I recently read Brilowski and Wendler's (2005) concept analysis of caring. As I had conducted a concept analysis of caring using Rodgers (2000) evolutionary framework, I read this JAN paper with interest. From my work using Rodgers’ model to analyse the concept of caring (and actually working with her on this project), there were several areas of departure between my understanding of Rodgers’ philosophy and that of these authors.

When authors assert that ‘there are no published concept analyses of caring in the nursing literature’ (p. 648), there is always the possibility that they simply may not have found a published report, which in this case may be what happened. My own paper accounts for at least three published accounts of concept analysis of caring using Rodgers’ evolutionary model (Sadler 1995, 1997, 2000).

Rodgers’ method begins with first identifying the domain to be studied based on the research question. Rodgers described the total indexed literature identified by the search as the population (p. 88). My understanding of Rodgers is that the researcher then selects the 20% sample randomly from this population, leading to ‘increase[d] likelihood of a credible sample’ (p. 88). A 20% random sample of the professional nursing publications (the domain of my study) between 1988 and 1992, using caring as the keyword, resulted in 341 articles/books/dissertations (Sadler 1995, 1997, 2000). Brilowski and Wendler broadened their time frame to 15 years of professional nursing literature (the reader must assume this as the domain as it was not clearly defined in the article) and, therefore, one would expect a larger rather than smaller sample. Brilowski and Wendler described identifying the population as 6000 articles, but then limited their attention to ‘those written in English and appearing in core journals’, and then further limited articles focusing on what was identified in my analysis as an attribute of caring: namely, the instrumental work of caring. Through an iterative process of eliminating articles, the population was reduced from 6000 to 283 articles. From the identified 283 articles, the authors sampled 25% and then eliminated seven more articles. It seems that the authors may have so narrowed the population that they were unable to fully capture the attributes and consequences of the concept of caring, as well as describe the use, significance, and application of the concept. Furthermore, the authors’ statement that they had difficulty identifying antecedents may be related to this lack of a representative sample from which to draw valid conclusions, as well as not addressing references ‘the actual situations where the concept is being applied’ (Rodgers 2000, p. 92).

When the authors provided dictionary definitions of the attributes, rather than develop a definition from the data, they made another clear departure from my own understanding of Rodgers’ model. ‘The attributes of the concept constitute a real definition, as opposed to a nominal or dictionary definition that merely substitutes one synonymous expression for another’ (Rodgers 1989c, as cited in Rodgers 2000, p. 91). While one may look for similarities between the two studies, the differing approaches to sample selection and analysis limit this comparison.

Rodgers described data analysis as ‘review[ing] the literature to identify data relevant to the attributes of the concept, its contextual features, ‘…surrogate terms and related concepts, along with data pertaining to applications of the concept (references)’ (p. 91). My understanding of this process is to read and reread the articles discovering the data within the article. In contrast to this process, the authors reported that ‘the articles [emphasis added] were grouped according to emerging themes’ (p. 643). It would seem that categorizing the entire article as data would obscure the identification of the attributes, references, antecedents, consequences and surrogate terms within the article.

While not purely ‘constructed’, the authors provide personal experience as exemplars, which raises questions. How accurate was the memory? Were portions constructed? Rodgers (2000) stated, ‘…exemplars should be identified rather than constructed by the investigator’ (p. 96).

In presentating the limitations of their study, Brilowski and Wendler noted that seminal works were omitted by the ‘sampling method’. Rodgers provided instructions for this very situation. Rodgers stated, ‘works considered to be ‘‘landmark’’ or ‘‘classic’’ and special searches to identify books, book reviews, dissertations, or theses may be included’ (p. 90).

Rodgers presented a clear philosophical basis to her approach that supports a rigorous concept analysis that will contribute to nursing knowledge and to solving the conceptual problems in the discipline. The areas of departure from the philosophy of evolutionary concept development and the methods proposed by Rodgers that support this philosophy are the basis of my questions about Brilowski and Wendler's study.


  1. Top of page
  2. References
  • Brilowski G.A. & Wendler M. (2005) An evolutionary concept of caring. Journal of Advanced Nursing 50(6), 641650.
  • Rodgers B.L. (2000) Concept analysis: an evolutionary view. In Concept Development in Nursing: Foundations, Techniques and Applications, 2nd edn (RodgersB.L. & KnaflK., eds), Saunders, Philadelphia, PA, pp. 77102.
  • Sadler J. (1995) Analysis and Observation of the Concept of Caring in Nursing. PhD Thesis, The University of Wisconsin-Milwaukee, Milwaukee, WI, 272p. UMI order no. PUZ9528682.
  • Sadler J. (1997) Defining caring in professional nursing: a triangulated study. International Journal for Human Caring 1(3), 1221.
  • Sadler J. (2000) A multiphase approach to concept analysis and development. In Concept Development in Nursing, 2nd edn (RodgersB.L. & KnaflK.A., eds), Saunders, Philadelphia, PA, pp. 251283.