Document Preview:

CO5124 Group Assignment 1
Group Report (25%)

Group Topic:

Very Poor Poor Average
Good Excellent Comments Mark Non-statistical description
Description of the issue (in clear, everyday language) and why it’s significant

/20 An overall answer or conclusion (in clear, everyday language) Relationship Does the model make any sense?

/30 Is there a relationship, or a definite lack of a relationship? R-squared? F-test confidence? Sanity checking Residuals: non-random? Heteroscedasticity?

/30 Co-linearity: do any X variables have high correlation, or is VIF more than 5 or 10? Presentation skills
Grammar, spelling

/20 Writing style, layout References have a consistent and complete style with author/title/year/etc.

TOTAL / 100 for 25%
MARK
Some popular questions, in case nobody in the audience asks anything:

Does X cause Y, or Y cause X? (Or X1 or X2 or whatever variable looks the most dodgy)

Could there be some lurking variable Z that causes all of Y and X1 and X2 and so on?

Your prediction presumes the future will be similar to the past; could some out-of-the-blue unforeseen event happen that is wildly different? (e.g., dinosaurs versus meteors)

Ethics: is what you propose unethical / murderous / worse than Hitler / going to kill thousands of babies?

Could we get a better fit by adding / dropping an X-variable, such as [something]?

What other factors might be causing the variability?

How difficult would it be to get a bigger sample? (Or equivalently: This result is significant only at the 5% (or whatever) level, so there’s a small chance it could be just lucky. How might we get better confidence?)

This result is only significant at the 5% (or whatever) level. Are the consequences of being right or wrong about the same, or is…