change techniques employed. We can see, therefore, that the list of characteristics
provided in Table 9.1 is a useful starting point but detailed intervention planning can
involve many different intervention characteristics.
One application of taxonomies of intervention content is a type of ‘retrospective
or archaeological process evaluation’. This is distinct from process evaluation proper
conducted experimentally as part of a trial, as described above. The idea is to compare
many published outcome evaluations and investigate whether the content of the
reported interventions is associated with reported effectiveness using meta-analysis.
Albarracín et al.(2005) provide a good example. Using 354 intervention descriptions
and 99 control groups, spanning 17 years, these researchers asked which of 10 change
techniques were most effective in promoting condom use among different target
groups. The researchers identified five types of persuasive communication, which they
referred to as ‘passive’ because these techniques could be employed without active
involvement of the recipients (e.g. through a health promotion leaflet). These were
provision of (1) information; (2) arguments designed to change attitudes; (3) arguments
designed to change normative beliefs; (4) arguments designed to persuade recipients
that they could perform successfully prerequisite tasks, that is, to enhance self-efficacy;
and (5) threat or fear-inducing messages. In addition, the researchers considered five
techniques that could be used in interventions involving ‘active’ or face-to-face
interaction with recipients. These included three types of skill training, namely (6)
condom-specific skill training; (7) self-management or self-regulatory skills training;
and (8) interpersonal or social skills training. In addition, the researchers considered
(9) provision of condoms; and (10) HIV counselling and testing. Results showed,
perhaps unsurprisingly, that active interventions involving interaction with recipients
were more effective in promoting condom use. The most effective interventions
provided information, attitudinal arguments, behavioural skills arguments and provided
self-management (or self-regulatory) skills training. In addition, provision of condoms
and HIV counselling and testing enhanced intervention effectiveness.
Overall, these results offer support for targeting the change processes specified by
the theory of planned behaviour, the information, motivation and behavioural skills
model and social cognitive theory. By contrast, they suggest that the change processes
specified by protection motivation theory may be less useful to intervention designers
trying to promote condom use (see Chapter 7). It is worth noting too that given an
average d of 0.38 for active interventions and assuming that, on average, 36 per cent
of people in a target group use condoms at least sometimes (figures reported by
Albarracín et al. (2005) – see Research methods 8.1 on dvalues) then an additional
17 per cent will use condoms at least sometimes following such active interventions.
This is a sizable increase in the number of users following intervention and could impact
on the prevalence of STIs and unwanted pregnancy rates. Interestingly, Webb and
Sheeran (2006) conducted a similar review investigating how effective interventions
that promoted intention formation (or goal setting) were in changing behaviour across
behavioural targets. They found that medium to large changes in intention (d = 0.66)
resulted in small to medium change in behaviour (d = 0.36); an effect size very similar
to that observed for active interventions by Albarracín et al.
Albarracín et al.(2005) also found that some techniques were associated with
effectiveness for some recipients but not others. Normative arguments targeting
subjective and descriptive norms were found to promote behaviour change in audiences
CHANGING BEHAVIOUR 221