FRENITESDE

Research & Evidence

Search for a method

Search titles, questions, territories and MSC identifiers.

40 results
  1. MSC-P-001How do you turn a marketing claim into a testable question?Decision Science↗
  2. MSC-P-002Correlation or causality: what can an analysis actually support?Marketing Measurement↗
  3. MSC-P-003How should uncertainty in a marketing result be expressed?Decision Science↗
  4. MSC-P-004Statistical significance or effect size: which result should be interpreted?Decision Science↗
  5. MSC-P-005How do you measure a marketing construct that is not directly observable?Market Research↗
  6. MSC-P-006How do you design and validate a measurement scale?Market Research↗
  7. MSC-P-007Alpha or omega: how should scale reliability be assessed?Market Research↗
  8. MSC-P-009PCA, EFA or CFA: which method should you choose?Market Research↗
  9. MSC-P-010When should you run a marketing experiment?Marketing Measurement↗
  10. MSC-P-011How do you design an A/B test that actually estimates an effect?Marketing Measurement↗
  11. MSC-P-012How many observations does an experiment need?Decision Science↗
  12. MSC-P-013How do you measure campaign incrementality with a control group?Marketing Measurement↗
  13. MSC-P-017How do you detect selection, contamination and attrition in an experiment?Marketing Measurement↗
  14. MSC-P-018Predictive or causal regression: what are you trying to estimate?Marketing Models↗
  15. MSC-P-019How do you diagnose a marketing regression before interpreting it?Marketing Models↗
  16. MSC-P-022How do you estimate price elasticity and its uncertainty?Pricing Science↗
  17. MSC-P-026Logit vs Probit: how do you choose for purchase probability?Customer Science↗
  18. MSC-P-029Which customers have the highest probability of churn?Customer Science↗
  19. MSC-P-027TAM, UTAUT or UTAUT2: which framework should be used to study technology acceptance?Market Research↗
  20. MSC-H-001Measurement and causality: how can a marketing effect be established?Marketing Measurement↗
  21. MSC-H-002Marketing response models: shape, delay and saturationMarketing Models↗
  22. MSC-H-003Pricing science: connecting price, demand and contributionPricing Science↗
  23. MSC-H-004Customer and choice science: behavior, value and heterogeneityCustomer Science↗
  24. MSC-H-005Measurement science: building valid indicatorsMarket Research↗
  25. MSC-H-006Statistical decision methods: choose, quantify, validateDecision Science↗
  26. MSC-P-008How do you validate a marketing measurement scale?Market Research↗
  27. MSC-P-014How do you design a marketing geo experiment?Marketing Measurement↗
  28. MSC-P-015How do you estimate an effect with difference-in-differences?Marketing Measurement↗
  29. MSC-P-020How do you address price endogeneity?Pricing Science↗
  30. MSC-P-021Fixed or random effects: which panel model should you choose?Marketing Models↗
  31. MSC-P-023How do you estimate a demand function?Pricing Science↗
  32. MSC-P-024How do you simulate a price-volume-margin scenario?Pricing Science↗
  33. MSC-P-028How do you estimate CLV with BG/NBD and Gamma-Gamma?Customer Science↗
  34. MSC-P-030How do you analyze retention with a survival model?Customer Science↗
  35. MSC-P-031How do you build a useful customer segmentation?Customer Science↗
  36. MSC-P-032How do you test segmentation stability?Customer Science↗
  37. MSC-P-033How do you validate a marketing forecast?Decision Science↗
  38. MSC-P-034How do you build a Monte Carlo simulation for a marketing decision?Decision Science↗
  39. MSC-P-035How do you model saturation and adstock?Marketing Models↗
  40. MSC-P-039Which statistical test should you choose?Decision Science↗
← All methods
METHOD DOSSIERMSC-P-015Experimentation and causality

How do you estimate an effect with difference-in-differences?

Difference-in-differences compares change in a treated group with change in an untreated group. Identification relies on credible parallel trends without treatment.

Direct answer

Estimate an ATT for groups and periods covered by the design.

Difference-in-differences compares change in a treated group with change in an untreated group. Identification relies on credible parallel trends without treatment.

Model or deliverableDiD = (ȲT,post−ȲT,pre) − (ȲC,post−ȲC,pre)

Callaway & Sant’Anna, 2021Roth et al., 2023

ESTIMAND

Estimation profile

Unit of analysis
Assigned unit × group × periodControlled scientific terminologyunit × group × period
Exact estimand
ATT in a two-group, two-period design
Model or deliverable
DiD = (ȲT,post−ȲT,pre) − (ȲC,post−ȲC,pre)

01—08

Verifiable analysis framework

  1. 01
    Question

    How do you estimate an effect with difference-in-differences?

  2. 02
    Estimand

    ATT in a two-group, two-period design · unit × group × period

  3. 03
    Data

    unit_id · group · period · treatment_start · outcome

  4. 04
    Model or deliverable

    DiD = (ȲT,post−ȲT,pre) − (ȲC,post−ȲC,pre)

  5. 05
    Declared calculation

    DiD = (ȲT,post−ȲT,pre) − (ȲC,post−ȲC,pre) → ATT in a two-group, two-period design

  6. 06
    Uncertainty checks

    Pre-trends · Anticipation · Use group-time ATT for staggered adoption · Clustered uncertainty

  7. 07
    Method-specific validation

    Check data, estimate stability and interpretation limits. Pre-trends · Anticipation · Use group-time ATT for staggered adoption · Clustered uncertainty

  8. 08
    Limitations

    Prove future parallel trends from pre-period data alone.

Required variables

  • unit_id
  • group
  • period
  • treatment_start
  • outcome

Checks · Controlled scientific terminology

  • Pre-trends
  • Anticipation
  • Use group-time ATT for staggered adoption
  • Clustered uncertainty
Possible decision

Estimate an ATT for groups and periods covered by the design.

Not supported

Prove future parallel trends from pre-period data alone.

Scientific sources

2 sources

  1. Callaway & Sant’Anna, 2021
  2. Roth et al., 2023

Dataset · Tool

Dataset · MSC-005Promotion incrementality→Tool · MSC-T02DiD calculator→

Method connections

Parent territoryMeasurement and causality: how can a marketing effect be established?RequiresCorrelation or causality: what can an analysis actually support?Compare withHow do you design a marketing geo experiment?

Read next

MSC-H-001Measurement and causality: how can a marketing effect be established?→MSC-P-002Correlation or causality: what can an analysis actually support?→MSC-P-014How do you design a marketing geo experiment?→MSC-P-017How do you detect selection, contamination and attrition in an experiment?→