Commonwealth University Belize :: Developing and Testing Hypothesis

A hypothesis is a statement of concrete relationship between variables. It is a statement describing empirically testable relationship between concepts. In most academic research projects, it is intended that the scholar should exhibit competence in the handling a critical building block or the foundation of the research project. The research proposal would normally include elements of the direction of the hypotheses. As the researcher embarks on the development of research instruments such as the questionnaire, it is the hypothesis more than anything else that influences which questions are to be asked and how they are asked. This is because the essence of data collection is to generate information to be used to test the research hypothesis.

“The researcher should realise that questionnaires are not constructed aimlessly. They must take their bearing from somewhere.”

The author counsels that to construct a meaningful and relevant questionnaire, the researcher must bear in mind that every single question in the questionnaire is related to the research questions and or the research hypotheses.

It provides direction for a study – no beating about the bush.
It eliminates trial and error research.
It helps to rule out intervening and confounding variables. Since hypotheses focus research to precise testable statements, other variables, whether relevant or not, are excluded.
It allows for quantification of variables (Wimmer and Dominick, 2000:257)

Whether stated as declarative statements or as alternative courses of action, hypothesis should have the following qualities:

It should be compatible with current knowledge.
It should be logically consistent.
It should be succinct.
It must be clearly defined or stated, such that it can be tested.
It must be capable of providing an answer to the problem generated by the enquiry.
It must be verifiable
It must be stated in a simple manner. It should obey the law of parsimony, that is, it should be as simple as possible.
It must, in addition to explaining the present problem, open the door to further knowledge.

How to State a Hypothesis

I have seen how many young or inexperienced researchers make a mess of this in their research proposals. Some write it out in form of an essay, which they present as a hypothesis. This is totally wrong. Others combine so many variables into one hypothesis.

A good hypothesis should be simple and precise. It should state the relationship that is presumed in the hypothesis between the two variables. The hypothesis should also be based on a quantifiable statement, knowing fully well that it will eventually have to be empirically analysed.

Kress (1979:49) has provided two simple alternative ways of stating or formulating a hypothesis. He says that hypothesis can be made as:

• Humorous toothpaste ads on television result in a significantly greater sales response than do non-humorous toothpaste ads.

• Viewers retain the message in humorous toothpaste ads for a significantly longer period than do for non-humorous toothpaste ads.

Hypotheses as Alternative Courses of Action

• More gain will be made from the excess capacity by leasing time to smaller banks in the surrounding communities.

More gain will be made from the excess capacity by leasing time to some of the smaller business forms in the communities.

2. There will be more gain if the excess capacity is used by leasing time to the schools in the community

Null (Alternative) Hypothesis

The null hypothesis (also known as the hypothesis of no difference) asserts that the statistical differences or relationships being analyzed are due to a chance or random error. The null hypothesis is the logical alternative to the research hypothesis (Ho). See Wimmer and Dominick (2000:258).

It is common practice in research to state the hypothesis in two opposing terms. When this is to be done, the researcher should state the null hypothesis, (that is to say, the hypothesis to be tested) in such a way that its rejection leads to the acceptance of the desired statement, i.e., the alternative hypothesis.

H? There is no difference in the results produced from two sales training methods (Method “A” and Method “B”)

H? Sales Training Method “A” produces results that are significantly different from the results from Method “B”

Due to errors, which are associated with probability sampling techniques, hypothesis testing is necessary. The researcher wants to see whether the difference between the samples or the sample and the population properties are too large, i.e., whether the differences are significant to consider them as real differences other than sampling errors. The properties we are talking about include the Mean, Median, Mode, and Variance.

The typically most frequently used statistically methods or tools employed by researchers to test hypothesis are:

In many situations, researchers develop studies that are based on existing theory and are, thus, able to make predictions about the outcome of the work (Wimmer & Dominick, 2000: 256). The real questions that are addressed by hypothesis according to Turkey (1986) are:

Hypothesis testing is concerned with the degree of significance. As Chisnall (1973:59) suggests, it enables probability statements, based on samples, about the characteristics of a population to be tested statistically. It involves the use of test of significance, which are important in the theory of decisions for example in the determining whether the differences noted between two samples are the result of a chance variation or whether they are actually significant.

As Stacks and Hocking (1999:31) rhetorically ask, “How do we know if a theory predicts accurately?” According to the authors, we test it empirically. This is the role of the test of significance, otherwise called hypothesis testing.