Reliability And Validity Of The Test

In this research, the main measures being used are reliability and validity. The difference between the two is that reliability determines whether or not other researches can reproduce results if provided similar conditions. Validity, however, is whether or not the measure being used is doing what it says it is doing. If the test is focused on one construct, it is called convergence, which means that the test is valid. If the test is focused on several constructs, the results cannot be significant, and this is called divergence, which means that the test is not valid. For example, a spelling test that has five mathematic related questions on it is given to three samples of students. The test should be reliable if all three groups produce consistent results, but the test will not be valid because the test uses mathematical questions to attempt to determine ability in spelling. One important reason to focus on validity and reliability is to see if the results give the ability to make significant conclusions and therefore discuss correlations. If the test is not valid or reliable, then it is likely that the results should not be reported due to flaws in the survey.

Methods

For this survey, there were 100 student participants. Among these college students, 17 were boys and 83 were girls. The minimum age for these 100 people was 18 and the maximum was 27. There were three people missing in the age question, and the reason for this is un-known. With that in mind, results showed a younger audience (M = 19.76), and these participants had generally low variability (SD =1.07). The scale that was used is called the Rosenberg Self-Esteem scale. The scale had ten items, the mean of the response to the ten items gave a representative scale o...

... middle of paper ...

...One way that we can measure the reliability of the Rosenberg self-esteem scale would be to collect a group of people from the survey that had relatively low responses to the construct, inform them of the construct and what it is, and have them retake the survey to see the correlation between the original test and the new test, this is called In the data set, there is little to no correlation between height and self-esteem. This supports the construct validity of the survey, because height would be a separate construct that should not be related to the data set at all. If height was a separate construct, then there would be a nonsignificant difference between height and self-esteem, rather than a positive self-esteem shown by the entire sample. If there was a correlation, that would indicate that the data set is not measuring what it is supposed to be measuring.