|
The Comparative
Immigrant Entrepreneurs Project (CIEP) was designed to examine the
prevalence of transnational entrepreneurship in immigrant
communities and to provide basic information about its empirical
contours and correlates. The research design targeted newer Latin
American immigrant nationalities in their primary areas of
residential concentration in the U.S. and includes questions
concerning the characteristics and regularity of the ties they
maintain with their countries of origin, Specifically, the coverage
includes Colombians in Queens, NY; Dominicans in Washington Heights
(Upper Manhattan) and Providence, RI; and Salvadorans in the
Washington DC metro area (the Adams Morgan section of DC and
Alexandria, Virginia) and Los Angeles (Pico Union).
The sampling
strategy began with establishing contacts randomly in these
communities with key informers to identify the local co-ethnic
entrepreneurs. These points of entry were followed through with a
snowball sample to identify all the ethnic entrepreneurs in each of
these communities. This ensured sufficient coverage of the
entrepreneurs whose business operations depend on maintaining
regular ties to their countries of origin and those who operate
primarily or entirely in the host society. However, these
communities are not populated exclusively by entrepreneurs.
Therefore a two-stage clustered residential sample was implemented.
The first stage identified city blocks as primary sampling units in
these neighborhoods and a simple random sample without replacement
was drawn. Within the primary sampling units the second stage
consisted of a systematic sampling of housing units using either a
one-in-four or a one-in-five sampling fraction. A screener question
for place of birth was asked of the head of household at each
selected residence to determine eligibility. This sampling design
yields equal probability of selection within sampling sites.
Sampling probabilities across sites, however, were not equal due to
differences in the sizes of the immigrant communities within each
site. Therefore, population weights were developed on the basis of
the 1990 census to correct for 1) unequal sampling probabilities
across sites; and 2) the higher ratio of entrepreneurs to
non-entrepreneurs in the study due to the use of the snowball
sampling techniques. Consistent criteria based on screening
questions were used to identify entrepreneurs (respondents who
indicated they were proprietors of firms or otherwise self-employed
in response to questions about occupation and place of employment)
and also to distinguish transnational entrepreneurs from
entrepreneurs who operate domestically ('The success of my firm
depends on regular contact with foreign countries' and 'The success
of my form depends on regular contact with [Colombia/Dominican
Republic/El Salvador, according to respondent's country of
origin]). The unweighted sample numbers 1,202 for all sites while
the weights project the unweighted sample to the population sizes of
each immigrant group within its respective U.S. community.
|
