Skip to content
No Vape

A vaping-policy synthetic control is a constructed comparison

Find the donor areas, earlier fit and actual outcome before treating a comparison line as a second observed policy world.

No Vape · Updated

Books arranged on a shelf
Illustrative image · Pramod Tiwari / Unsplash

Locate the treated unit and donor pool

The unit might be an area rather than an individual. Ask which place received the intervention, when it began and which other places could contribute to the comparison. Keep the report’s selection rules and outcome definition. A graph with two lines cannot tell you whether the comparison is a single actual area or a weighted combination.

HM Treasury’s Magenta Book annex describes constructing a synthetic comparison using historical relationships with other units and continuing it after implementation. This source defines the method, not a current vaping-policy result. Its historical tobacco example is not repeated as evidence about your quitting outcome or an app’s benefit.

HM Treasury and Evaluation Task Force: Magenta Book Annex A: A2.6 Synthetic control methods

Distinguish earlier fit from later difference

Look at what information was used to construct the comparison before intervention. Then identify the period used to describe divergence afterward. A good-looking later gap cannot by itself establish that the earlier comparison was appropriate. Nor does a similar starting value demonstrate a similar underlying trajectory.

Imagine a fictional region introducing a service arrangement while several other regions supply a weighted comparison. Before the change, their combination follows the region’s reported attendance pattern closely. Afterward the lines separate. That illustrates the structure of the question; the invented graph has no data, justified donor pool or causal result, and attendance is not an abstinence outcome.

Read what could make the donor comparison less credible

Ask what the authors say about changes affecting donor units, differences in measurement and other events in the treated place. Do not infer that every relevant difference was balanced simply because weights were used. The observed history helps construct a comparison; it cannot make the unobserved alternative future directly observable.

Also locate any reported checks of the result and retain their stated scope. A sensitivity display is not a fresh independent population. You need not select donors, calculate new weights or impose a visual pass threshold yourself. If the comparison’s construction is unavailable, say that the graph’s counterfactual basis remains unverified rather than treating its smoothness as proof.

Keep policy inference separate from personal choice

A faithful reading note identifies the treated unit, donor construction, outcome, intervention date and main assumption or unresolved concern. This differs from an ordinary historical-control study using an earlier comparison population and from simply subtracting two groups’ changes. Keep the actual method name and unit with the reported estimate.

A policy estimate concerns its documented population and setting. It cannot supply your personal chance of quitting or validate any product that happened to be available in that period. For treatment decisions, seek qualified interpretation of the relevant evidence. The useful reading output is knowing what the constructed comparison represents and why its later line must remain attached to the authors’ assumptions.

Sources and scope

Sources inform the guidance. The worksheets and examples are No Vape editorial suggestions, not tested treatments.

  1. HM Treasury and Evaluation Task Force — Magenta Book Annex A: A2.6 Synthetic control methods
  2. No Vape — No Vape — app features and store links

Written with AI assistance and checked against the cited sources. No clinical review is claimed. This guide covers practical planning, not individual medical treatment. How we research and write