Reliability and Validity questions Background info Our rm cu

Reliability and Validity questions:

Background info: Our rm currently needs to hire a number of administrative assistants. The demands our lawyers place on our administrative assistants are very high; they often yell at them or ask them to do very difcult things in a short amount of time. We have found that the best predictors of these assistants staying more than a couple of months is for them to have very low levels of neuroticism. We used to buy a test (called the WIN) that would measure this, but it has gotten too expensive, and so I’ve started to develop my own measure (I call my test the Neuroticism Assessment Form, or the NAF). To design my measure, I spoke to a few experts in the per-sonality eld, looked at some of the items on other tests of neuroticism, and wrote questions that made sense. I asked a number of current administrative assistants to ll out the neuroticism test we formerly purchased as well as the one I am developing. I also have job performance ratings for these administra-tive assistants. Here’s what I’ve found so far:

-The correlation between their scores on the NAF in July and their scores on the NAF in August is .89.

-The correlation between their scores on the WIN and the NAF is .75.

-The correlation between their scores on the NAF and a test of extra-version is .34.

-The correlation between their scores on the NAF and their performance ratings is .30. So here are my questions

questions:

1. Is my test a reliable measure of neuroticism? If not, what else do I need?

2. Is my test a valid measure of neuroticism? What additional information should I collect, if any? **please talk about all types of validities and reliabilities, thank you

Solution

1 Is my test a reliable measure of neuroticism? If not, what else do I need?

The correlation coefficient tells about the strength of the linear relationship and if the test leads to the finding that the correlation coefficient is not significantly different from 0 (it is close to 0), we have to conclude that the correlation coefficient is \"not significant\". Since the values of the different correlations are varying and decreasing and approaching 0 one has to conclude that the devised test is not a reliable measure of neuroticism. We needed to estimate the relationships among variables by doing a regression analysis as well.

2. Is my test a valid measure of neuroticism? What additional information should I collect, if any?

Since the correlation coefficient is approaching 0 and since the regression analysis has not been done between the between the dependent variable and the independent variables this method cannot be considered trustworthy or of performing consistently i.e. reliable and logically or factually sound i.e. valid.