When you write about the past impacts of your research, you will need to provide evidence to substantiate your claims. When you design research projects to deliver future impacts, you should embed ways to collect evidence and track those impacts as they occur.
Evidence is necessary to substantiate:
You only need to provide enough evidence to substantiate a claim. One piece of evidence per claim is a useful rule of thumb.
However, some claims will benefit from more evidence, to drive the point home. A case in point is invited talks as evidence of the scholarly impact of your work on the discipline or field of research. Here, examples of three invited talks would be more compelling than one.
But you certainly don’t need to provide all the evidence on the public record.
Evidence can be quantitative (e.g. statistics) or qualitative (e.g. testimonials). In highly competitive funding schemes such as NHMRC Investigator Grants, quantitative evidence tends to trump qualitative.
Evidence should be verifiable by the reader or assessor. This means it should be publicly available. A testimonial in a private email is not ideal, because no one can check its veracity. (If it’s all you have, you might as well use it anyway; it’s better than nothing.)
Most evidence of impact will come from other people or organisations. In general, your own publications will not provide evidence of impact. The exception is if you have published on the impact of your own research, e.g. a cost-effectiveness analysis of an intervention you implemented.
