Page 665 - Encyclopedia of Business and Finance

eobf_R 7/5/06 3:20 PM Page 642

Research in Business

repeatedly about possible developments of a particular 2. determining the sample size that is large enough to
product/service, technology, workforce, or business strat- represent the population yet small enough to be eco-
egy in varied scenarios. After the first round of inquiries, nomical
the experts review the responses of their peer panelists, 3. selecting a sampling method that ensures that each
revise their own if needed, and then start the second member in the population has an equal chance to be
round of inquiries. In most cases, through three to four selected
rounds, the experts’ assessments converge toward a set of
4. determining whether a questionnaire or structured
conclusions that form the basis for solutions to the prob- interview is more appropriate for gathering informa-
lem(s). This method is effective for predicting uncertain tion
future business environments and developments, espe-
cially when quantitative forecasting based on past trends is 5. developing effective questions by using clear, simple
not reliable. language, nonleading or non-multifaceted questions,
mutually exclusive choices, and group categories for
sensitive information such as age, salary, and morals
Experiments. Experiments are designed to test the cause-
and-effect relationships that researchers suspect exist in a For a questionnaire survey, researchers first decide
business environment. There are two categories of experi- whether mail, e-mail, or online is the appropriate medium
ments: laboratory and field. A laboratory experiment takes for ensuring a high response rate. Second, researchers need
place in an artificial setting created by researchers—such to send a persuasive cover letter with the questionnaire to
as testing a new drug on mice or testing automobile safety persuade readers to complete and return the questionnaire.
in collisions using dummies. By contrast, a field experi- If the first mailing (or e-mailing) generates a low response
ment takes place in a natural setting, such as testing the rate, a follow-up mailing is necessary to remind nonre-
effect of implementing a total-quality-management sys- spondents to complete and return the questionnaire.
tem on product quality and employee productivity in a For an interview survey, researchers first decide
company. whether a telephone or face-to-face interview is more
Whether in a lab or in the field, researchers exercise appropriate. Interviewers must be trained before conduct-
high control when they are able to manipulate independ- ing interviews. The interviewer begins by making the
ent variables, assign subjects randomly to control and interviewee feel comfortable and important by giving a
experiment groups, and control extraneous variables. For friendly greeting, explaining the purpose of the interview
example, a company wants to know whether newspaper and the importance of the interviewee’s participation, and
coupons for its product affect sales. The company ran- asking if tape recording is permitted. During the inter-
domly selects ten communities from the fifty communi- view, the interviewer should always use an interview guide
of predeveloped questions and follow-ups for each inter-
ties having daily newspapers within its sales region and
viewee. At the end, the interviewer should express appre-
then randomly assigns them to control and experiment
ciation to each interviewee for participating.
groups, with each having five communities. On Tuesday,
the control group receives a newspaper advertisement for
the company’s product without a coupon, while the Content Analysis. Content analysis is a method for mak-
experiment group gets the same advertisement with a ing inferences by systematically and objectively identify-
ing characteristics of messages embedded in the texts and
coupon for a 15 percent discount. Sales figures from the
the causal relationships of message contests and outcomes.
two groups are totaled on the following Saturday (the
This method is widely used in the research of business and
coupon’s expiration date). Through a statistical analysis of
the sales figures, the experiment shows whether or not the managerial communications, negotiations, and Web-
coupon has significantly affected the sales. based e-business. Content analysis employs a systematic
procedure of selecting texts, developing content cate-
Surveys. Survey research is a method of systematically
gories, and coding and analyzing data. Using computers
questioning a sample of respondents representing a spe-
to store and analyze data, researchers are able to conduct
cific population regarding their beliefs, attitudes, and
large-scaled content analysis with massive amounts of data
behaviors. This method is useful for collecting informa- without much difficulty.
tion about a population that is too large for every mem-
ber to be studied. To ensure the generalizability of the
BUSINESS RESEARCH PROCESS
findings, researchers design a survey by taking the follow-
ing steps: To conduct business research effectively, researchers follow
a general research process of five phases and must under-
1. defining the population of the study stand their interrelationships (see Figure 1).

642 ENCYCLOPEDIA OF BUSINESS AND FINANCE, SECOND EDITION