The journal *Nature Immunology*, now 5 years old, has made some worthy changes to its author guidelines.

Firstly, authors are encouraged to post *their accepted research manuscripts* (rather than preprints) in the appropriate archives hosted by their institutions and funding bodies (rather than on a personal site). "Depositing unrefereed preprints in electronic archives" also does not count as prior disclosure, which I would like to interpret as meaning that discussing your work in weblog posts is also allowed.

Secondly, the contributions of individual authors to the work will be described, so readers will know who did the major parts of the work.

Most interestingly, there is now a set of guidelines (and a checklist) for statistical interpretation of data which authors must follow. I think that no journal should be accepting papers using less stringent rules than these (edited here for brevity):

- Every article that contains statistical testing should state the name of the statistical test, the n for each statistical analysis, the comparisons of interest, a justification for the use of that test (including, for example, a discussion of the normality of the data when the test is appropriate only for normal data), the alpha level for all tests, whether the tests were one-tailed or two-tailed, and the actual P value for each test.
- Data sets should be summarized with descriptive statistics, which should include the n for each data set, a clearly labeled measure of center (such as the mean or the median), and a clearly labeled measure of variability (such as standard deviation or range). Ranges are more appropriate than standard deviations or standard errors for small data sets. Graphs should include clearly labeled error bars. Authors must state whether a number that follows the Â± sign is a standard error (s.e.m.) or a standard deviation (s.d.).
- Authors must justify the use of a particular test and explain whether their data conform to the assumptions of the tests.

It might be useful to have a template paragraph covering all of these points, which authors could modify to suit their needs. There is also a real need for some kind of learning module to cover statistical analysis - it shouldn't be difficult (it's mostly just jargon), but all the statistics resources I've read online have been confusing and incomplete.