Establishing the internal and external validity of experimental studies

College of Pharmacy, The University of Arizona, Tucson 85721-0207, USA.
American Journal of Health-System Pharmacy (Impact Factor: 1.88). 12/2001; 58(22):2173-81; quiz 2182-3.
Source: PubMed


The information needed to determine the internal and external validity of an experimental study is discussed. Internal validity is the degree to which a study establishes the cause-and-effect relationship between the treatment and the observed outcome. Establishing the internal validity of a study is based on a logical process. For a research report, the logical framework is provided by the report's structure. The methods section describes what procedures were followed to minimize threats to internal validity, the results section reports the relevant data, and the discussion section assesses the influence of bias. Eight threats to internal validity have been defined: history, maturation, testing, instrumentation, regression, selection, experimental mortality, and an interaction of threats. A cognitive map may be used to guide investigators when addressing validity in a research report. The map is based on the premise that information in the report evolves from one section to the next to provide a complete logical description of each internal-validity problem. The map addresses experimental mortality, randomization, blinding, placebo effects, and adherence to the study protocol. Threats to internal validity may be a source of extraneous variance when the findings are not significant. External validity is addressed by delineating inclusion and exclusion criteria, describing subjects in terms of relevant variables, and assessing generalizability. By using a cognitive map, investigators reporting an experimental study can systematically address internal and external validity so that the effects of the treatment are accurately portrayed and generalization of the findings is appropriate.

74 Reads
  • [Show abstract] [Hide abstract]
    ABSTRACT: Although it is generally appropriate for a healthy adult to consume 2000 to 2500 ml per day, the literature does not address evaluating any standard. The objective here was to develop a weight-based hydration management intervention and evaluate the impact of this on the incidence of acute confusion (AC) using an N = 98. The intervention consisted of a fluid intake goal based on 100 ml per kg for the first 10 kg, 50 ml/kg for the next 10 kg, and 15 ml for the remaining body weight. The treatment group received instruction and assistance on the fluid goal and the control group received routine care. Measurements included serum electrolytes, bioimpedance analysis, urinalysis, Mini-Mental State Exam, and the NEECHAM. There was no difference in the incidence of AC between treatment and controls, but those individuals with > or = 90% compliance demonstrated higher ECF volumes and also lower urine leukocyte counts.
    Western Journal of Nursing Research 04/2003; 25(3):251-66; discussion 267-73. DOI:10.1177/0193945902250409 · 1.03 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The radiotherapeutic management of painful bone metastases is controversial, with several institutional and national guidelines advocating use of single-fraction radiotherapy. We aimed to determine patient choice of fractionation schedule after involvement in the decision-making process by use of a decision board. Advantages and disadvantages of two fractionation schedules (24 Gy in six fractions v 8 Gy in one fraction) used in the randomized Dutch Bone Metastasis Study were discussed with patients using a decision board. Patients were asked to choose a fractionation schedule, to give reasons for their choice, and to indicate level of satisfaction with being involved in decision making. Sixty-two patients were entered. Eighty-five percent (95% confidence interval, 74% to 93%) chose 24 Gy in six fractions over 8 Gy in one fraction (P <.0005). Variables including age, sex, performance status, tumor type, pain score, and paying class were not significantly related to patient choice. Multiple fractionation was chosen for lower re-treatment rates (92%) and fewer fractures (32%). Single-fraction treatment was chosen for cost (11%) and convenience (89%). Eighty-four percent of patients expressed positive opinions about being involved in the decision-making process. Decision board instruments are feasible and acceptable in an Asian population. The vast majority of patients preferred 24 Gy fractionated radiotherapy compared with a single fraction of 8 Gy. These results indicate the need for further research in this important area and serve to remind both clinicians and national or institutional policy makers of the importance of individual patient preference in treatment decision making.
    Journal of Clinical Oncology 06/2003; 21(11):2156-62. DOI:10.1200/JCO.2003.10.112 · 18.43 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: To examine the concepts of external validity and generalizability, and explore strategies to strengthen generalizability of research findings, because of increasing demands for knowledge utilization in an evidence-based practice environment. The concepts of external validity and generalizability are examined, considering theoretical aspects of external validity and conflicting demands for internal validity in research designs. Methodological approaches for controlling threats to external validity and strategies to enhance external validity and generalizability of findings are discussed. Generalizability of findings is not assured even if internal validity of a research study is addressed effectively through design. Strict controls to ensure internal validity can compromise generalizability. Researchers can and should use a variety of strategies to address issues of external validity and enhance generalizability of findings. Enhanced external validity and assessment of generalizability of findings can facilitate more appropriate use of research findings.
    Journal of Nursing Scholarship 02/2004; 36(1):16-22. DOI:10.1111/j.1547-5069.2004.04006.x · 1.64 Impact Factor
Show more

Similar Publications