In: Statistics and Probability
3.14
The first application listed in Section 3.7 involved British insurance claims.
The 2005 median was £1570. A random sample of 14 claims from a large batch
Received in the first quarter of 2006 were for the following amounts (in £):
1175 1183 1327 1581 1592 1624 1777 1924 2483 2642 2713 3419 5350 7615
What test do you consider appropriate for a shift in median relative to the 2005 median. Would a one-tail test be appropriate? Obtain a 95 percent confidence interval for the median based upon these data. If all amounts were converted to, say, Burma or to $138, would your conclusions be the satire?
I NEED TO GET TO THIS ANSWER: TESTING Ho = 1570, TWO TAIL P= 0.570. ONE TAIL P= 0.0287. 94.3 PERCENT CONFIDENCE INTERVAL BASED ON SIGN TEST IS 1581. BUT THE EVIDENCE IS NOT STRONG THAT CHANGING CURRENCY WOULD ALTER THE CONCLUSION.