Policy Analysis Using Interrupted Time Series Course

Editorial Take

The University of British Columbia's course on Policy Analysis Using Interrupted Time Series, hosted on edX, offers a targeted and technically grounded approach to evaluating policy interventions. Designed for public sector analysts, researchers, and graduate students, it fills a niche in evidence-based policy training by focusing on a powerful quasi-experimental method. With a clear emphasis on practical application, the course equips learners to design, implement, and communicate findings from ITS studies effectively.

While not a broad data science curriculum, it stands out for its methodological precision and relevance to real-world governance challenges. The integration of communication skills ensures that technical findings are accessible to decision-makers, a rare but critical component in policy analytics education. This review dives deep into the course’s structure, strengths, limitations, and strategies for maximizing impact.

Standout Strengths

• Methodological Clarity: The course clearly differentiates interrupted time series from other causal inference methods like regression discontinuity. This helps learners choose appropriate designs based on data availability and policy context. It builds strong conceptual foundations early.
• Policy Relevance: Content is directly applicable to public health, education, and social policy evaluation. Real-world examples ground theory in practice, making it easier to transfer learning to government or NGO roles. Case studies enhance engagement.
• Data Communication: A major strength is teaching how to interpret results for non-technical audiences. This includes simplifying statistical outputs and avoiding misinterpretation, a critical skill for influencing real policy decisions and stakeholder buy-in.
• Visualization Focus: The course emphasizes producing compelling figures that highlight pre- and post-intervention trends. Well-designed graphs are taught as essential tools for storytelling and credibility in policy reports and presentations.
• Structured Learning Path: The five-week format progresses logically from theory to application. Each module builds on the last, ensuring learners develop both analytical and communication skills in tandem. Pacing supports part-time learners.
• Public Sector Alignment: Designed with policy analysts in mind, the curriculum aligns with common evaluation needs in government. It prepares learners to assess program effectiveness, compliance, and long-term impacts using observational data.

Honest Limitations

Software Depth: While modeling concepts are covered, hands-on coding in R or Python is limited. Learners expecting detailed programming tutorials may need supplementary resources. The course prioritizes understanding over implementation.
• Prerequisite Knowledge: Assumes familiarity with basic statistics and regression. Beginners may struggle without prior exposure to time series or causal inference concepts. A refresher on fundamentals would benefit new learners.
• Narrow Scope: Focused exclusively on ITS and RD designs, it may not suit those seeking broad data science skills. The specialization is a strength for policy analysts but a limitation for generalists.
• Assessment Quality: Quizzes and assignments may lack depth in testing applied modeling skills. Without graded coding exercises, learners must self-validate their understanding through external practice.

How to Get the Most Out of It

• Study cadence: Dedicate 4–6 hours weekly to absorb lectures and complete exercises. Consistent pacing prevents overload, especially in modeling weeks. Align study time with work schedules for better retention.
• Parallel project: Apply concepts to a real or hypothetical policy question. For example, analyze crime rates before and after a law change. This reinforces learning and builds a portfolio piece.
• Note-taking: Summarize key assumptions and pitfalls of ITS in your own words. Visual diagrams of time series setups help clarify when the method is appropriate versus when it may mislead.
• Community: Engage in edX discussion forums to compare interpretations of results. Peer feedback improves communication skills and exposes you to diverse policy contexts and challenges.
• Practice: Recreate course figures using free tools like Excel or Google Sheets. Reproducing visualizations builds confidence and reveals nuances in data presentation and scaling.
• Consistency: Complete modules in order without skipping ahead. Each concept builds on prior knowledge, especially when moving from model setup to interpretation and communication.

Supplementary Resources

• Book: 'Impact Evaluation in Practice' by Gertler et al. complements this course with broader evaluation methods. It provides additional case studies and policy contexts for deeper understanding.
• Tool: Use R with the 'segmented' or 'lm' packages to practice ITS modeling. Free and widely adopted, R offers robust support for time series analysis and visualization.
• Follow-up: Explore UBC’s other public policy or data science offerings on edX. Advanced courses in program evaluation or causal inference build directly on this foundation.
• Reference: CDC’s ITS guidelines offer real-world templates for public health evaluations. These provide structured approaches to data collection and reporting standards.

Common Pitfalls

• Pitfall: Misinterpreting correlation as causation in ITS. Without proper controls, trends may be mistaken for intervention effects. The course teaches safeguards, but learners must remain vigilant.
• Pitfall: Overlooking data granularity issues. Using monthly instead of weekly data can mask short-term effects. Always assess the appropriate time scale for the policy being evaluated.
• Pitfall: Failing to account for external events. Other concurrent policies or economic shifts can confound results. The course emphasizes robustness checks, but real-world application requires diligence.

Time & Money ROI

• Time: At 5 weeks and 4–6 hours per week, the time investment is manageable for working professionals. The focused scope ensures no wasted effort on irrelevant topics.
• Cost-to-value: Free to audit, the course delivers high value for budget-conscious learners. The cost-to-skill ratio is excellent, especially for public sector professionals.
• Certificate: The verified certificate has moderate career value, primarily within government or academic circles. It signals methodological competence but may not boost industry tech roles.
• Alternative: Free alternatives exist, but few integrate communication skills with technical modeling. This course’s blend of analysis and presentation is uniquely valuable for policy roles.

Editorial Verdict

This course is a strong choice for policy analysts, public health researchers, and graduate students seeking to strengthen their evaluation toolkit. It successfully bridges the gap between statistical methodology and real-world policy communication, a rare and valuable combination. The focus on interrupted time series provides a rigorous approach to causal inference when randomized trials aren't feasible, making it highly relevant for evaluating real-world programs. While not designed for data science generalists, its targeted curriculum ensures depth over breadth, delivering actionable skills in a short timeframe.

We recommend this course to learners committed to evidence-based policy work who want to move beyond descriptive analytics into causal evaluation. The free audit option lowers the barrier to entry, allowing professionals to assess its value before upgrading for a certificate. To maximize return, pair it with hands-on practice using real datasets and seek peer feedback on visualizations and summaries. With its strong institutional backing from UBC and practical orientation, this course stands out as a specialized but impactful resource in the policy analytics landscape.