10.A medical tech company has developed new software which enhances digital images produced by CAT, PET, MRI and X-Ray. The company claims this enhancement will not only reduce false alarms but also increase hit rates in detection of anomalies by radiologists; this would be a good thing. Their technology is extremely expensive though and in order to convince hospitals to invest, they need to show some evidence that detection rates are as good if not better with the software then without. To do this, the company conducted a very comprehensive examination of how well expert radiologists performed at detecting known anomalies in images using both standard and enhanced images. Obviously the company is looking for a favorable result. A contingency table analysis was conducted, this time implementing McNemar's test in the options of SPSS. The significance for the test was found to be .035 . Assuming we are using a standard alpha of .05 when evaluating our statistics, is the McNemar's test significant? Should the med tech company be happy with the results?

11.When conducting a 2 x 2 table analysis and using a Mantel-Haenzel method, what is the null hypothesis being tested? (Bonus if you can state the null in terms of association and in terms of odds ratios)

12.True or False: I can find that raters have high interrater reliability, but still find they are horrible/inaccurate in their ratings.

13.If I compute a Cohen's Kappa for two raters, and find the number to be negative, should I be happy or concerned about their work?