Details

Category: Statistical Software for Quantitative Analysis

Published: Tuesday, 20 October 2026 08:48

Written by: Matt Pye - HWD

Understanding 'Time to Outcomes' in Government

 

Overview

How long does it take someone to find work, access a service or reoffend? Which groups experience an outcome sooner, and what factors are associated with the differences? Understanding when outcomes happen can help government analysts build a fuller picture of policy and service performance.

This practical half-day course introduces survival analysis: a set of statistical methods for analysing the time until an event occurs. Its applications extend beyond health to areas such as employment, justice, business survival and public service waiting times.

Participants will learn how to analyse data when some outcomes have not yet occurred by the end of the observation period. For example, some people may still be unemployed or waiting for a service. The course explains how to use this information appropriately, rather than excluding these cases and potentially producing misleading results.

Topics covered include:

• Recognising time-to-event questions and preparing data for analysis in R.
• Handling incomplete follow-up and outcomes that have not yet occurred.
• Producing and interpreting survival curves and comparing groups.
• Using regression models to examine factors associated with earlier or later outcomes.
• Checking model assumptions and interpreting findings appropriately.

The course combines tutor-led explanation, live coding and guided exercises in R. Participants will leave with a working script they can adapt to their own data. Discussion will connect the methods to government analytical questions; differences between groups will be interpreted as associations, unless an appropriate study design supports a causal conclusion.

This course is suitable for government analysts, statisticians, economists and social researchers with some familiarity with R and basic regression. No previous knowledge of survival analysis is required. Participants will need a laptop with R and RStudio, or a similar development environment, and the survival, ggsurvfit and tidyverse packages installed.

 

Learning Outcomes

By the end of this course, participants will be able to:

• Identify government analytical questions for which survival analysis is useful.
• Explain censoring and why standard summaries can misrepresent time-to-event data.
• Prepare basic time-to-event data and produce Kaplan-Meier curves in R.
• Compare groups and interpret Cox regression results, including hazard ratios and uncertainty.
• Check key model assumptions and recognise limitations when reporting findings.
• Adapt a worked R script to begin analysing their own time-to-event data.