Hi I am looking at the chi-squared test completed by a colleague and I am not sure why we do not need to check that the total expected frequency equals the total observed frequency. Usually we would do this. Are we just saying we are not doing this because of the uncertainty within the first bucket. If we continue the buckets to up to $600 when we can then expect 0 in the buckets we get a different result. I am not sure why we can just stop to include the maximum of the observed claims. Thank you, Rachael
The short answer to this one is in Paper 2, unless you are told otherwise, don't look for errors in the existing calculations. If you were asked to do a Chi-square test in Paper 1 you would absolutely expect to include checks as you suggested. We could add an assumption about the maximum claim amount. Perhaps the highest value of phone covered by the policy is $500 and hence it's fine to use this as the maximum. There is certainly no need to make any changes to the spreadsheet but we could use this as an idea for some next steps. Sarah