Dear Colleagues,
I have four items on one scale, two are positively worded and two are negatively worded. If I revise the negative items, the scale earns a very low Cronbach's Alpha, but if I create a scale without revising the negatively worded items (include all the items as they were asked in the original questionnaire), the Cronbach's Alpha becomes powerful. I am getting confusion, is it okay to use negatively worded items without revising them?
Thank you for your answers in advance.
Janeth
A low value for Cronbach`s alpha may mean that there aren't enough items on the test. Adding more relevant items to the test can increase alpha. Poor interrelationships between items can also cause low values, so can measure more than one latent variable.
I am not sure that I have a clear enough picture of the question to answer it, so adjust the weighing of my answer accordingly. 1) Having both positively and negatively worded items adds to the validity of your responses. You're controlling for whether people are paying attention. 2) You likely want to reverse code the negative items, which I think is what you're asking. So, on a -3 to +3 scale, you're merely reversing sign. On a 1 to 7 scale, 1 becomes a 7, 2 becomes a 6, 3 becomes a 4, and the reverse is true. This will give you a truer measure of the underlying construct.