Page 124 -

110 CHAPTER 5 Surveys

US CENSUS—COUNTING EVERYONE
In the United States, a national census is taken every 10 years. Every person or
family in the United States is supposed to fill out a paper survey. Responses to
the Census Bureau are required by law, as the census count is used to distribute
budgets and seats in congress and to make many governmental decisions. When
a response is not received, individuals working for the Census Bureau visit
residences to try and collect data from those who did not respond to the paper
survey.
The Census Bureau tested a web-based form during the 2000 census. People
who received the short form (five out of every six Americans) had the option
of filling out the census form on the web. Each paper short form had an ID
number. To ensure appropriate counting, the respondent had to enter the ID
number on the web before filling out the actual survey.
Due to security and privacy concerns, the Census Bureau decided not to
have a web-based form in 2010. However, the Census Bureau used a web
survey for “reinterviewing” those who had already submitted their primary
census form. See https://www.census.gov/ for more information.

Instead of running a census, a structured method called “random sampling” (or
“probability sampling”) is often used. In a probability sample, it is known exactly how
likely it is for a participant to be selected for the sample, which is an equal, greater
than zero chance, and everyone selected in the sample receives the same invitation to
participate (Sue and Ritter, 2007; Müller et al., 2014).For instance, imagine that there
are 10,000 members of a population of interest (the sampling frame). Perhaps 500 of
these individuals are selected, at random, for requested inclusion in a survey study.
All of these selected individuals must meet inclusion criteria (characteristics that
they must have, such as being a nonsmoker or male) and not meet exclusion criteria
(such as not being a native English speaker) (Sue and Ritter, 2007). See the sidebar
on Random Sampling for an example of random sampling of a population of users.

A STUDY WITH RANDOM SAMPLING

When users are required to log into networked resources (such as an e-mail
system, intranet, or social networking site), random sampling methods can
be used, since a detailed list of who is considered to be within the population
of interest does exist. For instance, a research study focused on Beehive, an
enterprise social networking system from IBM. At the time of the study, it was
estimated that there were at least 38,000 registered users of the site. A total
of 500 users were randomly selected and invited to participate in the research
study, based on having logged into Beehive during the last week and having
enough data in their account so that friend recommendations could be made (the