QUALITATIVE AND QUANTITATIVE SAMPLINGEXAMPLE BOX 1
Quota Samplesfriends, family) among people in different social
classes in major Chinese cities. They selected house-
holds in four of China’s largest metropolitan areas
(Shanghai, Shenzhen, Tianjin, and Wuhan), identified
a set of neighborhoods in each, and then sampled
100 people per city. They had a list of thirteen occu-
pational titles that represented the full range of the
class system in China and 88 percent of all working
people in the four cities. Their quota was to get an
equal number in each city and a sufficient number of
households in each of the thirteen occupational cate-
gories for careful analysis. Thus, only 4 percent of the
people held the position as manager, but nearly 10
percent of the sample were managers, and 40 percent
of people held an industrial worker occupation, but
close to 10 percent of people in the sample were indus-
trial workers. The study goal was to test hypotheses
about whether a household’s social ties are with others
of similar or different social classes. They asked house-
holds to maintain a written log of social visits (in per-
son or via phone) with other people and recorded the
occupation of visitors. This process lasted a year, and
researchers interviewed people every three months.
The primary interest in the study was to compare pat-
terns of social networks across the various social
classes. For example, did managers socialize only with
other managers or with people from a wide range of
classes? Did industrial workers socialize with industrial
workers as well as people in various lower occupations
but not in higher occupations? Because the study goal
was to compare social network patterns across the var-
ious classes, not to have a representative sample that
described the Chinese population, it was a highly effec-
tive use of quota sampling.Two studies illustrate different uses of quota sampling
in quantitative research. In a study, McMahon,
McAlaney, and Edgar (2007) wanted to examine
public views of binge drinking in the United King-
dom. They noted that most past research was on
young adults and campaigns to curb binge drinking
had been ineffective. The authors wanted to learn
about public perceptions of binge drinking among
the entire adult population. They developed a survey
that asked how people defined binge drinking, the
extent to which they saw it as a concern, and reasons
for and solutions to it. They combined quota
sampling with another sampling method to interview
586 people in one city (Inverclyde, Scotland). For
quota sampling, interviewers approached potential
participants in the streets surrounding a shopping
center and invited them to take part in the survey.
The quota was based on getting a balance of gender
and six age categories. The other method was to go
door-to-door in several low-income neighborhoods.
The authors learned useful information about views
on binge drinking across age groups in both genders
in one city. They found wide variation in definitions
of binge drinking and support for a “false consensus
effect” in which a small number of the heaviest
drinkers see their behavior as normal and socially
accepted. Nonetheless, the sample is not represen-
tative, so findings on the extent of binge drinking in
the public and views about it may not reflect the
behaviors or views within the city’s overall population.
A second study in China by Bian, Breiger, Davis,
and Galaskiewicz (2005) employed a targeted use of
quota sampling. Their interest was in the difference
between the social networks and social ties (e.g.,
its geographical and temporal boundaries as well
as any other relevant boundaries.
Most probability studies with large samples of
the entire U.S. population have several boundaries.
They include adults over 18 who are residents of
the forty-eight continental states and exclude the
institutionalized population (i.e., people in hospi-
tals, assisted living and nursing homes, military
housing, prisons and jails, homeless and battered
women’s shelters, college dormitories). Ignoring
people in Alaska, Hawaii, and Puerto Rico and
excluding the institutionalized population can throw
off statistics—for example, the unemployment rate
would be higher if the millions of people in prison
were included in calculations (see Western and
Pettit, 2005). Many studies include only English