Univariate Statistics

Using Univariate Statistics in Communication Research

Communication researchers use univariate statistics as a tool to answer questions or test hypotheses about the world around them. For example, a researcher may read about the need for more organ donors and be interested in finding ways to increase the number of people who sign the organ donation registry. The researcher would look over previous research on the topic and how health messages inspire people to action. Suppose the researcher finds during his or her research that public service announcements (PSAs) have been shown to increase the number of people who take action on a specific health topic. The researcher may then develop a hypothesis that those people who watch a PSA advocating organ donation are more likely to sign the registry versus those who do not watch the PSA.

The researcher can then design an experiment to test the hypothesis. In this experiment, the independent variable would be the PSA (with two levels of shown or not shown). The independent variable represents the variable that is manipulated in the experiment to determine if it has any effect on the dependent variable. To conduct the experiment, the researcher could randomly select 100 people to participate in the experiment and randomly assign them to two groups. The researcher would show the PSA to half of the sample (Group 1) and not show the PSA to the other half of the sample (Group 2). At the end of the study, the researcher would ask participants in both groups to sign the donor registry. The dependent variable in this example would be signing the registry and it is used to determine if the independent variable (PSA) has any effect on the dependent variable (getting people to sign the registry).

The information collected during the experiment on who did or did not sign the registry and which group they were in (watch the PSA/not watch the PSA) [Page 1810]would represent the researcher’s experimental data. But how does the researcher use the data to find support for the hypothesis? This is where statistics become useful. After the researcher reviews the data and finds that there is a single dependent variable, he or she would then use univariate statistics to test the hypothesis. For this example, the researcher would run an independent samples t-test. If the researcher finds that there are significant differences, meaning the differences are greater than those expected by chance, in who signed the registry among the two groups of those who watched the PSA and those who did not, the researcher has support for the hypothesis and can conclude that using a PSA to attract more donors is a good idea for future health campaigns.

...