Summary tables for reporting analyses

The output for the same analysis from different software packages sometimes contains different information. Working out what to include in summary tables or text can be tricky.

The clarity (and sometimes accuracy) of the labelling of the output also varies, so you shouldn’t automatically copy a table of results from the output.

There may be some specific formatting requirements for tables in your discipline area, but here are some general tips on creating tables:

• Consider the number of decimals places for the statistics you report.  The output may include (many) more than you need.
• Avoid reporting two statistics in a single cell in a table; use a separate column (or row) for each.
• Avoid using brackets within tables.
• Avoid reporting different statistics in the same column of a table, unless clear column labels are used.
• Use clear labels, rather than abbreviated variable names that might be used in software.

Here’s an example of a table with poor formatting:

	Mean (SD)
Age	72.473 (6.54)
Gender – % male (n)	56.6667 (68)
Employment status – % retired (n)	74.17 (89)
Total_Cholest	194.1±37.1

Here’s a better version:

	Mean	Standard deviation
Age (years)	72.5	6.4
Total cholesterol (mg/dL)	194.1	37.1
	Number	Percentage
Males	68	56.7%
Retired	89	74.2%

These links provide examples of how to summarise results for particular kinds of analyses:

• Independent samples t-test and confidence intervals for the difference of means
• Linear regression

To come:

• Analysis of variance
• Logistic regression