TO: Students in CJ 161-00
FROM: R. B. Taylor
DATE: 4/23/03
RE: Comments on election fraud papers
----------------------------------------------------------------------------------------
On your returned paper you may see one or more numbers. They
correspond to the listed comments below. Please let me know if you have comments
or concerns.
COMMENTS ON STINSON MARKS PAPERS
- The null hypothesis is that there is no difference in the
local machine proportion voting for Stinson as compared to the proportion of
the absentee ballot vote for Stinson. No difference between the sample and
the population.
- Paper fails to mention level of measurement.
- Paper fails to specifically mention the z obtained in the
paper.
- On the question of level of measurement, at the individual
level it is nominal, if you said this you got credit. But given that we are
looking at the whole vote here, and we are looking at PROPORTIONS that can
go from 0 to 100, the LOM is ratio.
- For step 5 you want to indicate a specific Z CRITICAL
score.
- The z obtained should be around 25 or 26.
- If the z observed exceeds the z critical you want to REJECT
the null hypothesis of no difference, rather than FAIL to reject
- Conclusion is consistent with your results, even though
your results are not correct, and so you got major credit for that
consistency; OR your conclusion is consistent with Step 7, in which case you
also get partial credit
- The data of interest is not the COUNT but rather the
PROPORTION of machine and absentee ballots for Stinson.
- The initial opening statement needs to provide more context
for the data: what is the context and what is the question?
- You have confused z critical and z observed.
- You are sounding like you are doing a two tailed rather
than a one tailed test, which makes the z critical you have chosen
incorrect. If you are doing a 2 tailed test, alpha < .05 the z critical
would be + / - 1.96.