SECTION CONTENTS

In the previous section, we identified broad categories of descriptive (non-experimental) and causal (experimental) strategies for conducting research. The aim of Section 3 is to discuss these issues in more detail by examining various research designs and outlining their applications for conducting health sciences research. As in other areas of creative human activity, such as architecture or the fashion industry, designs refer to explicit plans for completing an object or an activity.

The conceptual basis for experiments is the need for control, as outlined in Chapter 4. In experimental research (see Ch. 5) we manipulate one or more variables (independent variables), while controlling the extraneous variables, by using appropriate controls such as control groups. If the experiment is well designed and properly conducted we are in a position to demonstrate the causal effects of the independent variables on the outcome or dependent variables. The random assignment of subjects to experimental groups is a common way to achieve control. However, health and illness outcomes often have multiple and interacting causes requiring multivariate experimental designs.

Another issue we will address is reactivity of human beings to social situations, such as being involved in research. Appropriate experiments attempt to control for the biases of both the participants and the experimenters involved in a research project.

There are situations where for practical or ethical reasons we cannot randomly assign subjects to control or treatment groups. Here we use quasi-experimental methods which involve comparing pre-existing groups undergoing different treatments (see Ch. 6). There are other research designs discussed in Chapter 6, including naturalistic comparisons, correlational designs and surveys. Naturalistic comparisons and correlational designs enable us to describe and predict relationships among variables, but should be used only with extreme caution in attempting to understand causal effects. Surveys that include the use of questionnaires and interviews to study a person’s knowledge, attitudes and beliefs concerning aspects of health and illness are essentially a tool for descriptive research. The accuracy of the information collected within the framework of non-experimental designs depends to a large extent on the use of appropriate sampling strategies for selecting the partipants.

An interesting and commonly used design in clinical settings is the n = 1 or ‘single case’ design. The advantage is that using n = 1 designs we may be able to demonstrate causal effects using only one or two subjects, without the need for separate control groups. The major limitation of n = 1 (and other types of clinical case studies) is that the findings may not be generalizable to other cases or situations.

Chapter 8 describes some of the elementary characteristics of qualitative field research designs. This chapter aims to compare and contrast quantitative and qualitative research, and describes the importance of qualitative research for understanding the personal experiences of our clients and/or patients. Qualitative data collection strategies (such as in-depth interviewing and participant observation) are outlined in Section 4 and quantitative data analysis in Section 6.