Interrupted Time Series Design

Graduate Depth 76 in the knowledge graph I know this Set as goal
Unlocks 1 downstream topic
quasi-experimental policy causal temporal

Core Idea

Interrupted time series (ITS) exploits sharp temporal boundaries when an intervention is implemented to the entire population. The design measures outcome trajectories before and after the intervention, estimating both level changes and trend changes. Unlike experiments, ITS requires no control group but demands careful specification of time trends and control for alternative explanations (simultaneous shocks). ITS is prevalent in policy evaluation, public health, and organizational change research.

Explainer

From your work with time series data, you know that outcomes in social settings rarely jump around randomly — they tend to follow trends, seasonal rhythms, and autocorrelated trajectories. Interrupted time series exploits this regularity. The core idea is simple: if a series was following a stable trajectory before an intervention, you can project that pre-intervention trajectory forward and ask whether the post-intervention data diverge from what would have been predicted. The divergence is your estimate of the intervention's effect. ITS makes the "what would have happened without the intervention?" question concrete by letting the pre-period trend answer it.

There are two distinct effects to estimate: the level change (did the series jump up or down immediately when the intervention hit?) and the slope change (did the trend accelerate or decelerate after the intervention?). Consider a city that bans tobacco advertising. Traffic-accident fatalities (an unrelated series) should show no discontinuity — but lung cancer diagnoses might show a slow trend change as smoking rates gradually fall. These two effects have different policy interpretations: a level change suggests an immediate shock; a slope change suggests a gradual mechanism. Your time-series cross-section background equips you to model both with a segmented regression that includes an indicator for post-intervention periods and an interaction term capturing the trend shift.

The critical threat to validity is simultaneous shocks — other things that changed at the same moment as your intervention. A natural experiment on a law passed in January will be confounded by anything else that happened in January. The classic defense is a control series: a comparable unit or outcome that was not subject to the intervention but would have been affected by the same historical confounders. If tobacco advertising was banned only in one state, the same-month trend in a neighboring state serves as a control; if lung cancer trends diverge only in the treated state, you gain confidence. When no control series is available, you must argue persuasively that no plausible alternative explanation coincides with the intervention's timing — a harder argumentative burden.

ITS has a natural home in policy evaluation precisely because many policies are implemented population-wide at a specific moment: speed limit changes, vaccine rollouts, sentencing reforms, financial regulations. Randomization is impossible; difference-in-differences may lack clean controls; regression discontinuity requires a continuous assignment variable. ITS asks only for a sharp temporal boundary and a long enough pre-period to establish the baseline trend reliably. The longer and more stable the pre-intervention series, the more credible the counterfactual projection — which is why ITS studies in public health routinely use 24–48 months of pre-data and why administrative data archives matter so much to the design's feasibility.

What did you take from this?

Topics in reflective domains aren't scored by quiz answers. Read, reflect, and mark when you've thought it through.

Quiz me anyway →

Prerequisite Chain

Counting to 10Counting to 20Understanding ZeroThe Number ZeroCounting to FiveOne-to-One CorrespondenceCombining Small Groups Within 5Addition Within 10Addition Within 20Two-Digit Addition Without RegroupingTwo-Digit Addition with RegroupingAddition Within 100Repeated Addition as MultiplicationMultiplication Facts Within 100Division as Equal SharingDivision as Grouping (Measurement Division)Division: Grouping (Repeated Subtraction) ModelDivision: Fair Sharing ModelDivision as Equal SharingDivision as GroupingBasic Division FactsDivision Facts Within 100Two-Digit by One-Digit DivisionDivision with RemaindersRemainders and Quotients in DivisionDivision Word ProblemsIntroduction to Long DivisionFactors and MultiplesPrime and Composite NumbersEquivalent FractionsRelating Fractions and DecimalsDecimal Place ValueReading and Writing DecimalsComparing and Ordering DecimalsAdding and Subtracting DecimalsMultiplying DecimalsDividing DecimalsDividing FractionsMixed Number ArithmeticOrder of OperationsInteger Order of OperationsVariable ExpressionsCombining Like TermsOne-Step EquationsTwo-Step EquationsSolving Multi-Step EquationsEquations with Variables on Both SidesAngle Pairs: Complementary, Supplementary, and VerticalParallel Lines and TransversalsCorresponding AnglesAlternate Interior AnglesTriangle Angle Sum TheoremExterior Angle TheoremTriangle Inequality TheoremSimilar Triangles: AA SimilaritySimilar Triangles: SSS and SAS SimilarityProportions in Similar TrianglesRight Triangle Trigonometry IntroductionTrigonometric Ratios ReviewRadian MeasureConverting Between Degrees and RadiansThe Unit CircleGraphing Sine and CosineGraphing Tangent and Reciprocal Trigonometric FunctionsDerivatives of Trigonometric FunctionsAntiderivativesIndefinite IntegralsBasic Integration RulesRiemann SumsDefinite Integral DefinitionProbability Density Functions and Continuous DistributionsCumulative Distribution FunctionsContinuous Random VariablesProbability Density FunctionsCausal Inference from Observational DataNatural Experiments and Quasi-Experimental DesignInterrupted Time Series Design

Longest path: 77 steps · 434 total prerequisite topics

Prerequisites (2)

Leads To (1)