Click below to go to:
Some Points on Writing a Paper
The following is abstracted from J. Scott Long, "Notes
on Writing Effective Papers for S650-2000: Categorical
Data Analysis," but is highly pertinent to this and any
other class you take in a university.
Long states and I agree that "...The care in which an idea
is presented affects the chances of an article being
accepted. Before you turn in your assignments, think
about whether you are presenting your material as effectively
as possible. This handout provides some suggestions that
can make your work for (this class...) and your later professional
work more effective. As with most rules, there are exceptions."
General Points
- Use a spelling checker and a grammar checker.
- Read the instructions carefully.
- If you don't understand the assignment, ask before you turn it in.
- When answering a specific question in an assignment, make
- sure it is clear which question you are answering.
- Put your name on all separate parts of your assignment.
- Number the pages.
- Use staples, and make sure they go through the paper.
- Use a fixed font for printing computer output. The columns
- must line up.
- Learn how to use the following:
i. "data are" not "data is"
ii. "effect" and "affect"
iii. "Ph.D." not "PH.D.", or "Ph.d.", etc.
- Stick to a few fonts of a reasonable size.
- Double or 1.5 space, except for tables and computer output.
- All tables should be set up so that the numbers are aligned. This
- is often easiest by making the table a fixed font.
- Don't use full justification if it creates large blanks in a line.
- When using relative terms, make sure that the comparison is clear.
- Don't use variable names in describing results unless it is
absolutely necessary. Say: "Being a female increases. Don't
say, "Increasing VAR04 increases . . . "
Organizing Your Paper
- Think of your write-up as telling a story. What
makes a story compelling for the reader?
- In general, don't tell the reader what you did not do.
- Don't make your paper a chronology of how you
got the results you got. The logic of
discovery and the logic of presentation are quite different.
- Start with the main point, then build the evidence
to support that point.
- Locate tables and figures where the reader can
find them and use them effectively.
Statistical Points
- Use the same N for all analyses unless there is a
reason for the N to change.
- A small P, a small t, or a small R2 does not imply
an incorrect model.
- Specify the scale of the variable. A change of 12
makes little sense if the reader doesn't have
a substantive understanding of what 12 means.
- Use 98% versus98 consistently.
In General
- Search for good examples of effective papers. Use them as a model.