Dear Colleagues, I am reviewing a number of primary studies on the impact
of new roads on residents’ disturbance. I’ve been putting together a list of
general criteria that can help me determine their methodological soundness.
Most of the studies are before and after interviews in which local residents
are asked to assess the amount of disturbance they experience due to road
traffic in their area. The questions tend to focus on issues such as noise, air
pollution, vibration and severance. The list of methodological criteria that I
have so far is printed below. It is doubtful that many or even any studies will
fulfil all the criteria but it would be useful to know if anyone who has looked
into/worked on studies of this kind thinks I have missed out obvious points,
or if anyone foresees any problems with the criteria I have put down.
Assessing the methodological soundness of disturbance studies by
looking for the following general characteristics:
Sample:random sample or entire population.
Responserate: >80% good response rate, >60% acceptable response
rate.
Percentageof Follow-up: high percentage (>80%?) of original sample
followed-up in subsequent interviews (or authors should be able to
demonstrate that respondents lost to attrition do not seriously confound
results).
Control/adjustmentfor confounders: prospective study with either
comparable control area or appropriate adjustment for relevant
confounding factors (such as age, gender, time spent at home, proximity
to road, sound barriers, double-glazing, alternative sources of
noise/pollution, smoking).
Ascertainingexposure: reliable measurements of physical environment
(eg. traffic rate, noise level, pollution level).
Adaptation:control for/exploration of effects of adaptation over time.
Secondaryroads: impact of new road on residents of secondary roads
considered.
Sufficientdata presented: enough data to verify author’s conclusions.
Multi-sitecomparison of new roads: on assumption that a rigorous
study of several new roads is preferable to a rigorous study of a single
new road.
Questions
shouldI set a percentage figure for ‘good’ and ‘acceptable’ follow-up
rates – clearly such figures are arbitrary but some sort of distinction
needs to be made between ‘high’ and ‘low’ rates of follow-up?
is my list of confounding factors appropriate – any omissions? Which
confounding factors are generally seen to be the most important?
shouldI specify a ‘good’ or ‘acceptable’ length of follow-up?
I would be very grateful to anyone who has any thoughts on the above. Any
assistance will of course be acknowledged.
With thanks and best wishes,
Matt Egan
MRC - Social and Public Health Sciences Unit,
University of Glasgow,
4 Lilybank Gardens, Glasgow G12 8R2.
0141 3577 530
http://www.msoc-mrc.gla.ac.uk/Evidence/Evidence.html
|