PSI's response to forum message by Chris and behavioral determinants

Navendu Shekhar - Friday 26 January 2007

I am sorry to be late in responding to specific issue on PSI programs.

On Chris's point about whether PSI as a matter of practice promotes both water
treatment and safe storage in all 24 countries(where PSI has a safe water
intervention program), the answer is no. PSI in most country programs promotes
water treatment at Point of Use as noted in Orlando's background paper (page
6, 1st paragraph).

PSI did introduce safer options of water storage vessels/ tanks in a few
countries (notably Nigeria, India, Uganda) but was unable to scale them
significantly. We however incorporate safe storage messaging into most of
our programs (brochures, demonstrations, etc) but its inclusion as a
structured program is not uniform. To give a rough figure, it would probably
be accurate to say that we have safe storage communication in about half of
the countries where we have water programs.

Let me also take this opportunity and respond to some of Orlando's
expectations from the e-conference (mentioned on page 17 of his background
paper)

*Issues in Orlando's paper (Page 17) *

· What are the strengths and limitations of the indicators listed
above for measuring behavioural determinants? (Please address each indicator
separately).

· Are there better ways to measure behavioural determinants that can
reduce the vagueness of those proposed and make them more explicit and less
generic?

*Response*

We at PSI, have some experience dealing with behavioural determinants.
Please see the following link.
http://www.psi.org/research/documents/behaviorchange.pdf

The concept paper on the link is largely for our internal use and elaborates
the behaviour change framework that we at PSI use. As the document
elaborates, our approach is very much in line with the model where program
activities lead to change in behavioural determinants which in turn impact
behaviour.

Our experience suggests that: 1) There may not be common determinants that
would predict behaviour across every country. In our experience, it is best
to identify significant determinants applicable to each country's/ target
group's context. We do this by using a multivariate logistic regression to
identify behavioural determinants significantly associated with use/ desired
behaviour.

(An example of such a study is
http://www.psi.org/research/smr/517%20Uganda%20PLHA%20SWS%20SMRS.pdf)

So, a limitation of the proposed indicators and
making them standard for everyone would be that we may force program
managers to invest their resources in a determinant that may not be driving
behaviour at all (in their country/ target group's context).

2) On issue of better ways to capture behavioural
determinants, our experience suggests that most determinants are slightly
complex concept and may require a few scale items to capture it fully. To
take an example, perceived severity can be considered a behavioural
determinant but needs to be captured through various angles such as
perceived severity towards self, towards children, perceived severity of the
different waterborne diseases and so on. That would mean not one but
multiple questions to be added. I am not sure if that's a very good idea.
3) That brings the point of why are we wanting to
measure determinants. In our experience, we have found that it is important
to capture determinants only when we want to design/change our intervention
activities or when we want to know what our communication should
say/emphasize. In such circumstances, it is important to target the right
behavioural determinant. Is that a need that this conference wants to
address or should address? May be for monitoring, it is a better option to
restrict ourselves to the behavioural indicators.

--
Navendu Shekhar
Regional Researcher, Southern Africa
Society for Family Health/ Population Services International
Metro Park, 2nd Floor,
8 Hillside Road, Johannesburg, 2193
South Africa

Phone: +27 11 484 5320 ext 229
Fax: +27 11 484 5802
Cell: +27 84 357 7419

New message
Reply to this message