Immediately before the ScienceOnline2001 meeting, well-known science writer Carl Zimmer wrote a piece on his blog titled Death to Obfuscation! In it he wrote,

[…] Scientists have a fierce passion for the passive voice. I suspect it has to do with the abject humility that they claim as a virtue of their profession. No one has the temerity to actually write, ‘We discovered X.’ Instead, ‘X was discovered.’

I agreed with the sound advice the article offered but I wondered a little about the claims about what scientists do and don’t.

So let me channel the spirit of this tweet from Tom Levenstein (retweeted via Gloria Lloyd and others):

Be as critical and curious about craft of writing as you would be about science.

My instinctive response to Carl’s example that no scientist would dare write ‘we discovered’ was to think it had two issues, not one. One the voice being used, active or passive, as he pointed out; the other being the use of the word ‘discovered’, which in many settings would be avoided as seeming too grand.

My next thought was ‘we observed’, a more modest replacement for ‘we discovered’. Wouldn’t that be common, I thought? Heck, I swore I’d read it in the last paper I was just reading. I looked back and sure enough, there it was.


I decided to informally examine if the accepted wisdom that scientists never used the active voice, or at least very rarely used it, by searching PubMed. (It is the accepted wisdom, I don’t doubt that, I just wondered how true it really was.)

PubMed is an archive of abstracts–the summaries at the head of research articles–hosted by the US National Library of Medicine, with links to the full articles, citations and whatnot. Nominally PubMed covers medical journals, but in practice it covers most, if not all, areas of biology.

A little trick* I’ve picked up is that you can search the PubMed abstracts from Google by using a search of the form:

site:www.ncbi.nlm.nih.gov/pubmed/ “we observed”

The first part specifies that site to be searched. The second part is a phrase being searched. This allows you to search on words that PubMed does not consider keywords and to use quoted phrases, which ask Google to match the complete phrase rather than search for the words in the phrase separately.

I first searched for ‘we’ and from the top ten pages of results collected the word followed the ‘we’, recording it if it was in the past tense. I then ran searches on those phrases. PubMed claims to have over 20 million citations. Here’s what I found, first the number of matches, then the word or phrase searched (Google always places ‘about’ before the count; I’ve omitted this for clarity):

8,440,000 we

443,000 “we found” (158,000 “we found that”)

200,000 “we investigated”

172,000 “we examined”

136,000 “we used”

79,800 “we compared”

73,900 “we observed” (19,900 “we observed that”)

54,700 “we measured”

47,500 “we determined”

31,800 “we reviewed”

24,700 “we aimed”

9,720 “we estimated”

5,240 “we discovered”

5,970 “we considered”

2,070 “we derived”

1,910 “we noticed”

1,820 “we saw”

1,590 “we assumed”

1,180 “we questioned”

1,110 “we modeled” (+ 261 “we modelled”)

581 “we increased”

525 “we improved”

500 “we reduced”

406 “we profiled”

228 “we learnt”

52 “we summarised”

39 “we realised”

21 “we helped”

Certainly less than a majority of abstracts use the active voice–there are over 20 million citations in the total sample–but that said it’s sizeable minority.

Would you conclude that a fair number of scientists ‘have the temerity’ to write in the active voice in the research literature? (One problem is the use of the ‘royal’ we, which I haven’t attempted to address.)

It would be useful to break this down by type of article (review, commentary, and so on) and research journal, but this will have to do for now.

Likewise, it would be interesting to compare the use of these phrases over time. You might hope that this is a growing trend, but it may prove that there has always been a minority favouring the active voice.


I missed Sylvia McLain’s (aka Girl, Interrupting) article and the ensuing discussion when I wrote the article above. Recommended reading for those interested in this topic.


* Yes, that word that some global warming critics have misread.

Some other science communication articles on Code for life:

Fact or fallacy, a survey of immunisation statements in the print media

Positive news

Banished from science writing. Words, that is.

Desk Guide for Covering Science, and academic conferences

How long does it take you to write a science blog post?

Note to science communicators–alleles not ’disease genes’

Science journalism–critical analysis, not debate