Hi there, this is Fiona from IELTS with Fiona, and this week I've been looking into online IELTS writing checkers.
Now, there are two reasons for this.
The first reason is that I started getting a lot of questions about how can I check my IELTS writing score.
And the answer was, well, you can just go and do the test.
The problem with that is the test is really expensive, isn't it?
And by the time you've paid for travel and hotels and everything, it costs quite a lot.
So I Googled it.
And of course, there are lots of online IELTS writing checkers and most of them free.
So I wanted to do some research into these.
But also, I discovered something new this week, and it's called testpredictor.com.
And that spelling is with a K. I'll put it on my website.
There is a blog for this podcast.
So if you go to my website ieltsetccom, the blog is called Which is the Best Online IELTS Writing Checker.
Yeah.
So this IELTS predictor.
No, it's called exam predictor.
Sorry.
Is it?
Hang on.
Test predictor.
Test predictor.
Yeah.
It does cost 30 pounds, but it's claim.
It says that it can save you the money of taking the test because it tells you if you're going to get the score you want.
So I'm interested in that and I wanted to see what the free versions could do.
Now, I think you all know Grammarly exists.
Grammarly, you know, it's a really useful tool.
And by the way, I'm really in favour of AI tools.
People say they're going to replace teachers and OK, maybe.
But I'm not against them.
I really encourage my students to use any tools they can to help them improve.
And I use this new chat GPT a lot.
And I often say to my students, you know, you don't need to ask me that.
You can check that yourself.
And it's better if you check it yourself, because you'll internalize the information better, I think, if you're doing the research.
So yeah, I'm absolutely in favor of students using things like Grammarly and even Google Docs, where they can point out your mistakes, you know, while you're preparing.
You can't always have a teacher helping you.
You can do so much nowadays by yourself.
But There are drawbacks.
I think Grammarly, the biggest drawback I found with that is that it tries to get you to write more naturally so it's easier for the reader.
And very often they don't like the passive tense.
None of these AI tools like the passive tense.
They say it's too formal.
Of course, for IELTS, it's not too formal.
So you've got to be careful about the advice that they give you.
And Grammarly, now I think you have to pay to get better advice.
I tested these with the same essay.
So you can see that on my site.
I put in quite a simple model answer that I'd written about mobile phones.
And I tested them with a different site.
Now, I did choose quite a simple answer deliberately because well, I didn't want it to be a perfect band nine.
I wanted to be able to assess the feedback.
And also sometimes simple answers.
Simple model answers can just help you improve your structure and coherence and cohesion.
So I used this model answer about mobile phones.
And yeah, so here are the results.
With Grammarly, sorry, I was talking about Grammarly.
It kind of gave me very limited feedback and then asked me to upgrade to get more feedback.
So I didn't do that.
I started with um, something called ieltscdtcom.
I can't remember what does cdt mean now computer delivered test maybe, i don't know.
Anyway, that was the first one that came up in google and i put this essay in it.
I think the essay could be a band nine on a test day in a real test because, as we know, um many of these model answers are written under you know optimal conditions.
They're they're not written under test conditions.
So in some ways when you see the band nine essays online they're not very realistic.
But if somebody could produce this essay in 40 minutes under test conditions, I think it probably would be a band nine or at least band eight.
And that's what I was expecting, really, from the writing checkers.
So the first one gave me a band seven.
Interestingly, it gave me 85 for task achievement, because I know that the essay answers the question without a doubt, but then gave me 65 for vocabulary, 6 for cohesion and 7 for grammar.
So, okay, it's an overall band 7.
It's not too far off, but I found that the feedback that I got from this checker was pretty useless, to be honest, because it's all about percentages.
It tells me usage of cohesive devices equals 7.55.
What does that mean?
How is that going to help me?
Logical progression equals 7.5 percent.
I just found it really unhelpful feedback and basically take it with a pinch of salt, as they say.
It's a very rough estimate and not very helpful.
The second one is ieltspodcast.com.
And I saw the exact same, it must be the same program with ieltsing.com.
So, I just tested the IELTS podcast one.
It gave me a very different set of figures and it was one band lower.
So this time it gave me overall band six.
And really interesting task achievement was seven.
It said that I didn't answer the question properly.
Vocab was six.
Cohesion and coherence 5.5.
I couldn't believe that because I thought five.
Actually, the cohesion and coherence was one of the best things about this essay.
And grammar, 6.5.
And range and accuracy, I know the accuracy is 100%.
So I'm not sure why it came out with that.
But it did give me very detailed feedback.
But I disagreed with all of them.
And if you go onto my blog, you'll see I've put a Google Doc link there where it just gives you more details, if you're interested.
I know it's not that fascinating for you, but it is for me, because I just need to know what these computer things are telling people.
And if you look at the IELTSpodcastcom, I do like the way it kind of underlines parts of the essay and you can hover over them and it tells you what's wrong when you hover over them.
The problem was that when I did that, the corrections were wrong.
So, for example, it didn't like the word many.
So, when I hovered over many, it said, instead, I should use a plethora of.
Now, I absolutely Absolutely can't abide that word, a plethora.
And I think it's really unnatural and I would never use it.
So I disagreed with that.
And then because the essay is about mobile phones.
It's very hard to avoid repeating mobile phones, and this is often the case with a key topic in the question.
If you're referring to that topic, many times it looks like it's repetitive, but sometimes you can't avoid that repetition.
There are no synonyms for mobile phones.
And the ones that they offered me were terrible.
They offered me dial, contact and call up.
You know again, it's a robot recognizing the word phone and thinking it could be a verb and saying oh, phone somebody is call up or contact somebody.
But it's not a replacement for mobile phones.
So that's a very typical AI problem.
And then the final thing.
This is a huge mistake that people make in IELTS and it's telling me it's the correct answer.
It wants to replace my simple but with on the contrary.
Now, on the contrary, go to my blog.
I've talked about this before.
It's not a synonym for but.
It's a different meaning.
And also grammatically, they made the wrong suggestion because they used a comma and it should be full stop.
You know, long story.
That's another thing.
But it was wrong.
All of the suggestions were wrong.
And I found that really disappointing, actually.
It gives you it tells you all the things you haven't done.
So it tells me that I haven't written an introduction or a conclusion, when it's pretty obvious that I have.
It tells me I haven't supported my main points with an explanation and then an example.
I absolutely have.
Anyway, so on.
Have a look at that Google Doc and you'll see what I mean.
And it says I haven't answered all parts of the question, but I have.
So I was a little bit disappointed with that, though I did like the way that.
Imagine if I did have real grammar mistakes I would be able to hover over and find out what the problem is.
It did offer, you know, I can upgrade again if I wanted more detailed feedback.
Also, I thought maybe then the score wasn't accurate because they told me that I didn't answer the question properly.
Well, if I didn't answer the question, that would give me five for task response.
But they gave me seven.
So there's something not right there.
And by the way, also, I noticed they say task achievement.
But the task two band descriptor is task response.
And there's a slight difference.
Task achievement is for task one and task response is for task two.
I don't know if they realize this.
Anyways, probably not important.
I use them interchangeably.
So the third one I tried was probably the worst one.
And it was called EssayChecker.ai and it was the least useful feedback.
And the only feedback it gave me was wrong.
So it said this essay has few spelling mistakes, few, which means not many.
But it's in red and it says this will lead to lower scoring.
So they mean that I have a few spelling mistakes which could lead to lower scoring.
Now, I know that there are no spelling mistakes in that essay.
I mean, there could be the American, British version of spelling.
Possibly, though I can't see an example.
So, yeah, this was the most useless and confusing.
Now, finally, chat GPT.
This is the one I thought that gave the best results.
And I think chat GPT is still learning.
So the more you tell it and the better your prompt. the better feedback it will give you.
So I'm sure this will be a really useful free tool in the future.
And this is probably the one I would recommend.
So.
Chat GPT gave me a 7.5.
Okay.
Task response was 8.
Lexical resource 7.
I'm a bit disappointed with that.
Cohesion 8.
Yeah, I agree.
Grammatical range and accuracy 7.
All right.
So, it's saying it's a pretty good pass, but not perfect.
It gave me feedback on each of the criteria and then I asked it for more detailed feedback, so you can, you can tell it what what you want.
So, for example, it said for vocabulary, it said however, there is some repetition of words and phrases throughout the essay, so I asked it to give me more information about that.
And it gave me much more detailed information.
And it was right.
It did say, OK, you repeated mobile phones too many times.
I didn't ask it to give me a suggestion to avoid repetition.
I should have done that.
But it was definitely the best and most useful feedback.
It did try to... Oh, it said some of the phrases were awkward, so I asked it to rephrase them.
And it did... rephrase my passive examples.
It didn't like the passive.
So, for example, I said before mobile phones were invented, the time spent away from work was our own.
Now, time spent, that's quite a sophisticated structure, even though it looks simple.
So, you might say the time which was spent But time spent is a high level band-aid structure.
It says you should rephrase this to before the invention of mobile phones, we had more control over our free time.
Now, that makes it a little less formal because you're using pronouns.
We had more control over.
Whereas if you chose the passive option, the time spent away from work was our own.
It's more like a formal noun phrase sentence.
So I disagreed with suggestions, you know, and that is usually the case with these AI things.
They don't like the passive tense and they try to rephrase them.
And so then this is interesting.
Then I asked it to rewrite it as a band nine criteria.
So I wanted to see if it thinks my essay is band 75, what do they?
What does it?
How could it improve it to make it band nine?
So it rewrote my essay. in 377 words, which is kind of too much, to be honest, or too many words.
It's way over length.
And it's almost saying well, the only way you could rewrite this is to make it over length and include a whole other paragraph.
Now I think realistically in 40 minutes this would be very difficult to achieve, though I did like some of the ways that it had rewritten my simpler essay and I think I could learn some language from it.
However, I noticed a lot of repetition in its Band 9 version.
So when I asked it then to assess its own band nine version guess what?
It gave its own version a score of band seven or eight and it said it has it needs a few minor improvements, particularly in terms of grammatical accuracy.
So I left it there.
I didn't ask why.
Perhaps I should have.
But of course, you can do that yourself.
Go and try chat GPT.
It is completely free.
And well, you can upgrade again, but the free version to me seems absolutely fine.
Go and try it and see what results you get and maybe compare it to to those other ones that i mentioned.
Um, if there are any that you've tried that you think might work well, i know there's one called write and improve from cambridge.
It's not free.
Then, you know, do let me know.
And if you go to my blog, i also made a video.
It was a video i made a couple of years ago when a student who was getting band 8 in everything kept getting band 65 in their writing.
And you know I have this kind of bet that I can tell you within the first two lines why you're not getting band 7.
And you can see what I do there in a really quick, just 10 minute video.
I point out about three key things that are stopping this student from getting band seven.
And, you know, sometimes it's like a light bulb moment for the student.
They'll say, well, I never realized that.
I just didn't know that.
And it can take two minutes to identify.
In this student's case, it was...
Well, there were linking word issues, especially with whereas, which is a really common mistake, and some kind of level of formality issues like asking the examiner a question.
You know the rhetorical question which I wouldn't advise for IELTS writing.
So go and have a watch of that video as well.
You know, you could put that essay into chat GPT and see what else that it might tell you.
I think it is getting closer to understanding what IELTS expects you to do in that very short time.
And I think you can do it for free.
So, yeah, great.
Do let me know how you get on.
And if you've got any other recommendations, please go to my blog and leave a comment and I will check them out and add it to the blog.
That would be really helpful if you could do that.
OK, so do remember that...
I do have a daily tips as well.
And this podcast is weekly.
And if you want anything in more detail, you want my direct help then do come and join the Members Academy.
When you just buy any single course, you'll also get other courses included, such as my 100 plus private blog.
And with lessons that you can search, they're searchable lessons.
And you'll also get a pronunciation course and a coaching, IELTS writing coaching course.
You'll get the live monthly member meetups with me where you can ask questions.
And weekly email and the daily tips.
So that's all included if you just buy one course and you get all of that for one year.
So if you're struggling with your writing or any part of the IELTS test, that would be a good option for you.
All right.
Thanks for listening today.
Let me know if you've got any questions.
Bye bye.
