Sources of potential bias when combining routine data linkage and a national survey of secondary school-aged children: a record linkage study.

On average, data linkage schools reported a very marginally higher percentage of withdrawals (0.18%; 95% CIs: 0.14, 0.23) than non-linkage schools (0.14%; 95% CIs: 0.11, 0.16), and a marginally lower percentage compared to control schools (0.25%; 95% CIs: 0.19, 0.32); although these differences did not reach statistical significance, and opt-out remained sufficiently trivial as to be unlikely to introduce bias.

also in 111,027 other papers