These are a variety of comments I wrote on about 1/2 of the homeworks
turned in last year. I allow resubmissions to fix such errors. But
it would make both our lives easier if you followed these rules the
first time around!
Call your figures: figure 1, figure 2, etc.
Give only one digit of accuracy for p-values. (.0003)
Give SE's with ranges between 3 and 30. So .24, .04, not .395 or .01
Give estimates to accuracy of SE: If SE = .25 => estimate 1.05. If SE
= .3 estimate is 1.0.
Make one point per graph. Draw as much attention to the single point
you are trying to make in the graph. Anything that doesn't support
your point should be removed. Be sure to tell the reader what this
one point is in a caption! (So each graph should be highly edited to
show off the key point graphically and then it should be captioned to
tell the reader what to look for.)
For the discussion section you want to guide the reader
carefully. Avoid giving ALL the output. Just give what is necessary
to make your case.
The primary reason for putting figures at the end of a paper is
convience to the author--not to the reader. So if you CAN put them in
the text, that is always to be prefered. Now an appendix is a
different matter altogether! It contains stuff that no one wants to
read, but the author feels compelled to write.
When looking at returns, there isn't much point in connecting the dots
here since it doesn't "pass" through the intermediate points.
Contrast this with looking stock prices--where it should pass through
intromediate values. The way to see if it makes sense is to remove
every other point, and connect the line between them. If that
captures most of what happens, then it is a good thing to do.
Anything worth measuring is worth discussing the meaning of. So if
you bother to say that the beta is 1.05, then say what the beta
means. If you give an intercept say what we are supposed to learn
from it. (I.e. If the intercept were hugely positive--would we want
to buy the stock or not?) You should of course mention the SEs, and
hence they need to be interpreted also.