Research Design & Stats
Types of Research
The purpose of this section is to present an overview of the types of research designs most commonly utilized in the rehabilitation and medical sciences. As consumers of research literature, it is important that life care planners are equipped to critique journal articles and to determine whether study findings are relevant to the patients with whom we are working. Recommendations in the life care plan must be supported by research data and based upon the documented course of disability and patient-specific variables over time.
The manner in which a researcher designs a study determines the type and depth of information which can be concluded from the project. For this reason, researchers spend a significant amount of time considering the most effective design in addressing the topic of interest. First, a researcher determines the purpose of the study. Are they seeking to:
*Explore or describe a research problem? Then surveys, case studies, and developmental studies may be most appropriate.
*Explain or predict relationships between variables? Then relational, associational, correlational, or ex post facto designs may be most effective.
*Control, establish cause and effect, or induce a result? Then experimental or quasi-experimental are likely to yield the most useful results.
Descriptive Research
Descriptive research designs are most effectively applied to studies aimed at gathering additional information, learning more about an area of interest, or becoming more familiar with a topic. Researchers interested in identifying the prevalence or incidence of a disability or in describing the distribution and characteristics of a group of patients may adopt a descriptive research design.
Survey Research
The purpose of conducting survey research is to explore and/or describe an area of interest in greater detail. Typically, a researcher gathers information from a large group of participants either by mail, interactive online site, telephone interview, or personal interview.
Results from this type of study will produce a “snapshot” of the current state of the research issue. A survey does not investigate relationships between variables, patient-specific characteristics, or generalize findings to all members of a population. In other words, the purpose of survey research is not to establish cause and effect relationships. Most individuals have completed a survey of some type, whether a public opinion questionnaire, customer service form, or community investment poll.
Developmental Studies
Developmental studies generally follow a group of participants over time and document changes in status. These types of studies may be referred to as trend, cohort, or panel designs. In most cases, researchers collect data at specific points in time which are used as a basis of comparison and analysis.
Narrative or documentary-style reports may result from this type of design, but conclusions cannot be applied to other members of the target population. Developmental studies are useful in describing phenomena longitudinally and may be used as the foundation for further investigation.
Case Study or Qualitative Designs
The purpose of this type of research is the same as pursued by survey and developmental designs, but generally involves much smaller sample sizes. Rather than gathering information from a large group of people, as in survey designs, the researcher is interested in gaining deeper insights into a fewer number of subjects. This form of inquiry typically involves personal interview or direct observation throughout the data collection process. Examples of this type of design may be a study based on the relationships and interpersonal dynamics of a specific family, documenting the experiences of a small group of patients with traumatic brain injury in a rehabilitation facility, or observing the interactions of selected toddlers in a preschool setting.
Relational or Associational Research
The purpose of relational or associational research is to identify the relationship or association between two or more variables.
Correlational Studies
Correlational studies are appropriate when comparing two or more characteristics within the same group of participants (Ary, Jacobs, & Razavieh, 1999). The purpose is not to establish cause and effect but to:
*demonstrate how specific characteristics vary together, and
*assess the degree to which one characteristic can be predicted when another is known.
There are two types of correlational studies; concurrent and predictive. Concurrent correlational studies involve assessment of the relationship between characteristics that were collected by the researcher at the same point in time. For example, a researcher may be interested in assessing the relationship between patients’ emotional well-being and levels of impairment when beginning a vocational rehabilitation program.
Predictive correlational studies may be utilized when a researcher is interested in determining whether knowing a previously documented characteristic (or set of characteristics) can lead to the prediction of a later characteristic (or set of characteristics). For example, a researcher may record patients’ emotional well-being when beginning a vocational rehabilitation program and then again after 60 days of job placement to ascertain whether earlier data was predictive of later data.
Ex Post Facto Studies
Ex post facto, or “after the fact” designs, attempt to identify a natural impetus for specific outcomes without actually manipulating the independent variable. This type of design is often utilized when it is not possible to control the experience, exposure, or influences which may affect participants.
Experimental or Quasi-Experimental Research
The purpose of experimental or quasi-experimental research is to establish a cause and effect relationship between two variables. The researcher deliberately manipulates a treatment (or independent variable) and measures how it affects the behavior or reaction of subjects (the dependent variable). In order for this research design to be appropriately utilized, the researcher must be able to:
*randomly select subjects,
*randomly assign subjects to intervention and control groups,
*randomly assign treatment to intervention groups, and to
*control the treatment (or independent variable) and any extraneous
variables which may have an effect upon the dependent variable.
As may be apparent from the description, this type of design is very difficult to utilize when working with human subjects. Consider the ethical implications of withholding pharmaceutical treatment from a control group in order to determine the drug’s effectiveness. In order for the drug to be approved for distribution to the general public, it is necessary to establish its effectiveness. However, there exists the potential for harm to individuals who participate in the project, whether receiving the drug or not.
While this type of study occurs, and must occur if advances are to be realized, there are many safeguards, supervisory and administrative requirements, and limitations imposed upon studies involving human subjects. For this reason, quasi-experimental designs are commonly used in medical and allied health fields of study.
* Note: Causality can only be established by true experimental designs, so the term “cause” (or derivatives) should not be used in association with any other design type.
Quasi-experimental designs are used when the researcher is unable to control for a necessary variable, or set of variables, but follow stringent guidelines for controlling error. For example, trials for an experimental pharmaceutical may utilize a quasi-experimental design to ascertain the effectiveness of a specific medication. Subjects are carefully pre-tested and monitored as the independent variable (the medication) is manipulated. Assuming that it was not possible to randomly select or randomly assign subjects to treatment/control groups, a quasi-experimental design (or pre-experimental design) must be adopted.
Life care planners will analyze a variety of research data when developing plans to meet patient-specific needs and should be familiar with the basic assumptions and limitations of each design. After reviewing several research articles, discussing your thoughts with other professionals, and considering the practical applications of various studies, you will gain a level of comfort in evaluating research literature.
References for Further Reading
Ary, D., Jacobs, L., & Razavieh, A. (1999). Introduction to research in education. Fort Worth, TX: Harcourt Brace College Publishers.
Bellini, J. & Rumrill, P. (1999). Research in rehabilitation counseling: A guide to design, methodology, and utilization. Springfield, IL: Charles C. Thomas.
Bolton, B. (1979). Rehabilitation counseling research. Baltimore, MD: University Park Press.
Campbell, D., & Stanley, J. (1966). Experimental and quasi-experimental designs for research. Skokie, IL: Rand McNally.
Fowler, F. (1988). Survey research methods. Thousand Oaks, CA: Sage Publications.
Kerlinger, F. (1986). Foundations of behavioral research. New York: Holt, Rinehart, and Winston.
Kerlinger, F. (1973). Behavioral research: A conceptual approach. New York: Holt, Rinehart, and Winston.
Spector, P. (1981). Research designs. Thousand Oaks, CA: Sage Publications.
Trochim, W. (2001). The research methods knowledge base, (2nd Ed.). Cincinnati, OH: Atomic Dog Publishing.