Bernoulli

Today we talked a lot about Bernoulli distributions. I found these distributions incredibly interesting because we use them so much in the field of psychology. For example, at work I have to issue a GDI (Geriatric Depression Inventory) to all of my patients. The GDI consists of 30 questions about their day to day life and only allows them to answer "yes" or "no." When they are finished, depending on their answers, I can scale their score and decide if they are "normal", mildly depressed, or severely depressed. However, I have also found that a lot of my patients find it difficult to answer these questions because they want to say "sometimes" rather than "yes" or "no". Thus, in my opinion, I feel like my final result could have a slight error. Although the error may only be a point or two, it is still an error. Maybe the test should change the answers to having "yes", "no", and "sometimes". I think that with the third choice, we could have a slightly more accurate reading of the patients. Also, I think that because my patients have trouble only answering "yes" or "no", they become frustrated and start to have a nonattitude. I think that because they don't have the third choice, they settle for an answer regardless of if they think it actually represents their true feelings. Although this is not a poll persay, I feel the GDI has some similarities and should be looked at for some of the errors it could have.