The Research Enews team invites your research word!
We are currently seeking to build a glossary of research terminology that will serve the holistic nursing community. In each issue of Connections in Holistic Nursing Research, a new term will be featured and the glossary will be on the website as well.
If you have a research term you think would benefit AHNA members, and would like to be a guest contributor to our glossary, please contact
firstname.lastname@example.org with your recommendation.
Companion Animals Power Analysis
Concept Analysis Randomization
Correlation - Its Role in Research The RCT: Randomized Controlled Trial
Critical Incident Technique Theory of Compassion Energy
Data Collection in Nursing Research Theory of Integral Nursing (TIN)
Dissemination Threats to Internal and External Validity
Effect Size and Significance Triangulation
Grounded Theory Variables in Quantitative Research
Hawthorne Effect Validity
Phenomenology What Are the Four Levels of Measurement?
By Mara M. Baun, PhD, FAAN
Animals are playing many therapeutic roles in the lives of humans. Many people have had animals as pets, especially when they were children, and fondly remember their relationships. Today, animals not only are highly valued in many homes, but they have also assumed therapeutic roles with humans.
At present, animals frequently are visitors in health care institutions. Many hospitals and nursing homes allow animals, particularly dogs, to visit patients/residents. These visitors can be “therapy dogs” who are unknown to the patients or in some hospitals the patients’ own dogs. Certified therapy dogs generally visit a number of patients who request them. The patients’ own dogs usually just visit their owners and are known as “companion animals”. Various studies have shown that humans respond more positively both physiologically and psychologically to dogs to which they are attached than to unknown dogs (Baun, Bergstrom, Langston, & Thoma et al., 1984; Schuelke, Trask, Wallace, Baun, Bergstrom, & McCabe, 1991/92).
It is not uncommon for nursing homes to have “resident” dogs and sometimes other animals who wander freely among the residents. Besides dogs, the most common are cats, rabbits, small rodents, birds, and fish. Problem behaviors of persons with Alzheimer’s were measured for one week prior to and 4 weeks after the placement of a dog. Participants on the day shift exhibited significantly fewer problem behaviors across the entire four weeks after the dog was living on the unit (McCabe, Baun, Speich, & Agrawal, 2002).
In another study conducted in an Alzheimer’s unit, it was found that particularly during the sundown period when persons with Alzheimer’s can become very agitated, residents were much calmer when a therapy dog was present (Churchill, Safaoui, McCabe, & Baun, 1999). For example, one person with Alzheimer’s who was pacing in the hallway responded very positively when a resident dog took him by the sleeve and lead him back to his room.
Caged birds in the rooms of the elderly in rehabilitation have been shown to decrease depression. Residents reported that their grandchildren enjoyed coming to visit with them and stayed longer, and one resident said that her physician came in each morning and sang a song to the bird. Another resident who received a diagnosis of a terminal disease while at the institution said that she talked to the bird about it, and her depression level decreased (Jessen, Cardiello, & Baun, 1997).
Children seem to have special relationships with animals. Children having physical examinations in a pediatric clinic were much calmer and less stressed when a therapy dog was present in the examination room (Nagengast, Baun, Leibowitz, & Megel, 1997); Hansen, Messinger, Baun, & Megel, 1999). Likewise, children undergoing dental procedures in a dental office and who had a therapy dog present experienced less physiological arousal and behavioral distress during the procedures than those who did not have a therapy dog (Havener, Thaier, Gentes, Megel, & Baun, 2001).
In summary, animals, especially dogs, are very active in health care these days. In a variety of settings they are having very positive effects. More research, however, needs to be done on the health benefits of companion animals.
Baun, M. M., Bergstrom, N., Langston, N. F., & Thoma, L. (1984). Physiological effects of human/companion animal bonding. Nursing Research, 33, 3, 126 129.
Churchill, M., Safaoui, J., McCabe, B. W., & Baun, M. M. (1999). Effects of a therapy dog in alleviating the agitation behavior of sundown syndrome and in increasing socialization for persons with Alzheimer’s Disease. Journal of Psychosocial Nursing and Mental Health Services, 37, 4, 16-22.
Hansen, K. M., Messinger, C. J., Baun, M. M., & Megel, M. (1999). Companion animals alleviating distress in children. Anthrozoös, 12. 3, 142-148.
Havener, L., Thaier, B., Gentes, L., Megel, M., & Baun, M. (2001) The effects of a companion animal on distress undergoing dental procedures. Issues in Comprehensive Pediatric Nursing, 24, 2, 137-152.
Jessen, J., Cardiello, F., & Baun, M. M. (1996). Avian companionship in alleviation of depression, loneliness, and low morale of older adults in skilled rehabilitation units. Psychological Reports, 78, 339-348.
McCabe, B. W., Baun, M. M., Speich, D., & Agrawal, S. (2002) Resident Dog in the Alzheimer’s Special Care Unit. Western Journal of Nursing Research, 24, 6, 684-696.
Nagengast, S. L., Baun, M. M., Leibowitz, M. J., & Megel, M. (1997). The effects of the presence of a companion animal on physiological and behavioral distress in children during a physical examination. Journal of Pediatric Nursing, 12, 6, 323-330.
Schuelke, S. T., Trask, B., Wallace, C., Baun, M. M., Bergstrom, N., & McCabe, B. (Winter, 1991/92). Physiological effects of the use of a companion dog as a cue to relaxation in diagnosed hypertensives. The Latham Letter, 14-17.
By Anita Catlin, DNSc, FNP, FAAN, Research and Ethics Consultant
Concept analysis in nursing refers to a multifaceted analysis of a term. Examples for holistic nurses might be deeper looks into the concepts of therapeutic touch, caring or alternative medicine. The concepts chosen for analysis by nurses are often those felt deeply about and used in research and practice.
Concept analysis is initiated by thinking:
- "What is it that I am interested in?"
- "How would this term be defined in the Webster's or Oxford dictionary?"
- "What similar terms can describe this concept?"
This would be followed by a literature search. The intent is to find a term or terms similar to the one being looked at. Decisions are made on time frames. Data may come from long ago, when Florence Nightingale, for example, defined "hygiene," or from more recent times, such as "what is a rapid response?" Articles are sorted into which articles actually contributed to the meaning of the term, and which articles were not useful in the analysis. It is common to start off with more than 100 abstracts and hone down to 25 articles. The literature is then closely scrutinized to develop a definition of the concept for nursing. Defining the concept begins with several components:
- attributes of the concept ("What is contained in it?"),
- antecedents to the concept ("What has to occur for the concept to take place?"),
- and consequences of the concept ("When the concept occurs, what happens?").
Cases are then developed; a perfect example of the concept in narrative form, borderline cases, and opposite cases. At the end, a discussion occurs as to how the concept can be theoretically applied to nursing and implications for clinical practice.
After some years of conducting concept analyses, Schwartz-Barcott and Kim (1993) took the analysis further calling the process hybrid analysis and included fieldwork. It was suggested that after the literature review, definitions, attributes, antecedents and consequences were studied, the researcher must go into the clinical area and test the concept.
- "Did the concept appear in reality as it did in theory?"
- "Were the definitions correct?"
- "Do the people affected by this concept agree?"
- "What are their opinions of how the concept is used?"
- "How do nurses apply the concept?"
- "Will it work in clinical practice?"
Various methods of testing the concept can be done such as focus groups, surveys, and interviews. The researcher attempts to gather real time information from those who use or will use the concept. Clinical reality may influence, change and/or verify that the concept is correctly defined and useable.
An example of concept analysis in holistic nursing is Makaroff (2012) who provided an interesting analysis on the "unsayable." Going back to 1959, she reviewed 1557 articles and texts that might contribute to when there were topics in patient care which simply could not be talked about, were taboo, or were not consciously expressed through language.
In another example, Catlin and colleagues (2008) were concerned about patients at the end of life receiving technological life extending therapies which seemed to be causing suffering. The concern was that the nursing care being delivered was not of benefit to the dying patient and was perhaps causing harm. Conscientious objection to care orders that cause suffering at the end of life was considered. A concept analysis of the literature was conducted with much of it from the military rather than from nursing. After defining the concept, attributes, antecedents, and consequences of conscientious objection, these researchers surveyed 66 nurses concerning their feelings about conscientious objection to futile care in neonatal and pediatric patients. Recommendations for practice came from both the theoretical work and clinical reports of the surveyed nurses.
Concept analyses provide clarity when a nurse wishes to implement a new program or begin an intervention study. Holistic nurses will find that using concept analysis will benefit patients through increased knowledge of the evidence backing the care provided.
Bonis, S. A. (2013). Concept analysis method to enhance interdisciplinary conceptual understanding.Advances In Nursing Science, 36(2), 80-93.
Catlin, A., Armigo, C., Volat, D., Valle, E., Hadley, M., Gong, W., Bassir, R., & Anderson, K. (2008). Conscientious objection: A potential neonatal nursing response to care orders that cause suffering at the end of life? Study of a concept. Neonatal Network, 27(2), 101.
Cronin, P., Ryan, F., & Coughlan, M. (2010). Concept analysis in healthcare research. International Journal of Therapy & Rehabilitation, 17(2), 62-68.
Makaroff, K.L.S. (2013). The unsayable: A concept analysis. Journal of Advanced Nursing, 69(2), 481-492.
Rodgers, B.L. & Knafl, K.A. (1993). Concept development in nursing. Philadelphia, PA: W.B. Saunders.
Schwartz-Barcott, D. & Kim, H. (Chapter 8, 1993) in Rodgers, B.L. & Knafl, K.A. (1993). Concept development in nursing. Philadelphia, PA: W.B. Saunders.
Walker, L.O. & Avant, K.C. (2000). Strategies for theory construction in nursing. Upper Saddle River, NJ: Pearson, Prentice-Hall.
Wilson, J. (1963). Thinking with concepts. London: Cambridge University Press.
Correlation - Its Role in Research
By Sue Roe, DPA, MS, BSN, RN
Correlation is used in a variety of contexts and is a term heard in many casual conversations. Someone might say, "Did you know there is a correlation between tall people and the type of car they drive" or "There is a correlation between popcorn lovers and the number of movies they watch". While used often, correlation is a term with special meaning.
The word correlation has an interesting background. It has been part of the English language since the 16th century. The Latin origin for correlation is the root "co" which means with, together, or join. By the 19th century, correlation became a term many disciplines adopted for their own uses. Therefore, it is important to be aware that when one employs correlation in research and statistics, it has its own unique use.
The best way to define correlation is to say that it is a relationship existing between phenomena, sets of data, or variables which tend to vary, be associated, or occur together and are not expected to do so by chance alone.
The purpose of correlation in research and statistics is to provide information about the relationship between sets of data or variables. As a statistical measure correlation indicates the extent to which two or more sets of data or variables may or may not fluctuate together. When this happens, it does not imply "causation" but rather in correlation the reason for fluctuation is that the sets of data or variables occur at the same time. For example, when there is a strong relationship between the number of people who smoke and who also drink alcohol, it does not mean there is a causal relationship between the two but rather one is associated with the other in a particular way. So, in correlation the focus is not on cause but on the strength and also the direction of the relationship.
Possible correlations can range from +1 to -1. A zero correlation indicates there is no relationship between the data or variables. A correlation of -1 points to a perfect negative correlation, meaning as one goes up, the other goes down. For example, the more miles a person drives, the less fuel there is in the car. A correlation of +1 implies a perfect positive correlation, meaning both move together in the same direction. For example, as people age, their height increases.
These scatterplots of data below show different correlational directions. Also noted are the strength of each. The closer to +1 or -1, the greater the strength of the relationship.
Scatter plots from Pierce, Rod. (8 Feb 2014). "Correlation". Math Is Fun. Retrieved 9 Jul 2014 from www.mathsisfun.com/data/correlation.html
Statistically, correlation is appropriate for examining relationships between quantifiable data (e.g., temperature, weight) rather than categorical data such as gender or a person's favorite sport and it is measured using a coefficient of correlation. The statistical tool most familiar is Pearson Product Moment Correlation where a linear relationship (as above) is examined. Here interval level data are required. When measured in a population Pearson Product Moment Correlation is labeled by use of rho (ρ). When in a sample, by the letter "r"; sometimes called Pearson's r. For example, a resulting correlation coefficient r might be +.85 which would be interpreted as a strong positive correlation.
There are limitations to using correlation.
- While helpful in analysis, it does not completely speak to all one wants to know about the data and results.
- Outlying data/variables can strongly influence the correlation coefficient.
- It does not work as well with curvilinear relationships, those that do not follow a straight line. For example, if the relationship being examined is age and the use of complementary therapies, while they may be related, the relationship may not follow a straight line. Younger adults may use a variety of different therapies than those who are older.
- If the sample size is small, correlation should not be used alone. Here, significance level is important because it will provide information about how likely the correlation may be due to chance because of a random sampling error.
While, Pearson Product Moment tends to be the most used statistical tool when examining correlational relationships, there are others that are appropriate for different levels of data. If there are two ordinal level variables, Spearman Rank Correlation Coefficient (rho) or the Kendall Rank Correlation Coefficient (tau) can be selected.
Correlation is one of many ways of reporting research results. While, it determines the strength and direction of a relationship which is most helpful it can also offer a launching point for more detailed analyses in later studies.
Critical Incident Technique
By Leighsa Sharoff, EdD, RN, NPP, AHN-BC
The Critical Incident Technique (CIT) is a qualitative, systematic, open-ended technique for educing descriptive data from participants. It can be used as a primary source of data collection or as an evaluation technique to supplement the interview or questionnaire questions as well as provide data for triangulation. The CIT is a powerful methodological instrument that can foster reflection and promote personal expression. CITs can be used to assist nurses and healthcare professionals to gain a deeper and more personal understanding of clients, acquiring their perspectives and concerns as well as a more personal awareness of oneself. It can be developed to conform to any area of nursing and provide a more comprehensive awareness of what nurses do and the needs of our clients.
Read the entire article Here.
Data Collection in Nursing Research
By Pamela Crary, PhD, RN, AHN-BC
Common data collection and measurement methods used in nursing research include physiological measures, observation, interviews, questionnaires, and scales.
Observational measurement involves an interaction between the subjects and the observer(s), where the observer has the opportunity to watch the subject in a specific setting. Unstructured observations involve spontaneously observing and recording what is seen in words. Structured observations are carefully defined, with specific guidelines about how the observations are to be made, recorded, and coded.
An interview involves verbal communication between the researcher and the subject. There are a variety of approaches, ranging from unstructured interviews, in which the content is completely controlled by the subject, to structured interviews with carefully designed responses.
A questionnaire, or survey, is a self-report form designed to elicit information through written, verbal, or electronic responses. They are distributed and administered in a variety of ways. The information obtained from questionnaires is similar to that obtained by an interview, but the questions tend to have less depth; however, questions are presented in a consistent manner and the opportunity for bias is less.
A scale is a more precise means of measuring than a questionnaire. Most scales measure psychosocial variables; there are some used with physiological variables, such as pain or nausea. Items on summated scales are summed to obtain a single score.
Commonly used scales include rating scales, Likert scales, and visual analog scales.
A rating scale lists an ordered series of categories of a variable that are assumed to be based on an underlying continuum. A numerical value is assigned to each category, and the fineness of the distinctions between categories varies with the scale.
Likert scales, the most commonly used of scaling techniques, are designed to determine the opinions or attitudes of study subjects. This scale contains a number of declarative statements with a scale after each statement. Response choices in a Likert scale most commonly address agreement, evaluation, or frequency.
The visual analog scale (VAS) is typically used to measure strength, magnitude, or intensity of individuals' subjective feelings, sensations, or attitudes about symptoms or situations.
By Diane Wind Wardell, PhD, RN, WHNP-BC, AHN-BC, CHTP/I
Dissemination is the final step in the research process. In an evidenced based practice project this would be the implementation step. It is critical that the information be shared. It is the responsibility of the team to disseminate the results especially when individuals (both human and/or non-human) have given of their time and energy to participate in the project. It is a moral responsibility.
Many granting agencies request specifics on the audiences and mechanisms that will be utilized to meet this goal. Usually the proposal (grant, dissertation, project, etc) identifies the areas were this information will be clinically significant, for example, by suggesting the journals that would be interested in the findings. No matter the size of the study, a review of the findings can be made in some public venue at the local, regional, national or international level. Posters may be a first step as criteria for acceptance are usually less stringent. Posters provide an opportunity to synthesize the study from the conceptual framework, question to be answered, methods used, the sample characteristics, findings, through the discussion. Poster sessions give individuals an opportunity to ask questions and share their ideas on a one-on-one format. Podium presentations still require the same information but there is a specific timeframe in which the content is to be delivered. A larger audience may be reached in this context and often the abstract and/or presentation is published in the meetings proceedings. Both of these steps provide the opportunity to organize the material in a meaningful way and helps in the preparation of a manuscript. Manuscript preparation should involve the team with each person contributing in some way to the final product. It is good, however, to designate the primary author up front so this person can be responsible for keeping the writing on task. It is also helpful to have a journal in mind when writing as it can help target specific requirements that might be request (i.e. nursing journals often want a conceptual framework which might not be asked for in other professional journals). If asked to do major edits to the manuscript by the editor it is usually wise to attend to this request by addressing each question specifically. Identify what has been changed or why not in the response to the editor recognizing that the person who made the suggestion may very well review this response. The wider audience you wish to reach with your findings can be influenced by the journals availability through a national search engine like PubMed. Additionally, journals are given an "impact score" which is assigned by such things as the numbers of readers/subscriptions and commonality of citations.
Dissemination can be made more difficult when the findings are not significant or do not provide a clear explanation of the results. However, it is still critical that this information be shared. Posters and presentations are still available. Publication is also important to identify both the strengths and weaknesses of approaches (i.e. timing of and modifications to the intervention, sample population and size). More attention can be placed in the discussion section to explain the possible reasons for the results.
This final step in the research process is an invaluable contribution and one that requires fortitude and integrity. Holistic nurses often have complex studies to present as they are interested in the body-mind-spirit and how it is sustained for healing of the self and others. The greater the body of research literature we have
the more support there is available for others to provide holistic care.
Effect Size and Significance
The Effect Size (ES) in a study is a relative number expressing the strength of the relationship between statistical populations (sample and control) and the interventions they were exposed to. This measure of association is complex. The researchers and the reviewers must ensure that the study design has internal validity, free of bias and accounts for confounding variables and random error. ES is expressed as small, moderate or large using numbers between -1 and +1. Because the ES is a relative number it complements other statistical measures. There are also several measures of ES dependent upon the type of study. There are three common approaches to determining the effect size; 1) statistical significance, 2) practical significance using the raw mean differences of experimental groups, and 3) relative size of the effects based on standardized estimates.
Reference: Effect Size Guidelines - Effect Size Substantive Interpretation
Significance in statistics represents a causal relationship, rather than chance occurrence (McGraw-Hill, 2002). A "statistically significant" result is one that would occur by chance less than a certain percent of the time. Usually significance level is set at .05 (95 percent likelihood not due to chance) but other levels such as .01 are commonly used (99 percent likelihood not due to chance). It has been argued that significance level should always be accompanied by effect-size statistics to understand the size and importance of the difference (Wikipedia, 2011). Historically, Lehrer (2010) explained that .05 as cutoff was a somewhat arbitrary decision made by English mathematician Fisher, in 1922, as it made pencil and slide rule calculations easier. Often the .05 level is chosen because it is conventional (Wikipedia 2011).
Lehrer, J (2010). The Truth Wears Off: Is there something wrong with the scientific method? Retrieved:
Statistical Significance (2002). In McGraw-Hill Concise Dictionary of Modern Medicine. Retrieved:
Statistical Significance (2011) in Wikipedia. Retrieved:
Grounded Theory: Applications in Holistic Nursing Research and Practice
By: Wendy Stiver, RN, BSN, MA
Introduction and History
Grounded Theory (GT) is an inductive methodology that leads to the generation of theory from systematic research processes. GT was developed by sociologists Barney Glaser and Anselm Strauss in the 1960s, and is a general method that can be applied in any field, including nursing. GT is most often discussed within the context of qualitative research. Other types of qualitative research include phenomenology, hermeneutics, ethnography, and historical research (Zahourek, 2013).
Andrews and Scott (2013) point out that the term Grounded Theory refers to the research method and the resulting theory grounded in the data. The first step in GT methodology is to identify the substantive area, or the area of interest for exploration. The second step is to collect data relevant to the substantive area. One of the features of GT is that data can be a mix of qualitative and quantitative, including interviews, observations, reflective journals, media reports, and even emails or smartphone texts. The researcher approaches the subject with an open mind and receptivity to whatever answers are found. "GT helps us to see things as they are, not as we preconceive them to be....GT orients us to seeing our behavior and the behavior of others as data" (Glaser, 2014, p. 48). A typical GT study may include in-depth interviews of varied numbers of participants, along with other data sources. A key principle of GT is that everything is data, and the researcher analyzes and then codes written data to identify the core category, which explains the main concern of the participants.
The researcher uses the constant comparative method of data analysis to compare elements or themes identified among the interviews or data sources. Another key dimension of GT is memoing. "Memos are the written records of the researcher's thinking, both conscious and preconscious realizations as the research and the researcher grows....There are no rules for writing them.....Memoing is not optional...memos ensure the quality of the emerging theory" (Glaser, 2013, p. 2). Memos contain the intellectual processes of the researcher and capture the evolution of the theory itself. Memos are sorted as the researcher moves toward the writing and rewriting phases of the research methodology. The sorting stage is also when the researcher may review pertinent literature to look at more data and to compare what has emerged from the coding process.
The outcome of the GT method is the emergence of a Grounded Theory that fits the data, works to account for how participants solved their main concern, is relevant to the field, and is modifiable in further development (Giske & Artinian, 2007).
Application to Holistic Nursing
Grounded Theorists strive to understand peoples' lives and concerns (Gelling, 2011), thus making GT relevant to holistic nursing research and practice. Both qualitative and quantitative research contribute to expanding our knowledge of human experiences and our "inner lives"; thus, a general method such as GT can be very useful in holistic nursing research.
Zahourek (2013) points out that holistic nursing research must incorporate holistic frameworks into the research process from conceptual origins through to interpretation of findings. An example of such research is Zahourek's article in 2005, "Intentionality: Evolutionary Development in Healing. A Grounded Theory Study for Holistic Nursing". The article show how the GT methodology was employed, the sampling processes, and the development of the theory of intentionality: the matrix for healing.
In summary, Grounded Theory is a systematic research approach which, when done well, results in a grounded theory that fits the pertinent data set and can speak to the broader human condition. GT has been embraced by a number of nurse researchers and has informed holistic nursing research and literature.
Andrews, T., & Scott, H. (December 18-20, 2013). Grounded theory seminar. Presented at Manchester Metropolitan University, UK. Retrieved from www.groundedtheoryonline.com.
Gelling, L. (January 31, 2011). What is the difference between grounded theory and phenomenology? NursingTimes.net. Retrieved from www.nursingtimes.net/nursing-practice/clinical-zones/educators 10/13/2014.
Giske, T., & Artinian, B. (2007). A personal experience of working with classical grounded theory: From beginner to experienced grounded theorist. International Institute for Qualitative Methods, 67-80.
Glaser, B. G. (2014). Applying grounded theory. The Grounded Theory Review, 13(1), 46-50. Retrieved from www.groundedtheoryreview.com.
Glaser, B.G. (2013). Introduction: Free style memoing. The Grounded Theory Review, 12(2), 1-13. Retrieved from www.groundedtheoryreview.com.
Zahourek, R. P. (2013). Holistic nursing research: Challenges and opportunities. In B.M. Dossey, L. Keegan, C.C. Barrere, & M. B. Helming (Eds.), Holistic nursing: A handbook for practice (Sixth Edition). (pp. 775-796). Boston: Jones & Bartlett Learning.
Zahourek, R. P. (2005). Intentionality: evolutionary development in healing. A grounded theory study for holistic nursing. Journal of Holistic Nursing, 23(1), 89-109.
By: Diane Wind Wardell, PhD, RN, WHNP-BC and Diana Guthrie, Ph.D, ARNP, BC-ADM, CDE, FAAN, FAADE
The Hawthorne Effect is a placebo type effect that involves a change in the dependent variable resulting from subjects' awareness that they are participants under study (Polit & Beck, 2012, p. 729). This effect was identified in research by Henry Landsberger in 1955 by analyzing data from experiments carried out in Hawthorne, Chicago between 1924 and 1932, by Elton Mayo at the Western Electric Corporation (Shuttleworth, 2009). It is the process where human subjects in an experiment change their behavior, simply because they are being studied. This is one of the hardest internal biases to eliminate or factor into a design. The fact that the workers in these studies were singled out or observed was enough to change the results of the studies. Consequently, what has developed is an attempt to control for this innate human response to attention (Hawthorne effect and placebo effects) that modifies results. Instead, researchers attempt to design studies in which the variables that are being manipulated are not influenced by this attention. In order to do so in modern day science it is believed that the participant needs to be "blinded" to their group assignment (intervention or control) so that this can not influence the outcomes that are being evaluated. Additionally, the control group can not simply be "no treatment" but one that has a similar level of attention provided. For example, if designing a meditation study one would have 10 minutes of meditation daily via tape and the control group would get equal attention by listening to 10 minutes of self-help information.
Not only does the "fact" of being observed or being a participant in a study influence the response of participants but their expectations about the effectiveness, (or non), personal experiences, and beliefs about if they are in the control or active treatment group can also influence the outcome. Luana and Miller (2011) recently wrote about these issues in relation to medication and behavioral studies and offered that it can be a complex function between participant, intervention, and presenting condition. They provide an interesting review of these factors and suggest that this information is an element of clinical practice
Luana, C. & Miller, F. (2011). Role of expectation in health. Current Opinion in Psychiatry, 24(2), 149-15. DOI: 10.1097/YCO.0b013e328343803b.
Polit, D., & Beck, C.T. (2012). Nursing research: Generating and assessing evidence for nursing practice, 9th Ed. Philadelphia, PA: Lippincott Williams and Wilkins.
Shuttleworth, M. 2009. The hawthorne effect and modern day research. Retrieved from http://www.experiment-resources.com/hawthorne-effect.html
By: Jen Reich, PhD, RN, ANP-BC
Phenomenology is both a philosophical tradition and human science method (Dowling, 2007, Wojnar & Swanson, 2007, Van Manen, 2002). Phenomenology seeks a deep understanding of lived experiences in our human world (Starks & Brown Trinidad, 2007, Van Manen, 1990). Van Manen (1990) noted that a real understanding of phenomenology can only be done by doing phenomenology. He explained that in the process of doing phenomenological research, we become connected to it, thus we “become the world.” (Van Manen 1990, p. 5). Heidegger terms this being in the world “dasein” (Koch, 1995).
A major concept in phenomenology is intentionality. Crotty (1998) explained that that intentionality in the phenomenological sense is means referentiality, relatedness, “aboutness”, rather than purpose or deliberation (p.44). Intentionality posits a relationship between conscious mind and object of consciousness (Crotty, 1998). Existential phenomenologists would explain this as a “radical interdependence of subject and world.” (Crotty, 1998, p.45). This unity of subject and object that intentionality posits requires a rejection of objectivism and subjectivism (Crotty, 1998)
Of the two major schools of phenomenology, Heideggerian and Husserlian, Heidegger’s version of phenomenology is more existential, seeking the meaning and understanding of our being in the world (Koch, 2005). Husserl’s method is descriptive, and stems from the Cartesian tradition, describing phenomena as brought through consciousness (Koch, 2005).
Since Husserl and Heidegger, there have been seven unique perspectives of phenomenology identified (Wojnar & Swanson, 2007). Max van Manen, a phenomenologist from the Utrecht (Dutch) tradition, has guided the research of many in the health profession and education fields (Dowling, 2007). Van Manen expressed that the ultimate goal of phenomenology “is to effect a more direct contact with the experience as lived” (Van Manen 1990, p.78). His work is considered a combination of descriptive and interpretive phenomenology (Dowling, 2007).
In both descriptive and interpretive traditions, phenomenology is intended to be an initial critique and not a be-all, end-all method (Crotty, 1998). Crotty explained that it is a valuable starting point in social inquiry, with research for the phenomenologist an attempt to “break free and see the world afresh.” (Crotty, 1998, p. 86).
Crotty, M. (1998). The Foundations of Social Research. London: Sage
Dowling, M. (2005). From Husserl to van Manen: A review of different
phenomenological approaches. International journal of nursing studies,
Koch, T. (1995). Interpretive approaches in nursing research: The
Influences of Husserl and Heidegger. Journal of Advanced Nursing, 21: 827-83
Starks, H & Brown-Trinidad, H (2007). Choose your method: A comparison of
phenomenology, discourse analysis and grounded theory. Qual Health Res 17: 1372-1380
Van Manen, M. (1990). Researching Lived Experience: Human Science for an Action
Sensitive Pedagogy. New York: SUNY Press.
Wojnar, D. M., & Swanson, K. M. (2007). Phenomenology: an exploration. Journal of
Holistic Nursing, 25(3), 172-180
By: Melodee Harris, PhD, APN, GNP-BC
Researchers use a power analysis to determine the sample size before conducting the study and to determine statistical significance after the study is completed. This is important when one is conducting a study that has as its purpose determining the benefit of one treatment/intervention over another. There are a variety of computer programs that may be used to calculate a power analysis. In holistic nursing research this may be more of a challenge as multiple factors may be contributing to the affect a particular treatment/intervention has on the participant.
Power is a function of effect size and sample size. Effect size denotes the degree of relationship between the research variables.. A power analysis is the combined effect size and sample size and is used to make a more precise prediction of the study results. A small effect size requires a larger sample size. If the intervention has a large effect size, fewer responses or participants are needed. Cohen sets a range to determine a small, medium, and large effect size. Effect size may be referred to as Cohen's d. Sometimes the ranges are used to estimate these values and sometimes the effect size is estimated from previous research.
A power analysis that meets ethical standards is performed prior to conducting the study in order to determine resources needed to carry out the research. A sample size that is larger than necessary wastes valuable resources and places an unnecessary burden on participants. A sample size that is underpowered will not determine conclusive results of a study. It is important to remember that a power analysis is only a calculated estimate that provides the researcher with an objective means for guiding a scientific basis for the statistical significance of the study.
Dr. Kristin Wicking, RN, BSN, MSN, PhD
There are two main kinds of randomization: random sampling and random assignment to treatment groups. In this issue we will only focus on random assignment, also called random allocation.
Randomization: "The random assignment of subjects to treatment conditions."
Random assignment: "A strategy used to assign subjects to experimental or comparison/attention control groups by probability (i.e. in a manner determined by chance alone). Each subject has an equal chance of being placed in to any one of the study groups." (Melnyk & Morrison-Beedy, 2012, page 462.)
Random assignment is concerned with how we as the researcher decide which subject or participant ends up in which one of two (or more) treatment groups, or treatment conditions. Who gets the experimental cancer drug and who gets the sugar pill? Who gets the special 1:1 tutoring session with the math expert and who gets nothing? Let's look at a lighthearted example to see why random assignment in to groups is so important in research.
Imagine that you had 10 high school boys lining up to play a game of basketball. You need 2 teams of 5 boys each. You decide to let them sort themselves in to the two teams.
So what might happen? They would self-select according to what mattered most to them. Perhaps the boys who always shoot hoops together on Thursday afternoons after school would clump themselves in to a team, and leave all the others behind. You might end up with one really strong team and one really weak team. Or perhaps all the boys who were happy to be on the 'skins' team were the ones with more athletic bodies, while those who chose to be on the 'shirts' team would be the ones who were less confident about showing their bodies. Maybe all the tall boys would end up on one team, or all the blonde boys. Anytime you leave it up to the 'participants' or 'subjects' in a study to 'self select' which group they will belong to, then you are introducing numerous possibilities for 'confounding variables' such as body types, hair color, body image, etc.
Let's say that you're not going to play an ordinary game, but instead you are going to run a little experiment. One team is going to get a special basketball lesson from a visiting basketball celebrity, while the other team is not. Then they will all 10 be given a chance to shoot 10 free throws and rated on how many baskets they scored out of ten attempts. Once you tell the boys about the celebrity coming, let's say Michael Jordan, then many of them would be very eager to meet and learn from Michael, so they might jostle to be on that "dream team." You could end up with the most assertive or confident boys all on the same 'experimental condition' team, because they pushed to be in the group that would meet/learn from Jordan; while the 'control ' team has all the boys who are less enthusiastic about basketball or about Michael Jordan, or who are less assertive about putting themselves forward.
Randomly allocating your 10 boys in to the two teams solves a host of problems. By using randomization, you are far more likely that the blondes, the athletic bodies, the assertive boys, the tall ones, will be evenly distributed across both teams. Then when you run your experiment, you can say with more confidence that it was the lesson from Michael Jordan that made the difference between the scores of the two teams, NOT the height, assertiveness, body type or hair color that accounted for the difference between the two teams. Your dependent variable is the free throw score, and your independent variable, or the experimental condition you are varying or manipulating, would be the lesson (or not) from Jordan.
But randomization is even more powerful. Because things like athletic bodies, hair color, even personality types, you might realize could affect your experiment as confounding variables, and so you could make a point of "splitting up" all the blonde, buffed, assertive boys and spreading them equally between the two teams.
But what randomization does for you, its hidden and even more potent power, is to ALSO help ensure that your two teams are equal on a host of other variables or features that you might not ever even think about as you are designing the experiment. Perhaps the fact that they have a dad who is a mad keen basketballer and practices with them every weekend could also be a confounding variable, but you can't tell that just from looking at the boys. Or perhaps some other physical trait like joint mobility or skeletal strength or peripheral vision could also confound your experiment. It's impossible to ask about or assess for every possible confounding variable, or even to know (guess) enough to think to ask about it.
But if you use randomization to decide which boys end up in which group (random assignment), then you sidestep all these problems, and greatly increase your chances that any feature or trait that might affect your experiment will be equally represented in both teams. If you have randomly allocated, then there will be just as many flexible, strong boys with excellent peripheral vision in the 'dream team' as in the non-celebrity team.
Randomization is easy to do and well worth the effort. You can generate random numbers in a simple computer program like Excel or even on a website like this one:
So always consider if you can randomly allocate your participants into groups, and do so whenever possible. And as any high school boy would tell you, leaving it to chance, or random assignment, means every boy has an equal chance of meeting and learning from the famous Michael Jordan. For a stronger research design, and to be as fair and ethical as possible, and to give everyone an equal chance of scoring a few moments with Michael, then use random assignment into treatment groups.
And let's play ball! :)
Melnyk, B.M. & Morrison-Beedy, D. ( 2012). Intervention research: Designing, conducting, analyzing, and funding. New York: Springer Publishing Company.
The RCT: Randomized Controlled Trial
Randomized controlled trial: (RCT) An experimental design in which individuals are assigned randomly to two or more groups: a treatment group (experimental therapy) and a control group (placebo and/or standard therapy) and the outcomes are compared. Someone who takes part in a randomized controlled trial (RCT) is called a participant or subject. RCTs seek to measure and compare the outcomes after the participants receive an intervention. Because the outcomes are objectively measured (instruments and/or physiological data) RCTs are quantitative studies.
The RCT is currently the most accepted scientific method of determining the benefit of a drug or a therapeutic procedure. It is one of the simplest and most powerful tools in clinical research.It can represent the "best" evidence available, which is integrated into the final decision about the management of a condition by healthcare practitioners in what is called evidence-based healthcare but is limited by its generalizability to groups.
In sum, RCTs are quantitative, comparative, controlled experiments in which investigators study two or more interventions in individuals who are assigned to receive an intervention in random order. Sources:http://www.medterms.com
MedTerms is the Medical Dictionary of MedicineNet.com; Millodot: Dictionary of Optometry and Visual Science, 7th edition. © 2009 Butterworth-Heinemann; www.wickipedia.com
Numerous forms of the RCT designs exist. This approach, while the 'gold standard' for scientific and often clinical evidence, has been criticized by many as a research approach that does not account for the numerous factors involved in behavioral sciences including holistic nursing, and in integrative, complementary and alternative practices. Currently there is national emphasis on multiple approaches to research that capture the complexities of clinical practice. (see report of integrative health care conference link).
Effectiveness versus Efficacy
These two terms have recently been discussed in evaluating the usefulness and applicability of research results. Research that aims for efficacy most often utilizes the RCT. Statistically significant results of a treatment or intervention in a controlled situation is an example. Effectiveness on the other hand, attempts to evaluate treatments and approaches in real clinical situations. Effectiveness is associated with 'what works most of the time'. Such interventions may not meet consistent statistically significant results when tested on controlled situations. People reporting greater comfort after a healing touch procedure while their blood pressures may not decrease significantly is an example.
Theory of Compassion Energy
By Dorothy Dunn, PhD, RNP, FNP-BC, AHN-BC
The Theory of Compassion Energy (TCE) evolved through a caring concept clarification process and theoretical evolution via a creative synthesis utilizing Rogers' Science of Unitary Human Beings (SUHB) and Caring Science (CS). The TCE is described as caregivers (formal and/or informal) who desire to care compassionately by intentionally knowing another through patterned nurturance with authentic presence (Dunn 2012; Dunn 2009a; Dunn 2009b). The premise of the TCE is that human beings are unitary or irreducible, in mutual process with an environment that is co-extensive with the universe, participating knowingly in patterning, and ever-evolving through expanding consciousness (Rogers, 1992; Newman, 1994). While caring is a quality of participating knowingly in human-environment field patterning (Smith, 1992).
When nurses engage with the nursed (patient) from a place of caring, compassion and presence the caring moment becomes energized and focused on meeting the needs of other which in turn energizes the nurse. The dynamism associated with this dyadic encounter is linked with positive outcomes for both nurse and nursed.
I have synthesized the meaning of compassion, nurturance, energy, caring theory, and intentionality which comprise the concept of compassion energy in my research studying what keeps nurses in nursing that revealed it is the essence and critical nature of nursing the nursed (patient) through a mindful, authentic presence that exudes a therapeutic energy which transform the caring interaction. However, many nurses set a default setting to protect themselves from the experience of the nursed (patient) by distancing self from the perceived vulnerability of suffering. However, in distancing self from other the caring interactions is non-existent and becomes a technical task devoid of healing potential and places the nurse at risk for compassion fatigue.
Rogers informed us that nursing is a humanistic science dedicated to compassionate concern for maintaining and promoting health, preventing illness, and caring for and rehabilitating the sick and disabled (Rogers, 1970, p. vii). Sr. Roach asserts that compassion is a way of living born out of an awareness of one's relationship (interconnectedness) to all living creatures (Roach, 2002, p. 50). Compassion means to suffer with and involves us in going where it hurts....be weak with the weak, vulnerable with the vulnerable, powerless with the powerless....full immersion in the condition of being human (Nouwen, 1983).
Compassion Becomes the Energy of Caring.
Often compassion is used interchangeably in nursing literature as sympathy, empathy, pity, altruism; I offer the following clarification:
- Sympathy: Ability to feel for the other, (mirror neurons let us be able to feel by watching another's experience).
- Empathy: Ability to imagine and share (understand) feelings of the other.
- Pity: connotes condescension, implies separateness (feel sorry for another).
- Altruism: love for another at the expense of oneself. Love others instead of our self.
- Compassion: world's richest energy source, strength from a shared weakness and shared joy.
"A sorrow shared is sorrow halved; a joy shared is a joy doubled" ~ German Proverb
Recently, I completed a secondary analysis of compassion data that revealed that caregiver's experience compassion satisfaction that transforms the caregiving encounter. Experiencing compassion, the caregiver seeks to know and understand interconnectedness to others to alleviate suffering and celebrate joy with the care recipient.
Compassion satisfaction is about the pleasure you derive from being able to do your work. For example, you may feel like it is a pleasure to help others through what you do at work. You may feel positively about your colleagues or your ability to contribute to the work setting or even the greater good of society through your work with people who need care (Stamm, 2005).
In the caregiving experience, the goal is to use opportunities to rise to occasions and use compassion strength with courage, knowledge, and skill rather than overcome the tendency to care at a distance. By focusing on compassionate care rather than on the tasks to get done or to do, one can stave off the risk of experiencing compassion fatigue. The act of understanding and nurturing self-generated vigor as compassion energy nurses (formal) and informal caregivers will find meaning in caring for self and other with the intent to alleviate suffering or celebrate joy (Dunn, 2009b).
Theory of Integral Nursing (TIN)
Tamisha Gatewood-Henderson, MSN, MHA, RN
The Theory of Integral Nursing (TIN) is a grand theory developed by Barbara Dossey, who expanded Ken Wilber's Theory of Integral Health. A grand theory is one that focuses on broad and general areas and concepts that can be used in a variety of settings and populations. The TIN is descriptive providing definitions of concepts and suggesting a new way of looking at the world of integration. The four-quadrant model of Ken Wilber describes the reality of experience. This further helps with "understanding the dimensions of how we perceive the world, coupled with Carper's theory of how we come to know what we know" (Frisch and Potter, 2015). The TIN includes five components:
©Roman Milert|Dreamstime Stock Photos
• (a) healing, at the center
• (b) the metaparadigm of nursing which includes nurse, health, person, and environment or society, touching or healing
• (c) six patterns of knowing (includes personal, empirics, socio-political, ethics, aesthetics, and not knowing)
• (d) four quadrants (Interior/Individual, Exterior/Individual, Interior/Collective, and Exterior/Collective) adapted from Wilber's integral theory, and
• (e) AQAL (all quadrants, all levels, all lines, and states, and all types) The concepts of AQAL (all quadrants, all levels) use the I, We, It, and Its which are elements of awareness.
Florence Nightingale was considered the first integralist - or "health diplomat". This person focuses on the individual and the collective, the inner and outer, and human and nonhuman concerns (Dossey, 2008). Not only was Nightingale concerned with the basic needs of humans, but she was also concerned with all aspects of the environment, including a connection with the Divine as an awareness.
This TIN has importance in holistic nursing because it is meant to address broad areas of nursing and expand our notions of who and what we are.
Dossey, B. (January/March 2008). Theory of integral nursing. Advances in nursing science. 31(1) E52-E73.
Frisch, N. & Potter, P. (2016). Nursing theory in holistic nursing practice. In B. M. Dossey, L. Keegan, C. Barrere, M.A. Blasko Helming, D. Shields, & K. Avino (Eds.), Holistic Nursing - A Handbook for Practice. (pp 111-120). Burlington, MA: Jones and Bartlett Learning
Threats to Internal and External Validity
Deborah Kramer, Ed.D., RN,CPNP, FNP
In choosing a research design the researcher wants to be free of bias from any threats to the validity-accuracy of the outcome of the research study. What are some risks that can affect the outcome of the study?
When conducting quantitative research the researcher is testing whether the independent variable is truly what is making the difference in the dependent variable. And that the results are generalizable to the population selected and replicated in other populations or environments.
To establish internal validity, knowing the dependent variable was impacted by the independent variable the researcher considers possible threats that could have influenced the outcome of the study. These threats arehistory, maturation, mortality, testing, instrumentation, and selection bias.
History - During the time that the research study is taking place another event can impact the outcome of the study. This event is the influence that results in the change of the dependent variable and not the independent variable. For instance a study is being conducted on eating a low fat, healthy diet in reducing heart disease for patients 20-25 years old with elevated cholesterol. The intervention is an educational video program, that is the independent variable and lower cholesterol levels is the dependent variable. During the time this study was taking place the most famous Rock Singer has a heart attack, and it is on the news and in the newspapers that he had a very high fat, unhealthy diet. It may be that event that influenced the study participants to change their diet and not the educational video.
Maturation - Participants of the study change over time during the course of the study. It can be physical (growth, healing, fatigue), cognitive (learning new information or skills), or developmental. These changes can be the influences on the dependent variable.
Mortality - Participants leave the study before its completion. If this leaves the groups being studied not equal, this may impact the outcome of the study.
Testing - Participants that are tested prior to the study are sensitized to the information or attitudes that are being studied. They may have altered scores or changed attitudes due to this rather than the independent variable.
Instrumentation - The actual instrument can change during the course of the study. If the instrument is a scale it needs to be tested and calibrated. The instrument can also be individuals scoring or rating an activity. Such as the judges at the Olympics scoring a triple toe loop jump in ice skating. Judges can go through training at the beginning of the study and and either become more proficient than other judges during the course of the study or less proficient as the time from the training elapses. The change in the instrumentation can impact the outcome of the study.
Selection Bias - This occurs most often when participants in the study are not randomly selected. There can be different characteristics in those who are hand selected or volunteer to participate in the study than those in the population that were not included. The differences in the groups can impact the dependent variable and it may not be due to the independent variable.
Research designs are selected to help minimize these threats to internal validity. An example is the Solomon four group. There are 2 experimental groups and 2 control groups. One of each group takes the pretest and the other does not. All groups take the post test. The results are analyzed to see if knowledge from the pretest impacted the outcome in the groups.
The ability to generalize the outcomes of the study isexternal validity. There are 3 threats to external validity, reactivity effects, selection effects, and measurement effects.
Reactivity - is also known as the Hawthorne effect. Participants respond because they are being studied and not due to the independent variable.
Selection - the sample from the population was selected may not be representative of the population being studied. Therefore, it is not generalizable to the population.
Measurement - the participants of the study are sensitized by taking a pretest to what is being measured in the study. This may impact the participants' outcome limiting the researchers' ability to generalize the outcomes of the study to the population being studied.
In selecting the research design, the researcher attempts to minimize the possibility of threats to external validity.
Polit, D., & Beck, C.T. (2012). Nursing research: Generating and assessing evidence for nursing practice, 9th Ed. Philadelphia, PA: Lippincott Williams and Wilkins.
Burns,N.& Grove,S.K. (2009). The practice of nursing research: Appraisal, Synthesis, and generation of evidence, 6th edition. St Louis, Missouri: Saunders.
LoBiondo-Wood,G & Haber,J. (2010). Nursing Research: Methods and critical appraisal for evidenced based practice, 7th edition. St Louis ,Missouri: Mosby
By: Sue Roe, DPA, MS, BSN, RN
You decide to "triangulate" in your next research study. What does this mean? How is it accomplished? What are the benefits and limitations?
Triangulation is an approach where two or more data sources are combined (such as methods, observers, or theories) in a single study. The presumption is that using a single data source does not provide enough information about a phenomena. The interconnected components of the totality of human beings, mind-body-spirit, recognized by holistic nurses, is triangulation at its best! These important "data sources" are collected in practice daily and used for assessment, action, intervention, and/or evaluation.
The utility of data triangulation is to add explanatory power, richness, and complexity to results. By its nature, triangulation can give more power to data; offer confidence in findings (reliability); provide cross-verification and corroboration; and it can facilitate validation which is essential in qualitative studies (Silverman, 2006).
"The combination of multiple methodological practices...adds rigour, breadth, complexity, richness and depth to an inquiry." Flick (2002)
A major concern or limitation in using triangulation is the possibility that findings from multiple sources will be divergent or contradictory. Mathison (1988) contends that there are three outcomes when using triangulation: convergence, inconsistency, and contradiction. Some also believe that triangulation is a way to compensate for a weakness in a primary method so a second or even third methodological approach is used. While these limitations are real, in research, remember, uncovering findings that are unintended or surprising allows the researcher to open up new thinking, garner deeper insights, and/or probe further to determine if data are flawed.
Triangulation is often considered the rationale for multi or mixed method research (use of a combination of quantitative and qualitative measures) and it is also thought of as a methodological approach.
A researcher can use within-method triangulation such as incorporating differing scales to measure a concept in a single questionnaire or the more robust between or across-method triangulation. In between or across-method triangulation a follow up interview might be conducted with some or all of the respondents after the employment of a questionnaire to that same group (Denzin, 1970; 1978).
As an early advocate for the use of triangulation, Denzin (1970) suggested four types of triangulation.
Investigator triangulation: Using multiple researchers or observers in a study. The goal is to avoid selective perception or blind spots by seeking different ways of "seeing" data.
Methods triangulation: Using a variety of different methods to complement data collection. These methods may include interviews, observations, questionnaires, and documents.
Source or Data triangulation: Using different aspects of the same method such as several samplings at different points in time.
Theory/Perspective triangulation: Using more than one theory to examine and interpret the data.
Triangulation, as a tool, has a long history (it was used to navigate ships before instrumentation). During the 1950s it emerged as an approach to consider when conducting research. Today, triangulation is an important methodological contributor to qualitative studies (interpretive research).
Triangulation has much to offer holistic nursing research as multiple data sources are very useful in garnering the insights needed to answer the many qualitative questions posed. Some examples of triangulation in holistic nursing research include a pilot study on Reiki for self-care of nurses and healthcare providers where methods were triangulated using a self-report caring scale and interviews (Brathovde, 2006). In another study, three nursing concepts were triangulated (art of nursing, presence, and caring) to determine qualitative convergence (Finfgeld-Connett, 2008).
Collecting and merging differing viewpoints is a key outcome for using triangulation in research. These viewpoints, when interpreted, provide that deeper understanding sought for topics or questions under study.
To find a useful listing of sources in triangulation consult:
Sage Research Methods Retrieve
By Dr. Sue Roe DPA, MS, BSN, RN
There are many important terms holistic nurse researchers must be comfortable with. Validity is one of those terms. From a research design perspective, having a valid study means that it accurately measures a specific concept or concepts the researcher is attempting to measure. For example, one study might be determining whether aromatherapy coupled with massage decreases mental fatigue.
In designing the study the researcher will first determine face validity. Face validity is how a measure or procedure appears "on the face of it". For instance, does this seem to be a worthwhile study? Is it well designed? Are the methods for collecting data reasonable?
The researcher will also need to ensure external and internal validity. Having external validity means the study results can be generalized across other populations, settings, outcomes, times, and treatments. This can apply to quantitative and qualitative research designs.
Internal validity deals with the accuracy of results. Is there sufficient evidence to substantiate the results? The focus is on controlling for possible confounding variables so the only factor(s) which affect the dependent variable is the independent variable. The question posed in internal validity is, "might there be an alternative reason for what was observed and/or for the results of the study?" There can be threats to internal validity. These might be bias or effects of the testing instruments used.
Internal validity offers confidence. In our example, this researcher will have high internal validity if it is found that aromatherapy coupled with massage decreased mental fatigue rather than confounding variables such as changes in nutrition or sleep habits.
Validity extends to statistics, and in particular, the validity of testing instruments. Here validity has a similar purpose - does a testing instrument measure what it claims to measure? Taking our example one step further, let's say this researcher decides she will use a fatigue scale to measure the sample's perception before and after the treatment of massage and aromatherapy.She will need to understand three types of test validity:
- Content Validity: Content validity is the extent to which a testing instrument reflects the specific and intended scope of content. For example, did the scale selected by the researcher cover all possible dimensions of fatigue?
- Criterion Validity: Criterion validity, also referred to instrument validity demonstrates accuracy by comparing it with another measure or procedure deemed valid. There are two types of criterion validity: Concurrent validity is accomplished when a testing instrument, such as the one selected by our researcher, is benchmarked with another fatigue scale measuring the same concepts and the result is a high correlation. Predictive Validity occurs when results from a testing instrument are able to predict future designated outcomes or results.
- Construct Validity: Construct validity seeks agreement between a theoretical concept and a specific measuring device or procedure. For example, in our study, the fatigue scale selected (or developed) should measure fatigue as it was defined for this study. It cannot measure other concepts such as sleep deprivation or stress. Construct validity has two sub-categories: Convergent validity and discriminate validity. Convergent validity means an agreement that the concepts expected to be related are in fact related. Discriminate validity is the reverse. There should be no relationship among concepts which theoretically should not be related.
A Few Sources:
Christensen, L.B., Johnson, R.B., & Turner, L.A. (2013). Research methods, design, and analysis. New York, NY: Pearson.
Creswell. J.W. (2013). Research design: Qualitative, quantitative, and mixed methods
approaches. Thousand Oaks, CA: Sage.
Garson, G.D. (2013). Validity and reliability. Blue Book Series. Statistical Associates Publishers.
Houser, J. (2013). Nursing research: Reading, using, and creating evidence. Burlington, MA: Jones & Bartlett Learning.
Polit, D.F. & Beck, C.T. (2011). Nursing research: Generating and assessing evidence for nursing practice. Philadelphia, PA: Lippincott Williams & Wilkins.
Variables in Quantitative Research
Pamela Crary, PhD, RN
A variable is a measurable characteristic that varies among the subjects being studied. As the definition implies, the characteristic or phenomenon under study varies in some way.
Independent Variable: a stimulus or activity that is manipulated or varied by the researcher to create an effect on the dependent variable. It is helpful to remember the independent variable as the treatment or intervention.
Dependent Variable: the outcome or response that the researcher wants to predict or explain. Changes in the dependent variable are presumed to be caused by the independent variable. It is helpful to remember the dependent variable as the outcome being measured.
Example: "Cancer patients who receive music therapy have less perceived pain than cancer patients not receiving music therapy."
Music Therapy is the independent variable. Pain is the dependent variable.
Descriptive and correlational quantitative studies involve the investigation of research variables.
Research Variable: quality, property or characteristic identified in the research purpose and objectives or questions that are observed or measured in a study. Research variables are used when the intent of the study is to observe or measure variables as they exist in a natural setting without implementation of a treatment. Thus no independent variables are manipulated, and no cause-and-effect is examined.
Extraneous Variable: a variable that exists in all studies and can affect the measurement of study variables and the relationships among these variables. Researchers try to control for extraneous variables so they do not interfere with measurement and outcomes. One way is using inclusion and exclusion criteria when sampling.
Confounding Variable: a type of extraneous variable that is not recognized until the study is in process, or is recognized before the study is initiated but cannot be controlled. Confounding variables weaken a study design and hinder interpretation of outcomes unless they are able to be controlled statistically during analysis.
Environmental Variable: a type of extraneous variable composing the setting in which the study is conducted. Examples of these include climate, family, healthcare system.
Demographic Variable: attributes of subjects that are collected to describe the sample such as age, gender, education, ethnicity, income, diagnosis, etc.
Conceptual Definition: provides the theoretical meaning of a variable.
Example: Stress is defined by Lazarus and Folkman (1985) as a perceived state when demands exceed resources to manage those demands.
Operational Definition: provides the measurement process for the variable.
Example: Stress will be operationally defined using the Perceived Stress Scale.
Burns, N., and Grove, S. (2011) Understanding nursing research, 5th Edition, Maryland Heights, MO: Elsevier Saunders.
Fain, J. (2013) Reading, understanding and applying nursing research, 4th Edition, Philadelphia, PA: F.A. Davis.
What Are the Four Levels of Measurement?
By Pamela Crary, PhD, RN
When collecting data in a quantitatively designed study, variables are conceptually defined with words similar to a dictionary definition. They are also operationally defined through ways of measurement using numbers. There are different levels of measurement depending on the research question being asked and the types of statistical analyses planned. Collecting the correct levels of measurement is necessary to assure that appropriate analyses can be done.
There are four levels of measurement; Nominal, Ordinal, Interval, and Ratio. One level of measurement is not necessarily better than another.
What is nominal level of measurement?
The nominal level of measurement is the most primitive or lowest level of classifying information. Nominal variables include categories of people, events, and other phenomena that are named, are exhaustive in nature, and are mutually exclusive. These categories are discrete and non-continuous.
Example: gender - Male or Female can be scored with 1 for Male and 2 for Female; likewise a patients' blood type could be categorized as 1=AB, 2=A, 3=B, 4=O.
No one category is more or less than another; they are simply categorized with a number for statistical analyses. They are not manipulated mathematically.
What is ordinal level of measurement?
The ordinal level of measurement is second in terms of its refinement as a means of classifying information. Ordinal implies that the values of variables can be rank-ordered from highest to lowest. Data are measured on an ordinal scale and subjects are ranked from lowest to highest and from most to least.
For example, household income: 1=$0-$4999, 2=$5000-$9999, 3=$10000-$19999, 4=$20000-$29999, and 5=$30000-$49999.
Ordinal data are not manipulated mathematically and the distance or interval between data is not always equal.
What is interval level of measurement?
Interval level of measurement is quantitative in nature. Interval level of measurement refers to the third level of measurement in relation to complexity of statistical techniques that can be used to analyze data. Variables within this level of measurement are assessed incrementally, and the increments are equal. Many nursing, social and psychological science studies measure data using tools or instruments that consist of a Likert type scale such as the one below.
For example: Respondents are asked to select from a series of statements that reflect agreement or disagreement on a 5-point scale. 1=strongly agree, 2=agree, 3=undecided, 4=disagree and 5=strongly disagree.
The individual units are equally distant from one point to the other. Interval data do not have an absolute zero.
What is ratio level of measurement?
Ratio level of measurement is characterized by variables that are assessed incrementally with equal distances between the increments and a scale that has an absolute zero. Ratio variables exhibit the characteristics of ordinal and interval measurement and can also be compared by describing it as two or three times another number or as one-third, one-quarter, and so on. Variables like time, length, and weight are ratio scales but can also be measured using nominal or ordinal scale. The mathematical properties of interval and ratio scales are very similar, so the statistical procedures are common for both of the scales.
Ratio level data meets all the rules of other forms of measure; it includes mutually exclusive categories, exhaustive categories, rank ordering, equal spacing between intervals, and a continuum of values. Ratio level measurement also includes a value of zero