Selecting research instruments to measure the reliability and validity of nursing research studies

AORN Journal, Nov, 1997 by Leslie H. Nicoll, Suzanne C. Beyea

Research instruments serve as measurement tools and are an integral component of any nursing research study. Research instruments may include physiologic or biologic measurements; self-report instruments, scales, or questionnaires; interviews; diaries; structured or unstructured observations; Q methodologies; psychologic measurements; or other specialized measures.

If a particular concept has been studied, research instruments to measure that concept can be identified and selected from clinical practice or from a literature review. When a concept has not been examined, part of the research process may include developing and testing new measurement instruments. A research instrument must be reliable (ie, consistent) and valid (ie, measure what it intends to measure) and contribute to the scientific integrity of a research project.

Decisions about research measures are made by the researcher and are based on the research question and the design of the study. The research instruments for an investigation need to accurately measure the phenomena under study. Identifying the right research instruments is critical to the validity of the design and the results. The researcher also must examine die fit of the research instruments to die methods. For example, interviews often are employed in qualitative research studies. Physiologic and biologic measures often are associated with quantitative research studies.

Quantitative measurements.

Physiologic or biologic measurements frequently are used in research studies because nurses are quite familiar with these research instruments. Measurements of temperature, blood pressure, pulse, and respiration are among those most commonly used and understood in clinical nursing practice and research. Other physiologic measures often used in research studies include laboratory test results and quantified numeric readings from patient monitoring devices. Such measurements often are described as objective and accurate for research studies. Even these values, however, can be subject to measurement error. For example, if a sphygmomanometer is not calibrated or the researcher has a hearing problem, blood pressure readings from subjects may not be accurate.

Physiologic data also can be obtained by observing the behavior of research subjects or having them complete self-reports or questionnaires. These types of instruments can provide information about physiologic parameters that are not measured easily by conventional methods. For example, subjective experiences such as vertigo, nausea, or pain are best measured by the individual experiencing the phenomenon. Developing consistent measurements and recording that data provides researchers with a basis to examine these patient experiences.

Qualitative measurements.

Self-reports and questionnaires frequently are used to obtain data about subjects' feelings, attitudes, or knowledge. Perioperative nurses often ask questions about their preoperative anxiety or knowledge of preoperative teaching. There also are a number of well-developed research instruments that measure anxiety or knowledge level. Common approaches to obtaining this type of information is through written surveys or interviews.

When using interviews, the researcher decides in advance to use either structured or unstructured interviews. Structured interviews are designed by the researcher to direct both the content and focus of the interview. By developing specifically worded questions and asking them in a predetermined order, the researcher ensures a certain level of consistency when conducting the interviews.

Unstructured interviews usually start with a broad open-ended statement such as "Please tell me more about your experience undergoing emergency surgery." In this instance, the researcher simply encourages the subject to describe in detail a specific experience. The interviewer's goal is to obtain as much information as possible without determining the direction of the interview. In this situation, the research subject has an opportunity to control the direction of the responses. The questions that are asked during interviews or as part of self-report instruments, scales, or questionnaires must be phrased in a manner that best asks the question.

Developing a questionnaire is a complex and sophisticated process. The researcher must first examine the concept or construct being researched. In the absence of a tested research instrument, the researcher must proceed with item and tool construction. Each item must be carefully worded and easily understood by the reader. The researcher also must be sensitive to factors such as gender, culture, sequence (ie, usually general to specific), and language. Many researchers have leaned the hard way that if the right question is not asked, the desired answers will not be obtained. Most researchers obtain expert consultation if they are faced with instrument or interview development.

When a researcher conducts interviews, the responses usually are recorded and later transcribed. Then the researcher uses qualitative research methods to analyze the words of the participants. The researcher often is trying to identify repeating themes or ideas among the responses. Remember that the level of analysis in qualitative research is words, not numbers.