I’m co-writing a short publication with a friend. She’s a PC user, I’m a Mac user. My version of Word is MSWord for Mac 2004, hers is MSWord 97 (according to the document info - I’m double checking this with her). We’ve reached our target total of 30,000 words but the word counts vary every time: mine is about 150 words greater than hers, for the very same document.
As well as being curious as to why this would happen, it concerns me slightly: I do a lot of translation work and charge by the word. I’d hate my clients to think I was tweaking wordcounts in my favour if they were to double-check and see a lower total.
So what could the reason be? Does mine count contractions like “don’t” as two words, or does hers count hyphenated word-pairs as one? We both have the footnote box unchecked and the document doesn’t have footnotes anyway.
I know from experience that Word counts contractions as a single word and does not omit any real words. I don’t remember any variations between versions on that.
Footnotes are one possible source of discrepancies, as it’s an option you can turn on or off.
Lists are sometimes a case where things get counted as words that really aren’t (such as the 1) 2) 3) or a. b. c. numbering). I’m not aware of any different in counting those between Word versions, though - I just know that if you Word’s list function, they are not counted, but if you type the numbering and then adjust indents manually, they will be counted.
Are you using the Word Count function under the Tools menu, or are you just relying on what shows up at the bottom of the document window? With documents as large as the ones you’re talking about, it isn’t efficient for Word to constantly recount the whole document. Instead, it does its best to add and subtract from what it already counted; there are sometimes small variations. If you use the Word Count function, it does the counting from scratch and you can be sure the number is accurate.
Sorry for this tangent, but this discussion made me recall that back when I was in elementary school a teacher mentioned that the statistic of “characters divided by 6” was used as a better approximation of the length of a piece of prose than strict word count; it was also probably easier to count in typed documents in the days before proportional fonts. I definitely can see that in a language like German where nouns like to stick together, word count is a terrible statistic. Often one can do similar things in English although not as easily and more rarely.
Is any of the text marked as Hidden? I just ran a test in Word 2007 wherein I started with 20 words total. I then selected 4 words and marked them as Hidden using the Font dialog. The resulting word count shows 16.
It may be possible that one version of Word is counting the hidden words while the other is not.
Another thing to look at is whether or not the Word Count function is counting the words in textboxes, footnotes, or endnotes. In Word 2007, this option can be selected in the Word Count dialog. It may be possible that you are counting said words whereas your colleague is not.
You might narrow the focus of your search a little by each selecting the same amount of text … say 1 page or several paragraphs … and each getting the word count. This might help you determine where the creep is happening. Try this with and without selecting any text that includes Captions, textboxes, footnotes, endnotes, etc.
This is pretty easy to test. Just make a very short document with something like
“hyphenated-word comma,splice period.splice number1nthemiddle colon:splice a I”
etc.
Word processors don’t all use the same metrics for determining which characters count as ending words and which don’t. And despite the fact that you’re both using Word, the Mac and Windows versions of Word are developed by completely different teams, and may have changed their mind about which characters are breaks in words and which aren’t.
There’s no hidden text in the document, no textboxes, no endnotes or footnotes. No numbering, bullet points or indents. I use the word count function in the Tools menu with the endnotes option unchecked.
I’m going to try counting it in sections now and will run the test suggested by iamthewalrus.