BLOG #4 Body Paragraph 1

               Content Validity Vs. Internal Consistency Reliability

    Validity is concerned about which characteristic is being assessed, sociability or selfishness for instance. One significant aspect regarding validity is content validity. It refers to component of questions in relate to a characteristic. An assessment with a high content validity has components of question items that are relevant to the characteristic that is being measured. Though it may seem easy to achieve a high content validity, choosing question items can be tricky and sometimes it is difficult to determine what kinds of content should be included in question items and what kinds of content should be excluded from them. Suppose that we are to measure sociability of individuals. We may include questions such as preference to spend free time with people rather than alone. But what about being good at speaking in public? Or tendency to initiate a conversation? Are those considered as one of the aspects of sociability or not? We can achieve high content validity by including question components that are relevant to a characteristic and by excluding those that are not relevant to it. Meanwhile, reliability refers to whether or not the data from an assessment is trustworthy. The most significant part in reliability is internal consistency reliability. It refers to whether question items of an assessment are accurately measuring an intended characteristic. Using the same example, we would choose three question items to which individuals rate their likings with the scale of 0 to 10; disco, nightclub and party at a friend’s house. However, those three items are only partially measuring one’s sociability. More importantly, they are assessing some other variables that are unique to those items. The first item may probably measure one’s liking of music and dance, in addition to sociability. The second item may probably measure one’s liking of drinking, in addition to sociability. Finally, the third item may probably measure liking of homey atmosphere, again, in addition to sociability. In order to maximize the internal consistency reliability, we have to increase either the number of question items or the correlation among the questions. For instance, we can add some more question items such as liking of local event or meeting new people. Now, the core elements of sociability are strengthened and unique elements to each question item are weakened. In this way, we achieve high internal consistency reliability. On the other hand, we can increase the correlation among question items. For instance, we can choose questions like preference to spend time with friends, preference to spend time with classmates and preference to spend time with coworkers. Then, the correlation among these three items becomes quite high and it indicates that the characteristic is most likely presenting sociability, not the liking of music or alcohol. However, the solutions come with a new problem. Trying to achieve a high internal consistency reliability, by increasing either the number of question items or the correlation among them, we are forced to sacrifice the content validity instead.   As we increase question items, the content of a characteristic is automatically being expanded. This means that the components of the question items become less and less relevant to the content of the characteristic , resulting in a poor content validity. Meanwhile, in increasing the correlation, we have to choose very similar question items. This means that the question items would represent only a small part of the characteristic. In other words, content of the characteristic is now very limited to only one specific aspect of the characteristic, also leading to a poor content validity. Content validity and internal consistency reliability cannot be high at the same time. Therefore, assessments always have to lack or compromise to some extent either content validity or internal consistency validity.