Whenevera measuring device is used in any experiment, the validity andreliability of the test is vital. Validity denotes the level a testis precisely representing what it is intended to reflect (Graham,2003).Thus, validity means credibility and the degree of believability.
For instance, the test question 2+2 =4 is valid as it truly measuresa student’s capability to perform basic addition in mathematics.But the same test turns out to be less meaningful as a measurement ofcomplex sum because it does not address all issues of knowledge thatare required in solving and understanding such a problem.
Furthermore,it becomes invalid when it is applied to measure the knowledge ofNorth America’s history as the capability to sum up two numbers hasnothing to do with it. Most people agree that 2+2=4 represents abasic summation, but can this question also apply to the construct ofintelligence? Other constructs are depression, motivation, anger, andany human trait or emotion(Franzen, 2000). Thus, construct validity refers to a test that measure any humantrait or emotion accurately. Let’s examine three types of validitythat are important in psychology concurrent, content and predictivevalidity.
Concurrentvalidity enables you to illustrate that an experiment is valid bycomparing it with an already proven measure. For instance, if a testfor adult intelligence reflected a high positive relationship withthe Wechsler Adult Intelligence Scale, it would mean that there is aconcurrent validity because the scale is a generally recognizedmeasure of intelligence (Heffner).
Contentvalidity deals with the ability of a test to include all aspects of asubject matter under study. (Graham,2003).The problem of 2+2=4 would be included in the scale of intelligence,but it would not reflect all information about intelligence(Heffner).Thus, to develop a valid test, it must cater for aspects ofintelligence. Predictive validity is used to compare the correlationbetween coefficients. For instance, it can be used to show how SATscore influences grades in college.
Ifyou conducted a same task on two occasions and had the same results,it means that the study is consistent and reliable (Graham,2003).
Onthe other hand, reliability refers to the consistency of observation,survey, test or other measuring devices. Imagine you had a weighingscale that produced different results every time you used it, basinga research on such scale would be unreliable. In psychology,reliability coefficient is used to test consistency. To show theamount of consistency, the same test is carried out separately to aset of individuals (Graham,2003).Ifthe outcome of the test is consistent, then the test is reliable(Graham,2003).For instance, if two people were to observe a child and one of themobserved eight aggression acts, we would expect that the same is alsoobserved by the second individual.
Understandingvalidity and reliability of results and data can help determineconstructively a correct diagnosis and establish a treatment plan.These terms are significant as they are interrelated when one iscalled in question in any research, then the entire study becomesdisputable. Thus, how these variables are manipulated and controlleddetermine the difference between a meaningless and solid research.Also, it is important to note that all data that is valid isreliable, but not all data that is reliable is valid (Graham,2003).
Franzen, M. D.(2000). Reliabilityand validity in neuropsychological assessment (2nd ed.).New York, NY: Springer.
Graham, J. R.(2003). Handbookof psychology: 10.Hoboken, NJ: Wiley.
Heffner, C. L.(n.d.). Chapter 7.3 Test [Web blog post].Retrieved fromhttp://allpsych.com/researchmethods/validityreliability/