Assessing Validity
Most validation studies begin by referring to content validity. Each health measurement represents a sampling of questions from a larger number that could have been included. Similarly, the selection of a particular instrument is a choice among alternatives, and the score obtained at the end of this multistage sampling process is of interest to the extent that it is representative of the universe of relevant questions that could have been asked. Content validity refers to comprehensiveness or to how adequately the questions selected cover the themes that were specified in the conceptual definition of its scope. For example, in a patient satisfaction scale, do all the items appear relevant to the concept being measured, and are all aspects of satisfaction covered? If not, invalid conclusions may be drawn. Feinstein has proposed the notion of sensibility, which includes, but slightly extends, the idea of content validity (108). Sensibility refers to the clinical appropriateness of the measure: are its design, content, and ease of use appropriate to the measurement task? Feinstein offered a checklist of 21 attributes to be used in judging sensibility. Indeed, content validity is seldom tested formally; instead, the “face validity” or clinical credibility of a measure is commonly inferred from the comments of experts who review its clarity and completeness. A common procedure is to ask patients and experts in the field to critically review the content of the scale. Alternatively, more formal focus groups and in-depth interviews may be arranged to explore whether the questionnaire is covering all aspects of the topic relevant to patients. Cognitive interviews involve having respondents verbalize their reactions to each question as they answer them to indicate how the questions are perceived by respondents (85). Occasionally, tests of linguistic clarity are used to indicate whether the phrasing of the questions is clear (109). It is difficult, perhaps even impossible, to prove formally that the items chosen are representative of all relevant items (110).
การประเมินมีผลบังคับใช้Most validation studies begin by referring to content validity. Each health measurement represents a sampling of questions from a larger number that could have been included. Similarly, the selection of a particular instrument is a choice among alternatives, and the score obtained at the end of this multistage sampling process is of interest to the extent that it is representative of the universe of relevant questions that could have been asked. Content validity refers to comprehensiveness or to how adequately the questions selected cover the themes that were specified in the conceptual definition of its scope. For example, in a patient satisfaction scale, do all the items appear relevant to the concept being measured, and are all aspects of satisfaction covered? If not, invalid conclusions may be drawn. Feinstein has proposed the notion of sensibility, which includes, but slightly extends, the idea of content validity (108). Sensibility refers to the clinical appropriateness of the measure: are its design, content, and ease of use appropriate to the measurement task? Feinstein offered a checklist of 21 attributes to be used in judging sensibility. Indeed, content validity is seldom tested formally; instead, the “face validity” or clinical credibility of a measure is commonly inferred from the comments of experts who review its clarity and completeness. A common procedure is to ask patients and experts in the field to critically review the content of the scale. Alternatively, more formal focus groups and in-depth interviews may be arranged to explore whether the questionnaire is covering all aspects of the topic relevant to patients. Cognitive interviews involve having respondents verbalize their reactions to each question as they answer them to indicate how the questions are perceived by respondents (85). Occasionally, tests of linguistic clarity are used to indicate whether the phrasing of the questions is clear (109). It is difficult, perhaps even impossible, to prove formally that the items chosen are representative of all relevant items (110).
การแปล กรุณารอสักครู่..
