Space, Place and Mental Health by Sarah Curtis

Author:Sarah Curtis [Curtis, Sarah]
Language: eng
Format: epub
Tags: Social Science, Human Geography
ISBN: 9781317051855
Google: F5veCwAAQBAJ
Publisher: Routledge
Published: 2016-04-01T05:08:06+00:00

‘Reverse causation’ and selective migration

As already observed, many of the research studies reviewed here are ‘cross sectional’ in design and it is therefore quite difficult to interpret the associations between area conditions and mental health. It is therefore important to consider processes of reverse causation and selective migration, whereby people already in poor mental health may be more vulnerable to material poverty and tend to congregate in poor areas.

Statistical research involving systematically recorded information for extensive samples usually relies on indicators which are approximations to the phenomena being measured and may also be proxies for other, unmeasured characteristics, which is one reason for extreme caution when interpreting statistical associations in terms of ‘cause and effect’ relationships. Also, in ‘cross sectional’ studies collecting all the information analysed at one point in time, there is no way of knowing which conditions were temporally antecedent and might have given rise to other, subsequent conditions. (Logically it is hard to see how a condition that arises later in time could give rise to differences in earlier conditions, except possibly if respondents are asked to give retrospective reports on conditions in the past and their recall is biased by their experience of their current situation.)

Although much of the literature in social epidemiology uses terminology which refers to ‘predictor’ variables having an ‘effect’ on an ‘outcome’ variable, this should generally be interpreted as a technical description of a statistical model, rather than an expression of actual causal relationships, especially in cross sectional studies. Research aiming to test causal processes giving rise to mental health variation is more rigorous if it is based on a longitudinal design, and can examine change in mental health, but even this method may not be able to measure all aspects of causal processes, so may give a partial picture of the chain of causation. Strictly speaking, discussion of statistical associations in terms of ‘one variable having an effect on another’ is therefore inappropriate in this kind of research.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.