Foundations Newsletter

“IELTS is all about luck”
“My friend took IELTS and she can’t speak English at all – and she got band 5!”
“My sister guessed on every question on the reading and got band 5!”

Do these comments sound familiar?

Recently my class of repeating students started flat out rejecting my reading practice on the grounds that they felt many students (their friends and classmates) guess on all the questions and get a band 4 or 5.  One student in my class insisted she guessed on every question and got band 4.5, and asked “so what’s the point of reading practice?” (and the rest of the class voiced their support).

So what are the chances of this actually happening?  What band can students realistically hope to get by guessing on every question on the IELTS reading section?  As teachers, we have no ammunition in this area to back up our claims that they really should not be guessing on the entire reading test.  As a result, our efforts to convince them are trumped by students’ first-hand anecdotes of many cases where guessing leads to a pass.

So I did some work to find the actual probability of getting band 4.5 or 5 on the reading section by guessing on every question.  I looked at eight past IELTS reading exams, from Cambridge IELTS Book 6 and 10.  I calculated how many questions they would get right on average if they guessed on every question, accounting for the changes in probability for the different question types and the varying number of types of questions on each of the eight tests.

Every test is different, but I discovered that the majority of the time, if students guessed randomly on every reading question, their average raw score over time would be about 7 correct (band 3).  Book 6 had outliers of 5.6 and 9.39, but book 10 was more consistent around the 7 average.  I used the more recent book 10, going with 7.26 as the average mark students could get by guessing.  I then made tables using both Poisson and binomial distributions, with similar results, but I’ll report the results from the Poisson distribution.  If any math teachers can help me out with what is a more accurate or appropriate method for this, please send me an email.  It’s also worth noting that not all IELTS raw scores convert the same way – some online converters are different, and even in the back of the Cambridge IELTS books, the raw scores can equal different general ability descriptions, depending on the difficulty of that particular test.  For now at least, I believe I have a reasonable approximation of marks students can expect to get by guessing on every question.

Here are the results:

To summarize:

• The vast majority of of the time, guessing will lead to no more than a band 3.5

• If a student took the reading test 132 times, guessing on every question, she would likely get a band 5 in reading one of those 132 times.

Interestingly though, band 4 doesn’t seem out of the question:

• If a student took the reading test 6 times, guessing on every question, she would likely get a band 4 in reading one of those times 6 times.

I showed my class the results, and the student who claimed she got band 4.5 by guessing everything then admitted she didn’t actually guess on ALL of the reading questions – she said she guessed on many of the hard questions, and attempted the easier questions, which is a reasonable strategy, given our students’ levels.

It would appear our students vastly overestimate the marks they can achieve by guessing, which makes them lose motivation to practice reading.  Who would want to put in the work if you think you can pass by guessing

When I talk to students, they seem to think their guessing odds are much greater than they actually are.  It’s likely that a few “guess for success” stories are skewing the students’ perceptions, especially since some of those students who claim they guessed on every question (and got a band 4.5 or 5) actually attempted the easier questions before guessing on the harder ones – but this part gets removed as the story spreads.  For students, the ‘effort’ part of the story isn’t what they want to hear (listener bias maybe?).

When I showed my class of repeating students some of my statistics and explained it to them, most of them then started reading the exercise I gave them without further resistance (and continued reading with their next teacher).  Taking the test 133 times to get a band 5, or even 6 times to get a band 4, sounds like more work than just practicing some reading.  The best strategy, they then told me, was to guess on the hard questions and do your best on the passages that are easier to understand.  Sounds about right to me. Sometimes it take some statistics to back up what teachers have been saying all along. 

Disclaimer:  The IELTS exam is not open with their grading system (especially raw score – band conversion, which varies), and the tests I looked at, although the latest released to the public, are still from years ago.  Also, every test is different, with varying numbers of multiple choice questions and varying probabilities.  As a result, there are many ways my statistics might be off.  So this is just to give a general idea.