Page 44 -

2.2 Research hypotheses 29

to navigate the site more effectively. For this research case, the null and alternative
2
hypotheses can be stated in classical statistical terms as follows:
• H 0 : There is no difference between the pull-down menu and the pop-up menu
in the time spent locating pages.
• H 1 : There is a difference between the pull-down menu and the pop-up menu in
the time spent locating pages.
From this example, we can see that the null hypothesis usually assumes that there is
no difference between two or more conditions. The alternative hypothesis and the null
hypothesis should be mutually exclusive. That is, if the null hypothesis is true, the al-
ternative hypothesis must be false, and vice versa. The goal of the experiment is to test
the null hypothesis against the alternative hypothesis and decide which one should be
accepted and which one should be rejected. The results of any significance test tell us
whether it is reasonable to reject the null hypothesis and the likelihood of being wrong
if rejecting the null hypothesis. We explain this topic in more detail in Section 2.5.
Many experiments examine multiple pairs of null and alternative hypotheses. For
example, in the research case above, the researchers may study the following addi-
tional hypotheses:

• H 0 : There is no difference in user satisfaction rating between the pull-down
menu and the pop-up menu.
• H 1 : There is a difference in user satisfaction rating between the pull-down menu
and the pop-up menu.
There is no limit on the number of hypotheses that can be investigated in one ex-
periment. However, it is generally recommended that researchers should not attempt
to study too many hypotheses in a single experiment. Normally, the more hypoth-
eses to be tested, the more factors that need to be controlled and the more variables
that need to be measured. This results in very complicated experiments, subject to a
higher risk of design flaws.
In order to conduct a successful experiment, it is crucial to start with one or more good
hypotheses (Durbin, 2004). A good hypothesis normally satisfies the following criteria:
• is presented in precise, lucid language;
• is focused on a problem that is testable in one experiment;
• clearly states the control groups or conditions of the experiment.
In the early stages of a research project, researchers usually find themselves con-
fronted with a broad and vague task. There are no well-defined research questions.
There are no focused, testable research hypotheses. The common way to initiate a
research project is to conduct exploratory descriptive investigations such as obser-
vations, interviews, or focus groups. Well-conducted descriptive investigations help
researchers identify key research issues and come up with appropriate control groups
to be manipulated as well as dependent variables to be measured.

2
Traditionally, H 0 is used to represent the null hypothesis and H 1 to represent the alternative hypothesis.