You are right ... guidelines tend to be for rural villages rather than urban areas. Also, they seem to assume that a rural village is a roughly circular group of dwellings. Most villages I have visited are more complicated than that. I have some sampling guidelines that you may find useful for urban settings (the issues are tricky to deal with in a text-only forum - on this topic pictures and maps really do paint thousands of words). I suggest that you contact the forum administrator for my e-mail address and e-mail me directly about this and I will send you what I have. If there is a demand ... I will prepare an article for Field Exchange on this issue. Let me know if there is an interest in this.

Dear anonymous, Thankyou for that contribution. I do concurr with you as regards the methodology. I happened to have participated in an nutritional anthropometric assess of informal settlements (urban slums) in Kampala, Uganda. Our success largely depended on obtaining as much information as possible regarding areas of study and corroborating this information as much as possible using secondary national data!! Two classic examples; i) We could not ascertain the actual populations within the areas selected but with the help of the Local area concillors, we were able to get some estimates to use cluster formation and sampling! ii) Even with our rough estimates, we noted that clusters were very large hence the need for secondary clustering!! iii) The structure of household arrangement was quite complicated but never the less, we were able to use a random selection method. In my view, the exisiting guidelines in ENA are just sufficient to conduct just about any kind of assessment. We for example believe we came up with a new sampling technique due to the complex structural arrangement of household patterns! This method is not reported in the ENA Methodology and only needed to be acknowledged. Besides sampling, I have not felt ineficiency on other sections of the methodology. Samuel

Correction of word! Please ignore the word "exiting" and replace by " existing" Thanks

You are right ... survey guidelines tend to be for rural villages rather than urban areas. They also seem to assume that a rural village is a roughly circular group of dwellings. Most villages I have visited are considerably more complicated than that! Early work on the EPI survey method on which the SMART and "30-by-30" methods are based showed that the assumption of a round village and the use of proximity sampling (i.e. a cluster of households around a random starting point) does produce a sample that is, for some purposes, good enough - provided that the number of clusters used is above about 25 (this is usually 30 as in the "30-by-30" method for nutritional anthropometry of the "30-by-7" EPI design) are used. I have had a go at producing a brief sampling guideline that you might find useful for urban settings. You can get this from [url]http://www.brixtonhealth.com/urbanSampling.pdf[/url]. If there is a demand for it I could prepare a longer article for Field Exchange on this issue. Let me know if there is an interest in this.

I would be very interested in any further information on sampling in Urban settings - the organisation I work for is starting to do more surveys in towns and cities, so any further information would be most welcome.

Sometime soon (it may already have happened!) over 50% of the human population will live in towns and cities. We have to face this and stop considering "urban" and "peri-urban" populations as something a little special. This means, I think, that we need survey methods better adapted to urban and peri-urban settings. Perhaps ENN / FE can take a lead on this issue. Do SMART have any plans to work on this?