1. All tables should have a number and a title. The number may be keyed to chapters (i.e.,  the tables in Chapter VII begin with 7.1 and are incremented). The title should either be academic ("Determinants of Voter Turnout in Bosnian General Election of 2006") or punchy ("How Religion, Ethnicity and Age Drove Turnout in the 2006 Bosnian General Election"). In presenting results, keep referring to tables by number so readers don't get confused.
   
   
2. It's customary to note the location of a table in the body of a paper with a placeholder. You insert something like "TABLE 7.3 ABOUT HERE" as a separate line of text where you want the table to be placed, usually at the end of the first paragraph in which you refer to that table. The tables are usually placed at the end of the paper. You should put only one table on a page and, if more pages are necessary, indicate that the table continues on the next page at the bottom of the first page on which the table appears.
   
   
3. The base of the table should indicate the source of the data in brief with a fuller reference in the text. Something like "Source: 1996 American National Election Study" will do nicely.
   
   
4.  The general rule of thumb is that you should present the table and then summarize the findings in the text. The table should be clear enough that the reader could skip the text and understand what you found and the text should be clear enough that the reader wouldn't have to look at the table to understand your results. This does not mean you have to report each value in the table in the text, just that you should summarize the findings of the table.
   
   
5. All tables must report the number of cases in the analysis. Ideally, the reader should be able to recreate the numbers from the information you provide.
   
   
6. The kind of analysis you're presenting will determine what information should be provided. In OLS regression models, it's customary to report the unstandardized regression coefficient (the b), the standard error of each coefficient, and something that indicates the level of statistical significance. This may be the Z score, the 95% confidence interval, or, more commonly, the p-value of coefficients. We customarily signal that p < .05 with one asterisk, p < .01 with two asterisks, and p < .001 with three asterisks. These are commonly associated with the coefficient. In a simple Chi-square analysis, we usually report the value of Chi Square, the degrees of freedom, and the significance level. Look at journal articles in your field as models.
   
   
7. Do not use mnemonics or other labels from your computerized data set to represent variables. In tables, indicate what the variable represents in plain English. So your composite variable, INTEFF, should be presented as Internal Efficacy. As a rule, we assume that higher values of a variable mean more of the phenomenon so that, for example, education is typically coded from 0 to the maximum number of years of formal schooling. If the coding of a variable isn't obvious, tell us. If you use a dummy variable (0,1) for, say, gender, indicate that (female=1) when you list the variable in the table.
   
   
8. As a rule, don't reproduce raw computer output but translate the output into a professional-looking table. Most computer routines now allow you to output your findings into a format that looks more like a table. You can also use a spreadsheet to produce a professional looking table, and paste it into your document (with some minor editing).
   
   
9. Two things help readers make sense of complex tables. In a crosstabulation, calculate percentages of each category of the independent variable, present them in a column, and indicate 100% and the number of cases at the bottom of the column.  If the column percentages do not sum to 100% due to rounding, note that. Similarly, it sometimes helps to footnote the first value in a table and explain it thusly--"This entry represents the impact of each additional year of education on the feeling thermometer score for President Bush." Use notes at the bottom of the table to explain anything that the reader needs to know to understand and interpret the table.
   
   
10. When in doubt, consult Jane E. Miller, The Chicago Guide to Writing about Multivariate Analysis.(Chicago: University of Chicago Press, 2005).

*with useful contributions from Michael D. Martinez