Sunday, 6 September 2009

Spell checkers - how useful are they?

People have said to me "Why bother teaching spelling nowadays? We all use spell checkers". This not only assumes that all writing is done electronically (which most of it is for some people but not all), but also that people know how to use them well (which some don't). A sentence in an essay from a student of mine started "The main thighbone is ...". I'm not sure what the learner had written but had obviously meant to write "The main thing is...".

I examined a corpus I gathered of 372 spelling errors from adult EFL learners with spelling problems. The mistakes came from handwriting so had not been through a spell checker. But I wanted to find out how useful a spell checker would have been to them. I ran the words through a Microsoft Word spell checker. Some interesting results:
  • 16% of the errors would not have been flagged by the spell checker at all because they were homophones or other real words.
  • 15% were shown as errors but the intended word (as indicated by context) was not offered.
So in over 30% of cases the spell checker would not have helped.

My student who wrote about thighbones probably just took the first alternative offered, without applying any brain cells to the decision. And this is another problem - very often you are offered several alternatives to choose from and the one at the top of the list may not be the one you need. Many of the errors offered up to five alternatives.
  • Of the 261 errors that were picked up by the spell checker and the right word was offered, 21% required the user to choose a word other than the first one shown, ie the second, third, fourth or fifth word.
So thinking still needs to take place. Some learners have admitted to me that they always take the first option shown. Blind faith!

The good news, perhaps, is that in 51% of the total sample, the intended word was the one offered first. But even then usually there were still other options that the writer needed to discount.

Maybe the easiest is when there is only one spelling offered. For example, a student who wrote 'culdn't' would have only been offered 'couldn't'. There were 62 cases where only one word was offered (16% of the total sample) . But of these 62, 11% didn't offer the correct word, so a writer can't even have complete confidence in this. An example to illustrate: a learner who wrote 'funately' meaning 'fortunately' would only have been offered one alternative, 'finitely', which would have made nonsense of his sentence if he'd accepted it.

So what's my point here? Spell checkers are undoubtedly useful. But if learners are too far from the original spelling or if they confuse two words the spell checker may not be helpful. We need to train these people in the use of them and encourage them to use other strategies too. Like much learner training, there is a danger of patronising learners - many will already know how to use a spell checker well.

An exercise like this may be the answer. Have a go and if it's useful, pass it on (but please link back to here).


There'll be more about using spell checkers and learning from them in my book, coming soon:

2 comments:

  1. Yes, Spell Checker is what we need, especially for blogger like us from Indonesia.

    ReplyDelete
  2. Thanks a lot for the test. I have got 100% and feel VERY motivated to learn more. ))

    ReplyDelete