Behavioural Surveillance Surveys - The Wisdom of Whores
Behavioural Surveillance Surveys - The Wisdom of Whores
Behavioural Surveillance Surveys - The Wisdom of Whores
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
What is to be measured ?<br />
Indicators must be chosen before the<br />
survey process begins, so that questionnaires<br />
can be checked to see that they do indeed<br />
yield information that can be used to construct<br />
the indicators selected. Standardized<br />
indicators have been defined for several<br />
sub-populations likely to be included in BSS.<br />
It is recommended that standardized indicators<br />
be used wherever possible. <strong>The</strong> use <strong>of</strong><br />
standard definitions and time reference<br />
periods greatly increases the comparability <strong>of</strong><br />
data across time and in different populations,<br />
and increases the cohesion <strong>of</strong> a national<br />
surveillance and monitoring and evaluation<br />
system for HIV.<br />
Detailed definitions and explanations <strong>of</strong><br />
standard indicators can be found in Chapter 9<br />
<strong>of</strong> these guidelines. <strong>The</strong>re may, however, be<br />
good reasons to include other indicators that<br />
are more situation-specific. This may be linked<br />
to behavioral patterns in a particular culture,<br />
or may be derived from some aspect <strong>of</strong> the<br />
national response. Use <strong>of</strong> a particular brand<br />
<strong>of</strong> socially-marketed condom may be an<br />
example. Survey designers must always bear<br />
in mind, however, that BSS is not able to<br />
evaluate the independent effects <strong>of</strong> individual<br />
prevention projects.<br />
Step 5 : operationalizing<br />
definitions <strong>of</strong> populations <strong>of</strong><br />
interest<br />
Once overall respondent groups have<br />
been settled on, the sampling universe must<br />
be defined. <strong>The</strong> universe is the population to<br />
which results may be extrapolated. An example<br />
<strong>of</strong> a respondent group might be sex workers,<br />
and the sampling universe could be those<br />
selling sex in the capital city <strong>of</strong> the country.<br />
After defining the sampling universe,<br />
the sampling domains must be identified.<br />
A sampling domain is a specific population<br />
segment or subset for which separate survey<br />
estimates are desired. For example, sex<br />
workers may be broken down into brothelbased<br />
and freelance sex workers. Further,<br />
brothel-based sex workers may be broken<br />
down into urban and rural brothel-based sex<br />
workers. In order to ensure that statistically<br />
significant changes can be monitored over<br />
time in particular sub-groups (or domains),<br />
sample sizes are calculated per domain.<br />
This means that it is very important to think<br />
clearly beforehand one word about what<br />
information the program really needs to<br />
improve its prevention efforts. It is no good<br />
calculating a sample size for all sex workers,<br />
and then deciding afterwards that it would be<br />
interesting to know whether changes have<br />
occurred just in sex workers aged under 25.<br />
Unless this is taken into account at the design<br />
stage, it is very likely that the sample size will<br />
be too small to calculate significant changes<br />
in a subset <strong>of</strong> the group, such as younger<br />
women only.<br />
Operational definitions <strong>of</strong> survey domains<br />
After deciding which groups and sub-groups<br />
are <strong>of</strong> interest, survey designers must specify<br />
exactly how they define a member <strong>of</strong> those<br />
groups. <strong>The</strong> importance <strong>of</strong> clear operational<br />
definitions <strong>of</strong> survey domains cannot be<br />
over-emphasized ; it is <strong>of</strong>ten quite difficult to<br />
achieve. In a culture with no overtly “gay”<br />
identity, for example, how should men who<br />
have sex with men be defined ? Where large<br />
numbers <strong>of</strong> women supplement their income<br />
with the occasional sale <strong>of</strong> sex, what constitutes<br />
a sex worker ? Unless domains can be defined<br />
in terms that are operationally useful for<br />
sampling and fieldwork purposes, survey<br />
estimates regarding such group’s behaviors<br />
can be expected to be subject to considerable<br />
error, and the group may have to be dropped<br />
from BSS.<br />
16<br />
C H A PTER 2 B EHAV I OR A L S U R V EI L L A NC E S U R V EY S