Reliability and validity are two important concerns in research, and, both reliability and validity are the expected outcomes of research. Validity means you are measuring what you claimed to measure. On the other hand, reliability claims that you will get the same results on repeated tests. There is a relationship between both characteristics of research, however, it does not mean that being valid also means being reliable and vice versa. For example, an invalid measure can be reliable because every time you test it, it produces the same (reliable) invalid results. Obviously, reliability and validity are equally important for any research.
In the research, reliability is the degree to which the results of the research are consistent and repeatable. Oftentimes, researchers repeat research in different settings to compare the reliability of the research. Finally, a reliable and reproducible research outcome yield new theories.
In experiments, you can achieve high reliability by repeating the experiments several times. Also, the researcher uses objectivity to improve the reliability f research. In case, the researcher developed a good research instrument and used it in the right way you will get reliable results. Because in quantitative research objectivity is easier to achieve.
In social sciences, the researcher uses different techniques to achieve more reliable results. However, in social sciences it is difficult to achieve reliability in the data collection, because, human behaviors are difficult to repeat even in similar situations. There are several external factors that influence the human behaviors and it is very important to know the effect of any such factor other than the independent variable.
There are certain techniques that can help improve reliability in a qualitative research. For example, rigor and control, sample representativeness, clarification, and justification are some ways to improve reliability in qualitative research.
Types of reliability
Test-retest reliability measures the reliability of the measure over a period of time. In social sciences, a test is administered more than one time over a period of time to check or retest its reliability. In natural sciences, the researcher conducts experiment more than one time to ascertain its reliability. The results of the tests and the inferences drawn have to be applied to natural settings, they should be reliable. But, this method of testing the reliability of the test is time-consuming, since the researcher has to wait for some time to re-administer the test.
Parallel forms reliability
It measures the reliability of the test by administering it in two different forms. Both forms of the test measure the same variables under study, but the format of the measure is different. The researcher must be able to formulate two different tests that measure the same variables. But the difficulty arises in formulating two tests that are similar in nature and measurement level. The researcher may also find it difficult to administer it to two similar populations. In social sciences, using parallel forms of same test is difficult and the issue of subjectivity arises.
More than one tester or judge uses inter-rater reliability check to measure reliability. The researcher asks more than one people to rate the reliability of his test. This type of reliability test is useful for subjective measures where more than one rater can best describe the reliability of the test.
Internal consistency reliability
Kumar R. (2000.a) in Research Methodology stated that he idea behind internal consistency reliability is that items measuring the same phenomenon should produce similar results.  It means that the items that test the attitude or behavior, are divided into half. Each half is tested separately and then their scores are correlated. For example, the researcher has developed questionnaire to test the attitude of people towards any state program. The researcher might divide the questions on the questionnaire in half and administer both questionnaires separately. The resultant scores of the test are correlated to know the internal consistency reliability.
Validity refers to the accurateness of the research as a whole and the accuracy of each step independently. It is the highest aim every researcher wants to achieve. When we measure what we have intended to measure we reach a conclusion that is valid and verifiable. To achieve validity one needs to use the right tools, processes, and methods in the research.
According to Kerlinger, ‘the commonest definition of the validity is epitomized by the question: Are we measuring what we think we are measuring’. Therefore, the first step to achieve validity in a research is to develop objectives that really target the questions that you have formulated. Kumar, R. (2000.b) in his book Research Methodology states that to establish the validity in research, the researcher should use logic and statistical evidence.  The concept of validity was formulated by Kelly (1927) who stated that a test is valid if it measures what it claims to measure.
There are two types of validity important to measure in a research: internal validity, and external validity. Let us see how;
Internal validity refers to the accurateness of the data collection tools, procedures, and techniques. In other words, it is the accuracy of the internal design that needs to be controlled by the researcher. An inaccurate research design and techniques cannot yield accurate and valid results.
McLeod, S. A. (2007) states that internal validity refers to whether the effects observed in a study are due to the manipulation of the independent variable and not some other factor. In-other-words there is a causal relationship between the independent and dependent variable. In experimental designs the researcher needs to be very careful in measuring the variables, the external variables can have an impact on the outcomes of the research and it can ruin the research.
External validity refers to the extent to which the outcomes of the research can be generalized to the population. In short, you can apply the findings of your research to a larger audience. Conducting experiments in natural settings can help improve external validity of the research. External validity aims at a far reach and scope of the research, you cannot limit your research to the laboratory where you have conducted the experiments.
Types of validity
Face validity as the name suggests shows the face-off value of the research or the measures used in the research. Although, it is not an authentic way to check the validity of the research. It is a general validity measure for the common people. So, people taking the test can measure face validity. Also, the researcher can ask some experts in his field to check the measure and its validity.
The researcher can hire a panel of experts to check the construct validity. They check whether the construct measures what it needs to measure. As the measure should measure what it is intended for and not some external factor.
Sometimes the researcher develops instrument to observe some criteria. The validity of the criteria can be judged by comparing it with another future assessment, if the future assessment proves to be successful it shows that the criteria or the test devised to test a behavior was valid and should be used again.
Reliability and validity are the two most important characteristics of the research. Therefore, every research whether it is in social science or physical science, literature or art there should be reliability and validity in the research.
- Kumar, R. (2000.a) Research Methodology. Sage Pub. London. Pp-140
- Kumar, R. (2000.b) Research Methodology. Sage Pub. London. Pp-137-138
- Kelley, T. L. (1927). Interpretation of educational measurements. New York: Macmillan.
- McLeod, S. A. (2007). What is Validity?. Retrieved from http://www.simplypsychology.org/validity.html