Various groups have established levels or hierarchies of evidence, usually based upon scientific merit in an empirical model. "The interpretations and decisions that nurses make as they review the research literature preparatory to developing policies, protocols, and guidelines, etc. are thus crucial and need to be based on sound procedures and explicit criteria. The rigor essential to such reviews is also important because of the aura of credibility that accompanies recommendations that purportedly are based on research." (Stetler, et al. 1998, p.195)

After the sources of evidence (clinician experience, patient preferences, and scientific findings) have been found, each source must be evaluated for its strength.
Strength is evaluated by asking: "how good is the evidence (meaning how valid and relevant) for this particular treatment-related clinical situation?"
To answer the "how good" question, researchers have created typologies of hierarchical evidence strength.

The ONS Levels of Evidence (Ropka & Spencer-Cisek, 2001), developed by an ONS project team, is a typology that is recognized as a good method for evaluating sources of evidence. A number of published approaches to categorizing levels of evidence were reviewed including the Agency for Health Care Policy Research (1994), U.S. Preventive Services Task Force (1996) Guide to Clinical Preventive Services (3rd edition), and work by Gray (1997), and Hadorn et al. (1996). The table below is an adaptation of the work by Hadorn et al. 1996.

In the ONS model, evidence is ranked from Levels I to III with subcategories. ONS Level I is evidence from meta-analysis of multiple, well-designed controlled studies; experimental studies; and well designed nonrandomized studies. ONS Level II evidence is obtained from systematic reviews of non-experimental studies, and ONS Level III is evidence derived from qualitative studies, case reports, and expert opinion. The strength of the evidence is strongest for Level I studies and weakest for Level III studies."