Reliability

• What is the difference between reliability and validity? Which is more important? Why?
 
• What are the different ways of assessing reliability?

• What are the different ways of assessing validity?

• What are the different ways of obtaining validity evidence?

Reliability and validity are the scales of measurement that is used in research methodology. However, both the terms sound as like they are similar in nature, but there are technical differences between them.

Reliability is the scale or instrument of measurement that is consistent with its result. It means if any test is taken using any instrument for a particular subject and the result of the test will be similar in every attempt then it will be considered as reliable. Reliability is more concerned with the consistency and the stability of the test result (Bryman & Bell, 2015).

Validity is the instruments or scale by which one can measure the outcomes of a certain research hypothesis about how strong the outcomes are. With the help of validity, one can answer the question such as are we right about the research methodology or the outcomes. Validity is more concerned whether the measure has been done with the right concept or not.

Reliability and validity both are related with each other in the process of measurement. Reliability is the stability of the test score and validity stands for the accuracy of the test score. Technically both of them are important for a research methodology. It is also notable that research methodology can have reliability without validity but to obtain the validity of the result or the outcome it is very important to have reliability. It means without the reliability there is no existence of validity so it can be said that reliability is more important than validity (Cooper, et al, 2006).  

The measure indicates about the extent of the biases and ensures measurements that are consistent across time and items of the instrument. There are different types of methods for assessing reliability is discussed below.

In this process, the coefficient of the outcome can be obtained by repeating same measures for a particular research. Test-retest method is done when there is a questioner that contains few items, which are supposed to use for a measurement of a certain concept that has administered to a set of respondents. If the same respondents are used after few weeks or more than six months, then the correlation obtained are from two different times. This outcome from the same respondent in two different times called test-retest coefficient. The higher the coefficient, the better the test-retest reliability and the result of the test will be more consistent and reliable (Abdulkadir Uzunöz, 2011).

Validity

When there are two comparable sets that will be tapping for measures the similar construct and they are highly correlated then the assessing of reliability will be called parallel form reliability. In this, case both the research methodology will contain similar things and the response formats for both the studies remain same. Here the change, which is done to maintain the difference in two methods, is the sequence of the question in the questioner. In this type of research method if the two results are highly correlated then it can be said that the results are fairly certain and reasonably reliable.

With this methodology of measures, one can indicate the homogeneity of the things that can tap the construct. In this case, the research methods should be together as a set. In spite of remaining in a set, the items should has the option of independently measuring the concepts that are similar so that when respondents provide the same meaning for each item. The help of examination and their outcomes can see this type of measures whether the subsets of different items are highly correlated. Internal consistency of the research methodology can be examined by the interitem consistency reliability and the split half reliability test (Avcu, 2015).

In this method of measuring, there is the consistency of the reliability test by the respondents who will provide an answer to every item of the measure. In this way of measuring the reliability of the items, there should be measures that are independent should have the same concept that will remain correlated with one another. There are also two types of interitem consistency reliability test, which are Cronbach’s coefficient alpha and Kuder-Richardson formulas. The first one is used for the multipoint scaled items and the second is used for dichotomous items. Here is also the higher coefficient is, the better measuring instrument (Aygun & Ergun, 2015).

By the help of this reliability measure, one can correlate the two halves of an instrument. The two halves of the measurements will justify the estimate of the methods. If the Cronbach’s coefficient alpha is lesser than the split half reliability in those circumstances, which are being, underlined more than one response dimension then the measure of certain other conditions will be met as well. Looking after these measures it can be said that in all the cases interitem consistency reliability gets the perfect and adequate indexing from Cronbach’s coefficient alpha factor (DeLuca, 2011).

Different ways of assessing reliability

Different types of validity assessing methods are Content validity, Criterion validity and construct validity.

In the assessment of validity where measures are adequate and include representative item set is in content. When the scale of measuring item represents the higher domain of universe of concept the greater the validity of the content is. The content validity will be well delineated by the different function of dimension and elements of the measurement process. One of the minimum indexes of content validity is the face validity. This is the indication of those items that are intended for a concept to measure (Endo et al., 2014).

This is the validity that is established during the measurement differentiates of an individual on a expected or a predictable criterion. This type of validity demonstrates the correlation of the statistical relationship between job performance and the test performance. The Criterion-related validity is of two type

Concurrent validity – If there are validity test at the same time when the test is given then it is known as the concurrent validity

 Predictive validity – If there is a validity test that is done before and then the criterion are given to the method then this is known as the predictive validity (Fikkers, Piotrowski & Valkenburg, 2015).

This type of validity measure involves the requirement of demonstration about the measure of the test that measures the characteristic of the claimed measure. This type of characteristic is important for the successful completion of the job. There are also two different type of construct validity.

Convergent validity – This type of convergent validity is generally established because there are two different measuring instruments are used which are similar and correlated with each other.

Discriminant validity – In this case they are based on the same theory, but the variables are different from each other. In this case, the variables remain uncontrolled because the variables are predicted for the measurement instrument (Haner & Rude, 2015).

•    Evidence based on content – measure is done by studying the construct, test content is examined then it is to decide about the representation of the adequate test content. Experts do this.

•    Evidence based on internal structure – determination of dimension, which will be, used known as factor analysis. Item will be examined for homogeneity and coefficient alpha will be calculated.

•    Evidence based on the relation with other variables – Test scores are related to certain criteria. Correlation is done between the test score of the same construct and another construct to obtain convergent evidence (Leung, 2015).

Reference List

Bryman, A., & Bell, E. (2015). Business research methods. Oxford university press.

Cooper, D. R., Schindler, P. S., & Sun, J. (2006). Business research methods(Vol. 9). New York: McGraw-hill.

Abdulkadir Uzunöz,. (2011). Reliability and validity towards environment attitude scale for secondary school students. Educ. Res. Rev., 6(16). doi:10.5897/err11.236

Avcu, R. (2015). Turkish Adaptation of Utley Geometry Attitude Scale: A Validity and Reliability Study. Eurasian Journal Of Educational Research, 15(58). doi:10.14689/ejer.2015.58.1

Aygun, O., & Ergun, A. (2015). Validity and Reliability of Sun Protection Behavior Scale among Turkish Adolescent Population. Asian Nursing Research. doi:10.1016/j.anr.2015.04.006

DeLuca, C. (2011). Interpretive validity theory: mapping a methodology for validating educational assessments. Educational Research, 53(3), 303-320. doi:10.1080/00131881.2011.598659

Endo, A., Suzuki, M., Akagi, A., Chiba, N., Ishizaka, I., Matsunaga, A., & Fukuda, M. (2014). Reliability and Validity of the Upper-body Dressing Scale in Japanese Patients with Vascular Dementia with Hemiparesis. Occup. Ther. Int., 22(1), 10-18. doi:10.1002/oti.1378

Fikkers, K., Piotrowski, J., & Valkenburg, P. (2015). Assessing the Reliability and Validity of Television and Game Violence Exposure Measures. Communication Research. doi:10.1177/0093650215573863

Haner, M., & Rude, S. (2015). Establishing the Reliability and Validity of a Performance Measure of Big Picture Appraisal. Cognitive Therapy And Research. doi:10.1007/s10608-015-9688-2

Leung, L. (2015). Validity, reliability, and generalizability in qualitative research. J Family Med Prim Care, 4(3), 324. doi:10.4103/2249-4863.161306