The philosophical foundations of the modern scientific research method can be traced back three centuries. As science historian Stephen Toulmin has noted [1], it was then that the Renaissance of Montaigne and Shakespeare, noteworthy for its celebration of humanity in all of its diversity, gave way to the demoralizing and enervating Thirty Years' War, a conflict fueled by religious and ethnic intolerance. In prostrate and exhausted postwar Europe, Descartes and Newton arose, promulgating rational thought as a mechanism for progress and a way to overcome humankind's foibles. Toulmin finds the origins of modern science in the latter part of the 17th century, in transitions from oral to written communication and from rhetoric to documentation as well as in the search for universal principles (of physics and ethics, for example) that would apply in all locales and times. This emphasis on the objective record and underlying "laws" underpinned the advance of modern science as, during the ensuing three centuries, science developed a progressively deeper understanding of the natural world, causal mechanisms, the rationale for fundamental interventions, and ways to develop and use the technologies born of these discoveries.

From a philosophical perspective, I contend that Newton's era and the Renaissance established two permanently important "world views," from which the traditions and methods of inquiry have proceeded. The first, the foundation of our dominant paradigm of modern science, posits an objective reality characterized by stable, predictable, and commensurable phenomena that operate through the laws of cause and effect. This reality is subject to universal laws and forces, although the expression of these laws and forces is affected by particular circumstances. The second "world view" emphasizes a subjective reality that is characterized by complexity, apparently infinite variation, and incommensurability without perturbation. The laws of cause and effect operate, but always within a unique set of circumstances determined by multiple factors. Appreciating the particulars of those circumstances in depth may constitute its own, sufficient aim of study.

In these two world views, the identity of the scientist differs. In the first, the scientist is an external observer; pursues work that is theory- and hypothesis-driven; emphasizes precision, reliability, and predictability; and seeks to make generalizable observations. In the second, the scientist is a participant as well as an observer, part of the phenomena of interest, and grounded in nature. In evaluating his or her work, this scientist emphasizes salience, richness, verisimilitude, trustworthiness, and the generative effect of findings on the work and understanding of others.

These two world views and the position of scientists within them roughly characterize the "quantitative" versus "qualitative" divide that is now usually posited as separating two communities of scientists and their methods. Portrayed in this manner, the schism between these traditions seems profound, unlikely to be bridged. In truth, the perspectives, methodologic approaches, and communities of scientists are different but interdependent, and both are apt to be instrumental in the progress of science at different points during the acquisition and use of knowledge. Rather than emphasizing the incompatibilities of "qualitative" and "quantitative" research traditions and risking an intellectual Thirty Years' War, I believe that it is appropriate to highlight the complementary natures of biomedical research (quantitative), epidemiology (quantitative), medical anthropology (largely qualitative), medical sociology (largely quantitative today but with an older qualitative tradition), medical humanities (qualitative), medical ethics (qualitative and quantitative), and other forms of inquiry appropriate to today's health problems.

Viewed from certain perspectives, quantitative and qualitative research share the values of science. As Lincoln and Guba have pointed out [2], both traditions emphasize truth, consistency, applicability, and neutrality while taking different procedural approaches to assuring the quality of their processes and products. In search of "truth," quantitative researchers emphasize standard measures, replicable findings, comparison to accepted gold standards, minimization of bias, and successful prediction. Qualitative researchers, in pursuit of "trustworthy results," emphasize the need for prolonged immersion in the phenomena of interest; study from several different points of view; detailed description; discussion of findings with experienced members of the relevant community; "field journals" that record procedures; active debate among members of the scientific community; and systematic exploration of the investigator's point of view before, during, and after research. In addition to sharing values, qualitative and quantitative researchers have similar criteria with which to evaluate the quality of their research publications [3].

From an "ecumenical" perspective that recognizes the need for both qualitative and quantitative research, I do not see that one method is generally superior to the other. Instead, I find it suitable to ask when, and under what circumstances, one approach is more appropriate than the other. A form of qualitative research, for example, would be desirable when asking questions about meaning, human value, or the understanding of social processes not previously explored or when searching for new theory grounded in the perceptions and traditions of social groups. Questions of magnitude, rate, incidence, or prevalence, on the other hand, generally yield only to quantitative methods.

In this issue, Kravitz and colleagues [4] report on their efforts to characterize patients' expectations for health care. Their aim of measuring the prevalence of expectations was well served by a quantitative survey. However, another of their aims—to understand the sources of patients' expectations—required original descriptive work grounded in patient reports. The product of their research is strengthened by the availability of data of both types and provides a richer understanding of how often and on what basis patients may judge care to be unsatisfactory. For the purposes of Kravitz and colleagues, then, a dual approach (quantitative and qualitative) was most appropriate. If this research team should next design an intervention intended to maximize the fulfillment of expectations and choose to implement it in an experimental study, the question will again arise, Would both quantitative and qualitative methods be appropriate? By engaging in both methods of inquiry in their present study, Kravitz and colleagues have not permanently camped in one domain of science or the other. Faced with similar circumstances, other investigators would do well to follow their lead. In science, almost nothing is as unintelligent as always applying one method to every kind of problem and question.