Lack of reliable data about street vendors, who are difficult to survey, ha
s hampered efforts to improve the safety of street-vended food. A two-phase
method for sampling vendors, surveying first in areas of concentrated vend
ing activity identified by local authorities and second in randomly selecte
d areas, was developed and implemented in two Guatemalan cities where stree
t-vended food had been implicated in cholera transmission. In a 4-day surve
y in Escuintla, 59 vendors (42 from phase 1, 17 from phase 2) were intervie
wed. They demonstrated good knowledge of food safety and cholera but unsafe
practices, implying that more effective, practical training was needed. In
a B-day survey in Guatemala City, 78 vendors (77 from phase 1, 1 from phas
e 2) were interviewed. Sixty-eight (87%) vendors stored water, usually in w
ide-mouthed vessels prone to contamination; this led to a field test of a n
ew system for safe water storage. Useful information for public health plan
ning and intervention can be gathered rapidly with this new method for surv
eying street vendors.