2

Is there any convention regarding the minimum information that should be included in a table showing a linear regression output?

The reason for this question is that I want to make the output as easy to understand as possible. I have graphed the main findings but I want to include the other outputs in a detailed table.

This is in a general social science context (i.e. a presentation; not journal publication) and my audience is not trained in statistics, so I want to keep everything as simple as possible (while still retaining the critical information).

Andy W
  • 15,245
  • 8
  • 69
  • 191
Amarald
  • 1,045
  • 5
  • 15
  • 21
  • There are many conventions for doing this. Different journals like different things; different programs do different things by default. – Peter Flom Oct 10 '12 at 22:10
  • Many social scientists will be used to APA format, using the style guide of the American Psychological Association. – Peter Flom Oct 10 '12 at 22:23
  • 1
    See [What is a good resource on table design](http://stats.stackexchange.com/q/3542/1036) and [Some notes on making effective tables](http://stats.blogoverflow.com/2012/02/some-notes-on-making-effective-tables/) for references of interest. IMO for presentations you want to focus information much more than typical. Tables with more than 4~5 rows of information are too difficult to read in that venue. – Andy W Oct 11 '12 at 02:44
  • You might find the two [Jane E. Miller books](http://www.press.uchicago.edu/ucp/books/author/M/J/au5172592.html) helpful. – dimitriy Oct 10 '12 at 22:48

1 Answers1

3

If a table is really called for, these information should be included:

  1. Regression coefficients
  2. An asterisk system or bold font to indicate statistically significant coefficients
  3. If space allows, add a column to show confidence interval.

Another useful technique to build in some animated boxes or arrows to direct their attention to the key information. That way you won't lose them.

But the important question is, if they are not trained in statistics at all, why show them a table? How would knowing the other coefficient help them? Instead, I'd suggest you to finely craft the interpretation of the results and then put it onto the slide. Just state the regression coefficient of the main predictor and whether it's significant. And then as a foot note on the slide, indicate what variables were controlled for. Remember to avoid causal inference.

Before the presentation, prepare a copy of the actual print out. If there is anyone asking in-depth questions, ask the person to talk to you after the event for more details.

Penguin_Knight
  • 11,078
  • 29
  • 48