Do different types of hospitals have different practices when completing death certificates?
A health official and her team compare the results of 24883 randomly selected autopsies to the causes of death listed on the certificates. Death certificates in the study came from three types of hospitals, public; private, for profit; and university hospitals. Of the 24883 death certificates evaluated, 6785 were confirmed to be accurate, 8140 contained errors but did not require recoding of the underlying cause of death, and 9958 were incorrect and required recoding.
Data are shown as a 3×3 contingency table.
|
Confirmed accurate |
Inaccurate, no change needed |
Incorrect, recoding needed |
All |
|---|
|
Public |
997 |
1206 |
1493 |
3696 |
|---|
Private |
2213 |
2670 |
3232 |
8115 |
|---|
University |
3575 |
4264 |
5233 |
13072 |
|---|
All |
6785 |
8140 |
9958 |
24883 |
|---|
The health official uses a chi‑square test of independence to evaluate her null and alternative hypotheses.
?0:There is no association between hospital type and death certificate status.?1:There is an association between hospital type and death certificate status.
The value of her chi‑square test statistic ?2 is 0.484 with four degrees of freedom.
Complete the analysis by computing the ?-value of the official’s test and then decide whether she should reject her null hypothesis when the significance level is ?=0.01.
Compute the ?-value of the official’s test using software. Report your results to three decimal places.