I’m sometimes asked about how to format a table of results from a multilevel model. As with many statistical matters, while I certainly think that there are correct and incorrect conclusions that we can draw from statistical evidence, I’m often less interested in identifying approaches as 100% right, or 100% wrong. I’m more interested in: “What is a principled approach?”
So it is with creating tables. What is a principled approach? What basic table format offers the essential statistical information in a principled way? Hence the table template below.
I think the template is always a very good starting place for submitting your work. Of course, particular journals, or particular conferences, may ask you to present your results in a different format, but the below always seems to me to be a good place to start.
| Variable1 | \(\beta\) | t | p | Confidence Interval |
|---|---|---|---|---|
| AAA | 1.23 | 1.11 | .15 | [-.01, 1.45] |
| BBB | -3.21 | 2.01 | .04 | [-4.56, -1.23] |
| CCC | 2.34 | 3.21 | .01 | [1.23, 4.56] |
| Random Effects | variance | |||
| var(\(u_0\)) | .012 | [.011, .051] | ||
| var(\(e_i\)) | .456 | [.123, .456] |
Of course these are just pretend variable labels, and you would include real illustrative variable labels instead.↩︎