Enter Analysis Inputs
Example Data Table
| Period | Treatment | Control | Phase |
|---|---|---|---|
| Week 1 | 120 | 98 | Pre |
| Week 2 | 124 | 100 | Pre |
| Week 3 | 123 | 101 | Pre |
| Week 4 | 126 | 103 | Pre |
| Week 5 | 128 | 104 | Pre |
| Week 6 | 127 | 105 | Pre |
| Week 7 | 142 | 107 | Post |
| Week 8 | 145 | 108 | Post |
| Week 9 | 149 | 110 | Post |
| Week 10 | 151 | 111 | Post |
| Week 11 | 154 | 112 | Post |
| Week 12 | 156 | 113 | Post |
Formula Used
1. Treatment change: Treatment Post Mean − Treatment Pre Mean
2. Control change: Control Post Mean − Control Pre Mean
3. Counterfactual post mean: Treatment Pre Mean + Control Change
4. Absolute effect: Observed Treatment Post Mean − Counterfactual Post Mean
5. Relative effect: (Absolute Effect ÷ Counterfactual Post Mean) × 100
6. Difference-in-differences estimate: (Treatment Post − Treatment Pre) − (Control Post − Control Pre)
7. Confidence interval: Effect ± z × Standard Error
This calculator uses a practical causal impact framework based on treatment and control trends. It is useful for structured impact screening, campaign evaluation, pricing studies, policy changes, and intervention reviews where a comparable control series exists.
How to Use This Calculator
- Enter a clear intervention name and the measured outcome metric.
- Paste the treatment series for both pre and post periods.
- Paste a comparable control series for the same periods.
- Make sure pre counts match, and post counts also match.
- Select the desired confidence level for interval estimation.
- Click the calculate button to view impact estimates.
- Review the observed, counterfactual, and uncertainty results.
- Use the CSV or PDF buttons to export findings.
Frequently Asked Questions
1. What does this calculator estimate?
It estimates the likely effect of an intervention by comparing treatment changes against control changes. This creates a practical counterfactual benchmark for the post period.
2. Why is a control series important?
A control series helps account for background movement unrelated to the intervention. Without it, ordinary trend changes can be mistaken for causal impact.
3. What is the counterfactual post mean?
It is the estimated post-intervention treatment outcome that would be expected if the intervention had not occurred, based on the observed control movement.
4. How should I choose the control group?
Choose a series that is unaffected by the intervention but behaves similarly before the intervention. Strong pre-period alignment improves credibility.
5. What does pre-period correlation show?
It shows how closely treatment and control moved together before the intervention. Higher positive correlation usually suggests a more reliable comparison structure.
6. Does this replace a full econometric model?
No. It is a strong screening and reporting tool, but complex studies may require regression, Bayesian structural time series, or additional diagnostics.
7. What if the confidence interval crosses zero?
That means the estimated impact is not clearly separated from zero at the chosen confidence level. The intervention may be weak, noisy, or uncertain.
8. Can I use daily, weekly, or monthly data?
Yes. The calculator works with any consistent time unit. Just ensure treatment and control periods are aligned and measured on the same scale.