coincide with your answer to question 3b in terms
of the discriminatory power of the fitted model?
c. Suppose the distribution of observed and expected
cases and noncases was given by the following table:Partition for the Hosmer and Lemeshow Test
mrsa¼1 mrsa¼ 0
Group Total Observed Expected Observed Expected
1 29 10 0.99 19 28.01
2 31 11 1.95 20 29.05
3 29 11 2.85 18 26.15
4 29 11 5.73 18 23.27
5 30 12 9.98 18 20.02
6 31 12 14.93 19 16.07
7 29 12 17.23 17 11.77
8 29 13 19.42 16 9.58
9 29 13 21.57 16 7.43
10 23 9 19.36 14 3.64
What does this information indicate about how well the
model discriminates cases from noncases and how well the
model fits the data? Explain briefly.
d. Suppose the distribution of observed and expected
cases and noncases was given by the following
table:Partition for the Hosmer and Lemeshow Test
mrsa¼1 mrsa¼ 0
Group Total Observed Expected Observed Expected
1 29 10 10.99 19 18.01
2 31 11 10.95 20 20.05
3 29 11 10.85 18 18.15
4 29 11 11.73 18 17.27
5 30 12 11.98 18 18.02
6 31 12 11.93 19 19.07
7 29 12 11.23 17 17.77
8 29 13 11.42 16 17.58
9 29 13 11.57 16 17.43
10 23 9 11.36 14 11.64
What does this information indicate about how well the
model discriminates cases from noncases and how well the
model fits the data? Explain briefly.
e. Do you think it is possible that a model might
provide good discrimination between cases and
noncases, yet poorly fit the data? Explain briefly,
perhaps with a numerical example (e.g., using
hypothetical data) or generally describing a situa-
tion, where this might happen.Test 385