Before scientists can begin working on a question that interests them, they need to formulate a research hypothesis. This is an important step in the scientific method because it determines the direction of the study. Scientists need to scrutinize previous work in the area and select an experimental design to use that helps them find data that either supports or rejects their hypothesis. There are four types of hypothesis scientists can use in their experimental designs: null, directional, nondirectional and causal hypotheses. The hypothesis chosen by researchers will influence the design of the study or experiment they go on to perform, and will direct the way that the study's results are communicated in academic papers. They can also drive further research by establishing research baselines and narrowing down the list of variables that may be affecting a relationship.

The Conventional Null Hypothesis

This is the conventional approach to making a prediction. It involves a statement that says there is no relationship between two groups that the researcher compares on a certain variable. The hypothesis also may state that there is no significant difference when different groups are compared with respect to a particular variable. For example, "There is no difference in the academic performance of high school students who participate in extracurricular activities and those who do not participate in such activities” is a null hypothesis. In many cases, the purpose of a null hypothesis is to allow the experimental results to contradict the hypothesis and prove the point that there is or is not a definite relationship between two variables. This allows for further study into a subject, and allows researchers to identify which variables are or aren't affecting a relationship – as the answer to a null hypothesis can eliminate a variable from consideration in future research.

The Exploratory Nondirectional Hypothesis

Certain hypothesis statements convey a relationship between the variables that the researcher compares, but do not specify the exact nature of this relationship. This form of hypothesis is used in studies where there is no sufficient past research on which to base a prediction. Continuing with the same example, a nondirectional hypothesis would read, “The academic performance of high school students is related to their participation in extracurricular activities.” A nondirectional hypothesis is useful when researchers are starting to examine the relationship between two variables: it leaves ample room for researchers to determine whether there is a positive, negative, or lack of correlation between two studied variables and use that result to launch further research projects.

The Trend-Based Directional Hypothesis

This type of hypothesis suggests the outcome the investigator expects at the end of the study. Scientific journal articles generally use this form of hypothesis. The investigator bases this hypothesis on the trends apparent from previous research on this topic. Considering the previous example, a researcher may state the hypothesis as, “High school students who participate in extracurricular activities have a lower GPA than those who do not participate in such activities.” Such hypotheses provide a definite direction to the prediction. These predictions provide easy metrics to determine whether the hypothesis is valid or invalid – but also run the risk of being adversely affected by confirmation bias when a researcher is determined to prove their prediction right regardless of actual results.

Influence Degrees and the Causal Hypothesis

Some studies involve a measurement of the degree of influence of one variable on another. In such cases, the researcher states the hypothesis in terms of the effect of variations in a particular factor on another factor. This causal hypothesis is said to be bivariate because it specifies two aspects – the cause and the effect. For the example mentioned, the causal hypothesis will state, “High school students who participate in extracurricular activities spend less time studying which leads to a lower GPA.” When verifying such hypotheses, the researcher needs to use statistical techniques to demonstrate the presence of a relationship between the cause and effect. Such hypotheses also need the researcher to rule out the possibility that the effect is a result of a cause other than what the study has examined.

Related Articles

About the Author

Blake Flournoy is a writer, reporter, and researcher based out of Baltimore, MD. Working independently and alongside professors at Goucher College, they have produced and taught a number of educational programs and workshops for high school and college students in the Baltimore area, finding new ways to connect students to biology, psychology, and statistics. They have never seen Seinfeld and are deathly scared of wasps.