Katy Stokes Demographic Methods Centre, ONS NTTS Conference, March - PowerPoint PPT Presentation

Measuring uncertainty in ONS population estimates: capturing variability in statistics from combinations of census, administrative and survey sources Katy Stokes Demographic Methods Centre, ONS NTTS Conference, March 2015

Uncertainty measures Purpose: to develop a statistical measure of uncertainty associated with local authority mid- year population estimates. Approach: • mirror the mid-year estimates in their basic structure – Cohort Component Method • simulations-based to measure variability around the mid-year population estimates

Cohort component method for population estimates Uncertainty estimates = Assume no variance Births – Deaths Internal + International + Population Population = + migration migration + Minor for year x for year x+1 - - adjustments Bootstrapping to create 1,000 simulations to derive 95% CIs for MYEs

Bootstrapping International Immigration Mid Year Estimates = International Workers MWS 348 Local Passenger Authority Students HESA/BIS/WAG Survey (IPS) estimates of Others PRDS National international Estimate immigrants UK Returners Census Uncertainty estimates = 1,000 1,000 simulated 1,000 Worker counts 1,000 simulated simulations admin counts international 1,000 Student counts from IPS for each immigrant 1,000 ‘Other’ counts migrant type in counts for each each LA LA 1,000 UK Returner counts Apply admin-based proportions to IPS Sum these to produce 1,000 LA estimate for each migrant type to derive counts totals. 26 th and 975 th ranked values of each migrant type in each LA provide uncertainty interval

LA level estimates – international immigration

Summary of uncertainty measures (as % of MYE)

Key points Methods and Techniques • Bootstrapping • Parametric • Non parametric • Imputation • Regression • Poisson regression Software used • SAS and Stata

Questions? for further information: http://www.ons.gov.uk/ons/guide- method/method-quality/imps/latest- news/uncertainty-in-la-mypes/index.html or contact: methodology@ons.gsi.gov.uk