Using personal pronouns in research writing


15 October 2009


Should you use “I” or “we” or neither in your thesis or paper?

Thoughts on this have changed over the years. Traditionally, using personal pronouns like “I” and “we” was frowned on. Instead of saying “In Section 3, I have compared the results from method X with those of method Y”, you were expected to write “In section 3, the results from method X are compared with those from method Y”. This is known as writing in the “passive voice”, and for many years it has been considered the “academic” way of doing things. I think it is favoured because of the tone of detachment and impersonality that it helps establish.

Sometimes the passive voice is awkward. For example

In studying ARIMA models, the effect of the estimation method on forecast accuracy was explored.

This is easier to express using “I”:

In studying ARIMA models, I explored the effect of the estimation method on forecast accuracy.

In my exponential smoothing monograph, one of the coauthors preferred to write everything in the passive voice, which led to some rather awkward phrasing. (I edited all chapters to consistently use “we” before it went to print.)

There are still some journals and research supervisors who insist that research writing must be in the passive voice. However, the situation is slowly changing and now many journals accept, or even encourage, the use of personal pronouns. The International Journal of Forecasting which I edit allows authors to use whichever approach they prefer.

A related issue for research students writing a thesis is whether to use “I” or “we”, especially when the material has previously appeared in a co-authored paper. In general, I prefer students to use “I” when they mean the author, as it is their thesis. (The royal “we” should only be used by monarchs.) However, it is very important to include a statement at the front of the thesis clarifying the role of co-authors involved with any parts of the thesis. If a chapter is essentially a co-authored paper, many universities require a signed statement from all authors.

One area where “we” is useful is in referring to the reader and author together. For example,

In the following theorem, we see that …

This is particularly common in mathematics.

In summary: