace controls
Live values

Connect a tool's inputs to see what it produces. You can try a branch before connecting it to Result.

Try your own inputs

Use a copy of the example to try your own data. Keep the problem's assumptions in mind, and predict the effect of a change before checking the output. You can also add another branch to compare two ways of doing the calculation.

Start at an output arrow, then click or drag to an input. Drag empty space to pan. Delete removes a selection. Ctrl/Cmd+Z undoes a change.
Context for this challenge

Select treated outcomes through the binary assignment and divide by their supplied propensity.

Botie has outcomes, binary treatment assignments, and valid supplied propensities.

Inspect the component calculation
Divide(
  Amount: Multiply(
    Value: Binary treatment assignments
    Multiplier: Observed outcomes
  )
  Divisor: Check treatment overlap(
    Propensity vector: Propensities strictly between zero and one
  )
)
Examples and results
Binary treatment assignments
  1. 01
  2. 10
  3. 21
  4. 30
Observed outcomes
  1. 06
  2. 14
  3. 28
  4. 35
Propensities strictly between zero and one
  1. 00.5
  2. 10.5
  3. 20.5
  4. 30.5
Required result
  1. 012
  2. 10
  3. 216
  4. 30
Binary treatment assignments
  1. 01
  2. 10
Observed outcomes
  1. 03
  2. 13
Propensities strictly between zero and one
  1. 00.5
  2. 10.5
Required result
  1. 06
  2. 10
Try a fresh construction

This clears this draft, including its undo history. Earlier successful attempts stay recorded. Build again without opening hints to record an independent attempt.

Stuck? Inspect a worked construction

The tools below form one possible solution. You can assemble it, follow its signals, and undo that change to return to your own construction.

Assumptions and test precision

This uses the unnormalised all-unit mean. Positivity and no unmeasured confounding cannot be established by the arithmetic. Numerical tests allow a difference of 0.001. Lists and structures must also have the right shape.

Your last tested pieces can start the full construction. An existing draft there keeps its own pieces.

Unequal assignment probabilities change the weights

This check uses the pieces from “Keep treated and control contributions separate”.

BOTIE

Botie has four observed outcomes, treatment indicators and each person's probability of treatment. Calculate the unnormalised inverse-probability-weighted treatment difference averaged over all four people. Treated outcomes use their treatment probability; controls use one minus that probability.
Binary treatment assignments
  1. 01
  2. 10
  3. 21
  4. 30
Observed outcomes
  1. 03
  2. 12
  3. 24
  4. 31
Propensities strictly between zero and one
  1. 00.75
  2. 10.5
  3. 20.5
  4. 30.75

Use a number, text, or an array matching the requested result. For a matrix, put each row in brackets. Saving your prediction runs your current construction on these inputs and checks the earlier examples too.

Build a construction in the workspace above first. This case will test those pieces.

Why would a propensity of zero or one make this stated procedure invalid?

Identification needs supported comparisons, and unstable weights warrant investigation before reporting an effect.

This uses the unnormalised all-unit mean. Positivity and no unmeasured confounding cannot be established by the arithmetic.

Work through the Python challenges →
Continue the courseSaving the Whole Prediction Procedure →
← Comparing Treatments in the Same Population