To consider the results of research valuable and applicable, they have to be reliable, and valid. Salazar (Pp.1-2) defined reliability of results as a reference term to repeatability that is if the study were to be conducted for a second time in the same circumstances, it would produce the same results. Reliability is a term describing a research outcome that applies also to study measures. Validity points to the trustworthiness of results (credibility) or are the results true as judging the assumption, the faster students in solving an exam are those with the highest IQ. Van Cantfort (Pp.1-9) suggested there are two aspects of validity, internal and external. Internal validity is where researchers use research procedures and instruments to measure what instruments and procedures are supposed to measure. In simple words did procedure D as an experimental manipulation instrument produces a significant difference? External validity means that the results of the present research can be applied or generalized to other similar conditions (Salazar, Pp.1-2). Salazar (Pp.1-2) described the relationship between reliability and validity as reliability is necessary to achieve validity; however, alone it is not enough justification for validity.
The department of Psychology at Georgetown University explained the factors that affect the internal validity of research results. These factors (threats) are Maturation (changes due to aging or development between or within groups), history (outside proceedings during the course of research). Third, is instrumentation valid during the course of the research, fourth is testing that is experience gained by the research team during the time course influenced results. Next, are the participants selected in each grouping the same to respond to the instruments and procedures homogenously (selection)? Sixth is the Hawthorne effect that is when individuals of the treatment group whose response represents the dependent variable change over time so they act differently. Seventh is contamination when the comparison group affects or becomes affected by the treatment group. The department of psychology summarized threats to external validity (factors that prevent generalization) as selection bias (nonrepresentative sample), reactive effects of experimental arrangements, or test or pretest sensitization. Multiple treatment interference is considered also a threat to external validity.
In some cases, a specific instrument assesses the average causal effect of the treatment group. This produces a diversity (heterogeneity) of casual effects, a situation that alerts to the trade-off between internal and external validity. Reading from the threats to validity previously explained, internal validity requires a valid instrument, and no contamination of treatment and control groups. On the other hand, external validity needs the causal effect can be generalized to a population beyond that included in the research, which is already affected by the defects of the internal validity instrument (Moffitt, Pp.91-108). Moffitt (Pp. 91-108) suggested that caution and proper research design are the keywords. Moffitt also suggested that in studies where causal effects are unknown, knowledge of other causal effects and evidence search from other studies may enlighten the path to estimate the causal effect of interest.
Works Cited
Department of Psychology. Validity and Reliability. Dept. home page. Georgetown University. Web. 2008.
Moffitt, Robert. “Remarks on the Analysis of Causal Relationships in Population Research.” Demography vol 42 (1) 2005. p. 91-108.
Salazar, Heather. “Introduction: Summary.” University of California, Santa Barbara. 2006. University of California, Santa Barbara. 2007. Web.
Van Cantfort, Thomas. “Validity.” 2008 Courses: On Line Reprints. 2006. Fayetteville State University. Web.