Crawford, Laura Jessie2019-05-162019-05-162017http://hdl.handle.net/10179/14634With administrative data increasingly being recorded electronically, data linkage has become a popular method of research. It involves the linkage of two data sets – survey and administrative data, in the current study – to create a wider and more varied data set, with which a greater number of research questions can be examined. Seeking consent from participants to link their data is an ethical and legal requirement. However, consent seeking may create systematic bias as likelihood of consenting may be associated with a variety of health and socioeconomic variables. Variables associated with consent were examined for linkage between the Health, Work and Retirement longitudinal study and Ministry of Health data sets in New Zealand. Unlike previous studies of this type, participants were older adults. Binary logistic regression revealed that Māori ethnicity (OR 0.68), Diabetes (OR 0.66), and participating for more survey waves (OR 1.88) were significantly associated with consent. The model explained 7 to 10% of the variance in consent, suggesting that older adults are not greatly influenced by these variables. Implications for research and policy are discussed.enThe AuthorMedical record linkage -- New Zealand -- Data processingMedical records -- Access control -- New ZealandOlder people -- New Zealand -- Longitudinal studiesSocial sciences -- Data processingLogistic regression analysisThesisQ112933526https://www.wikidata.org/wiki/Q112933526