26
Let’s talk a little more about sample and response bias. Sample bias
occurs most often because of nonresponse (selected respondents or units
are not available or refuse to participate, or some answers and
observations are incomplete). Response bias occurs because questions
are misunderstood or poorly formulated, or because respondents
deliberately equivocate (for example, to protect the project being
evaluated). In observations, the observer may misinterpret or miss what
is happening. Exhibit 9 describes each type of bias and suggests some
simple ways of minimizing them.
Exhibit 9.—Three types of errors and their remedies
Type Cause Remedies
Sampling Error Using a sample, not the entire
population to be studied.
Larger samples—these reduce but do not
eliminate sampling error.
Sample Bias Some of those selected to
participate did not do so or
provided incomplete information.
Repeated attempts to reach nonrespondents.
Prompt and careful editing of completed
instruments to obtain missing data;
comparison of characteristics of non-
respondents with those of respondents to
describe any suspected differences that may
exist.
Response Bias Responses do not reflect “true”
opinions or behaviors because
questions were misunderstood or
respondents chose not to tell the
truth.
Careful pretesting of instruments to revise
misunderstood, leading, or threatening
questions. No remedy exists for deliberate
equivocation in self-administered interviews,
but it can be spotted by careful editing. In
personal interviews, this bias can be reduced
by a skilled interviewer.
Statistically valid generalizations are seldom a goal of qualitative
evaluation; rather, the qualitative investigation is primarily interested in
locating information-rich cases for study in depth. Purposeful sampling is
therefore practiced, and it may take many forms. Instead of studying a
random sample or a stratified sample of a project’s participants, an
evaluation may focus on the lowest achievers admitted to the program, or
those who have never participated in a similar program, or participants
from related particular regions. In selecting classrooms for observation of
the implementation of an innovative practice, the evaluation may use
deviant-case sampling, choosing one classroom where the innovation is
reported as “most successfully” implemented and another where major
problems are reported. Depending on the evaluation questions to be
answered, many other sampling methods, including maximum variation
sampling, critical case sampling, or even typical case sampling, may be
appropriate (Patton, 1990). The appropriate size of the sample may also
differ when the different methodologies are adopted, with precision in
numbers based on statistical considerations playing a much larger role
for the quantitative approach.