Descriptive Designs

Descriptive designs result in a description of the data, whether in words, pictures, charts, or tables, and whether the data analysis shows statistical relationships or is merely descriptive.  If the topic is for Level I or II, the design will be descriptive. “What” questions invariably lead to descriptive designs.

No matter what method is chosen to collect the data, all descriptive designs have one thing in common: they must provide descriptions of the variables in order to answer the question.  The type of description that results from the design depends on how much information the researcher has about the topic prior to data collection.  Look at the design in the same way that you looked at the question.  Level I questions, with little or no prior knowledge of the topic, lead to exploratory descriptive designs.  Level II questions, where the variables are known but their action cannot be predicted, lead to descriptive survey designs.

Exploratory Descriptive Designs

When the purpose of a study is exploration, a flexible research design that provides an opportunity to examine all aspects of the problem is needed.  As knowledge of the variables increases, the researcher may have to change direction.  Ideas occur as data are collected and examined.  The key to a good exploratory design is flexibility.

The word “exploratory” indicates that not much is known, which means that  a survey of the literature failed to reveal any significant research in the area.  Thus, you cannot build on the work of other; you must explore the topic for yourself.

Even though we talk about the exploratory study as an entity in itself, it should be remembered that it is an initial step in the development of new knowledge.  Because of the flexibility of this type of design, very few, if any, variables are under the researcher’s control.  They are said to be under the control of the situation in other words, observed as they happen or as the researcher comes upon them.  As a result, no inferences can be drawn from the data.  The data may lead to suggestions of hypotheses for further study or to an idea for a conceptual framework to explain the action of the variables, but the exploratory question must be followed by higher-level questions if new knowledge is to be gained.

Your methods and questions will change depending on what you find out as you go along. The results of the study will provide detailed descriptions of all the observations made by the researcher, arranged in some kind of order.  Conclusions drawn from the data include some educated guesses or hypotheses for further study.  A relationship between the observations made and a concept might be proposed.  Further research would be required to test these proposals.  This is the purpose of exploratory research.

Descriptive Survey Designs

Questions at Level II ask, “What is the relationship between or among variables?” You know what the variables are, and you know how to measure them, so you are beyond the scope of an exploratory study.  The variables you are interested in have been studied before, either independently, as in an exploratory study, or with other variables, so that there is sufficient information to ask a question about the relationship between them.  You are able to relate the variables in your study to a concept or conceptual framework so that the study does build on previous work.  The major consideration is accuracy in the measurement of the variables.

Designs for studies at Level II require a descriptive survey.  The design dictates how the variables are to be measured in testing their relationship.  In this type of design, the variables are partly controlled by the situation, as they are in exploratory design, but they are also partly controlled by the investigator, who chooses a sample for the study.  You are not interested in controlling the other variables, merely in observing them.  For example, in a study of the relationship between educational level of drivers and ability to make sound judgments about traffic signs, the investigator controls the first variable by selecting a sample of drivers with all types of educational backgrounds.  The judgments of these drivers are then assessed.  The drivers’ judgments will be examined and evaluated as they occur.  The purpose of the study will be accomplished by seeing if the occurrence of sound judgment is related to educational background.

Many research questions ask about variables that cannot be subjected to experimental manipulation, either because the variables cannot be manipulated or because to look at them outside their natural setting would be meaningless.  For example, in looking at factors leading to mental illness, it would be unethical to isolate a single factor, such as poverty, and manipulate it to see if it results in mental illness.  An experimental design would require that subjects be assigned to groups placed at different levels of poverty.  After the specified length of time had passed, the groups would be examined to see if mental illness had developed.  The absurdity of such an approach is obvious.

Rather than using experimentation to discover the causes of mental illness, you start with the effect and select a sample of mentally ill patients.  Then you look for variables that might be related to mental ill patients.  You might find a significant relationship between poverty and mental illness.  You might also establish that poverty precedes mental illness in time.  However, you might discover that well-to-do persons are less likely to be diagnosed as mentally ill even when they have the same symptoms as person at the bottom of the poverty scale.  Alternative explanations are always possible in descriptive surveys.  Subjects differ on many factors, only a few of which can be controlled.

Although absolute proof of causality cannot be established in a descriptive survey, it is possible to accumulate extensive evidence to support causality. No experimental research has been done with human subjects to see if road accidents can be induced by introducing slippery road surface.  But, by showing that slippery road surface is the one variable preponderant in road accidents, support grows for the theory that the accidents can be caused by slippery road surfaces.

Many variables of interest to transportation researchers cannot be experimentally manipulated.  Attitudes, beliefs, or behaviours are concepts that are often thought of as causal in health, illness response to treatment and other effects.  The descriptive survey can be of great value in the study of these variables.