Word counts are increasingly used instead of column inch measurements. They might seem simple but there are a number of pitfalls. From painful experience, here are a few tips:
If you're comparing coverage between papers, use the same search parameters or the results won't be comparable.
If you're looking for stories about a subject, and time constraints mean you can't read through every article, restrict the search to articles with the key phrase in the headline/lead paragraph, as results will be more reliable
Skim through results to remove any articles that aren't strictly about the topic
Remove database headers and footers, such as the Lexis Nexis section and load-date fields, before you count.
Instead of using a calculator, paste the text into Word and use the word count facility - much quicker and more reliable!
Record the number of articles along with the word count - this gives a clearer picture as broadsheets tend to print fewer articles with more words, whereas tabloids print more, but shorter, stories.
Word counts by their very nature don't take into account accompanying photographs, which may skew the results, as tabloids use more pictures and less text
Explain your methodology when you submit your results - if anyone queries the figures this will help prove your point.
Remember that word counts aren't an exact science, but simply a snapshot of what is available on a database on a given day - and make sure your editor knows this
For recent examples of newspaper word counts seeThe Times and washingtonpost.com. Even blogs are getting in on the act, here and here. And there's a non-fan here.