Write the question before the percentage
A headline can divide returns by orders to leave and produce a neat number. Before calling it a success rate, ask what the numerator and denominator represent. Did the same people receive an order and complete a return within the measured window, or are two separate totals being compared?
Eurostat explicitly warns against directly relating people returned in a quarter to those ordered to leave in that same quarter. An order issued late in one year may lead to a return the next year. Its quarterly article also separates returns to third countries from total returns, which include readmissions between EU and Schengen-associated countries.
A fictional example makes the mismatch visible
Imagine an office issuing 100 orders in April to June and recording 30 returns during those months. Twenty of those returns relate to earlier orders; ten relate to the new group. Dividing 30 by 100 produces 30%, but only ten people in this invented new group have a recorded return so far. Neither calculation tells us what will happen to everyone else.
These are invented teaching numbers, not estimates for a country. The point is the missing link between events. A genuine cohort analysis would define the starting group, follow-up period and outcome rules before calculation. It would also explain unresolved cases and how repeated records are handled.
Keep an extraction card
For each comparison, save the original table and a short note. This is our suggested reading worksheet, not an official Eurostat reporting format.
- Indicator and table identifier, including whether the destination is a third country or any destination.
- Reference quarter, extraction date and geographical coverage.
- Unit counted and rounding or revision notes.
- Whether records are linked across time, or only aggregate totals are available.
- The exact conclusion the evidence supports, followed by the conclusion it cannot support.
Check whether the series changed
The Eurostat source notes changes in national series and definitions. Treat a break marker as part of the number. Before describing a change as operational improvement or deterioration, read the coverage note for both periods. If you cannot make the periods comparable, report them separately.
A useful editorial sentence is: the source records this many returns during this period, under this definition. A stronger claim about the fate of newly ordered cases needs linked evidence. A claim about why outcomes changed needs a further causal argument.
Use the data desk with its limits visible
Open our Europe desk, choose one return measure and record its period and source. Then compare it with an order measure only after writing down the distinction above. The desk is a dated snapshot; a new publication date on this explanation does not refresh its underlying figures.
Source reviewed on 20 September 2026. This article explains how to read aggregate statistics. It does not determine any individual immigration case.