This section will focus on the definition and the methods of evaluating reliability and validity in survey research.
Reliability and validity is a major issue when it comes to research, indeed failure to assure the validity and/or reliability of the findings may cause the research to be questioned even worse rejected as invalid.
Reliability refers to consistency and/or repeatability of the measurement; in other words, consistency can relate here to the questionnaires being clear and well define in order to not confuse the respondents and repeatability here means that if searchers have findings from a group they should be able to repeat the survey and get exactly the same results.
There are several ways to measure survey research consistency;
Consistency may be measured by using triangulation; this will be done using multi sources of information for example by taking different information from three different investigator or just by combining more than one methods of data collection in the same research; interview and survey and so on. This will actually fructify and increase consistency of the research and therefore reliability as well. The advantages from this approach are that is time consuming, expensive and request lots of effort.
Consistency could be measure by using one or more following strategies:
- Inter-Rater or Inter-Observer Reliability
- Test-Retest Reliability
- Parallel-Forms Reliability
- Internal Consistency Reliability
This will also result the findings to be reliable.
Validity refers to the degree to which the measurement procedure actually measures the concept that it is intended to measure.
Validity research has several ways to be obtained:
- Face validity
- Content validity
- Predictive validity
- Concurrent validity
The picture below shows us the relativity between reliability and validity; it gives us a well descriptive view of the relationship between reliability and validity.
As the picture shows, there are four possibilities, first possibility the measure can be reliable but not valid, second one the measure may be valid but not reliable, the third one is the worst case where the measure are neither reliable and not valid as well however the last measure is a perfect case where we have both reliability and validity.
To conclude from the picture, we noticed that reliability and validity have related ideas however one can work without the other one, or both at the same time.
William M.K. Trochim. (2006). Types of Reliability. Available: http://www.socialresearchmethods.net/kb/reltypes.php . Last accessed 18th Nov 2010.
Adri Labuschagne. (2003). Qualitative Research - Airy Fairy or Fundamental?. Available: http://www.nova.edu/ssss/QR/QR8-1/labuschagne.html . Last accessed 18th nov 2010.
Celia Taylor, Graham R. Gibbs and Ann Lewins. (2005). Quality of qualitative analysis. Available: http://onlineqda.hud.ac.uk/Intro_QDA/qualitative_analysis.php . Last accessed 18th nov 2010.
William M.K. Trochim. (2006). Reliability & Validity. Available: http://www.socialresearchmethods.net/kb/relandval.php . Last accessed 18th Nov 2010.