Capitalization and Comma Errors

Here are the answers to the "Common Mistakes" handout this week. Corrections are mostly in red:

1. I saw beautiful coral reefs and yellow, red, and blue tropical fish.

2. I like to watch CNN nNews on YoutTube.

3. I saw a short video about the origin of the universe on the TED tTalk app yesterday.

4. I like the characters, especially Elsa, and songs, such as 'Let it gGo.'

5. I watched famous comedy movies, such as "Breakfast at Tiffany's.”. (The period goes inside the quotation marks)

6. So, I watch entertaining movies, especially heroic movies or adventure movies, like 007.

7. Shuri-jyo was built in the 1400’s.

8. She was a faster swimmer than I was, and she inspired me to practice hard every week.

9. People often make mistakes, so it is necessary to double check reports.

10. We practiced throwing, running, and tackling everyday after school.

11. So, I agree with Francisco.

12. Also, I did yoga and swam, like freestyle and breastroke, at the fitness club.

Academic Discussion task 12

 I just took the Academic Discussion 12 writing test. This time, I wanted to organize my essay a little differently (we'll talk about this in class this week). Since this is scored by the AI, I can't be sure that you'll get similar results with the human rater. 

Results were typical: 7 minutes, 141 words, 5.0 score.

Academic Discussion #2 (again): another experiment

 Hans did another experiment on ETS's practice academic discussion #2. 

This has 120 words, and he deliberately tried to write in generalizations, with little concrete detail. E-rater gave it 4.0.

Academic Discussion 12

The content of this essay is deliberately bad. The question was the best way to evaluate teachers. I wrote that the best teachers allow their students to play video games instead of studying. Perfect grammar, dumb content. You can read it for yourself. 

155 words, 9 minutes, 5.0.

So you can fool e-rater with nonsense. The AI is not evaluating content, only grammar.

 

Academic Discussion 11

Here's my academic discussion essay for task 11. There's nothing surprising here. 153 words, 6 minutes, 5.0 score. I think I've got a pretty good handle on what the AI is looking for: you need 130+ words with few misspellings and grammar errors for a top score.

iBT Academic Discussion Writing Task (a summary)

 So I've written essays on the first ten topics that ETS has posted online (link: https://www.ets.org/toefl/ibt-enhancements/prep.html). Here are some results:

1. Every essay that has been 130+ words and grammatically accurate has scored 5.0.
2. One essay at 132 words, that had at least nine grammar mistakes and eight spelling mistakes got 4.0.
3. Essay 5 had 119 words and no grammar mistakes and got 5.0.
4. Essay 8 had 121 words and no grammar mistakes and got 4.0.
5. Every essay that had fewer than 119 words has scored 4.0.

Two easy generalizations:

1. You want to write more than 130 words in your essay to get 5.0 from e-rater.
2. You can't make too many grammar and spelling mistakes.

Next question: is e-rater assessing vocabulary level? 

• If you write 130+ words with only one and two syllable words, can you still get 5.0? 
• If you write 100 words and use a lot of long and academic vocabulary, can you get 5.0?

We'll find out soon.

Task 10 + grammar errors

 This time I took the basic content of the previous task 10 essay and re-wrote it, adding grammar and spelling mistakes. Five minutes, 132 words. I count at least nine grammar mistakes and eight misspellings.

Score: 4.0

Task 10

 Here's my task 10 essay. Six minutes to write and two minutes to get the word count down to 132 words.

Score: 5.0.

So it appears that you need 130+ words to get 5.0, with no grammar mistakes. Let check grammar next.

Task 9

 Here's my task nine essay. It took five minutes, is 120 words long, and has no grammar mistakes. 

Score: 4.0.

I guess when ETS says, "An effective response will contain at least 100 words," their definition of "effective" is 4.0.

Academic Discussion: Essay 8

OK, some interesting results.

Essay 5: 119 words, score: 5.0.

Essay 8: 121 words, score: 4.0.

When I look at these two essays, I don't see a difference in grammatical complexity or vocabulary level. I wrote in a conversation style. Maybe tomorrow I'll try to use more academic vocabulary.

Here's the essay:

Academic Discussion: Topic 7

 

115 words, no grammar mistakes, 4.0 again. So when ETS says, "An effective response will contain at least 100 words," they are considering 4.0 to be a good score.

Keep in mind, essay 5 was 119 words, no grammar mistakes, and got 5.0.

So we're recommending at TOEFL test takers aim for at least 120 words.

Academic Discussion: task 1

 Hans did an experiment on task one (how to attract people to rural areas) from ETS. He found that it took 16 grammar and spelling mistakes to lower an essay from 4.0 (110 words) to 3.0.

This is 4.0:

This is 3.0:

So yes, e-rater is evaluating grammar!

Academic Discussion: essay 6

 

In essays 5 and 6, I didn't mention the two other students. Essay 5 got 5.0 with 119 words (1 paragraph), and essay 6 got 4.0 with 111 words (2 paragraphs). There were no grammar mistakes in either essay.

I'm not deliberately trying to use academic vocabulary, but it would interesting to see what would happen if I limit myself to short, non-academic vocabulary.

Also, ETS has a new YouTube video on the Academic Discussion task:

ETS's descriptions of the scoring system for the academic discussion is better than it was for the independent essay, and much better than the speaking rubrics, but this topic deserves a blog post of its own. You can download the pdf of the writing task rubrics here.

Academic Discussion: Essay 5

 

This is my one paragraph essay for question 5. It took about four minutes to write, then another minute to edit it down to under 120 words (to be precise, 119). It was scored 5.0.

So essay 4 was 107 words and got 4.0. And essay 5 was 119 and got 5.0. I'll write the next one in 110 and see what happens.

This also suggests that ETS isn't necessarily looking for multiple paragraph essays.

Academic Discussion: 4

 Here's an essay I wrote for the fourth task (gap year). It was just 107 words (7 minutes), and got 4.0, so I'm pretty sure e-rater is valuing word count.

Based on the first four essays, somewhere between 107 and 130 words is the cut-off for a 5.0 essay.

Academic Discussion: task 3

 Here's the essay that I wrote for the third academic discussion. You can find them at https://www.ets.org/toefl/ibt-enhancements/prep.html

I wrote 173 words (7 minutes), so I needed to take 2 photos of the screen.

Notes:

I tried to "contribute to the discussion" by acknowledging Kelly and Paul's contributions, but also saying that water and air pollution aren't big problems in Tokyo.

Then in paragraph two, I explained that climate change (global warming) is a big problem in Tokyo. Anyone who was here 30 years ago knows that it's gotten much warmer, and summers can get dangerously hot.

Finally, at the end of paragraph two, I suggested that scientists should try to pull CO2 out of the air to solve/ameliorate the problem.

This one got 5.0 too. I'm going to try some shorter essays to test the recommended 100 words.

My school, E4TG, will begin offering TOEFL Writing July 31, 2023.

Academic writing: task 2

 Here's the essay I wrote for the second task published by ETS at https://www.ets.org/toefl/ibt-enhancements/prep.html

There is a first paragraph that isn't shown in the text box. This one also got 5.0.

Actually, this is the second essay that I wrote for this task. The first one was only 104 words, and got 4.0 despite being grammatically perfect. This suggests that ETS is weighting the length of the essay.

Academic Discussion Writing: First Experience

 So I just did the first academic discussion task from ETS.

Here's a picture of my essay:

You can see that I wrote 130 words in a little over five minutes. Here is a picture of the score:

And I got 5.0. So what does my experience suggest?

• ETS is obviously using e-rater or some other AI for scoring. It took about 10 seconds to score the essay.
• I did mention both students (Kelly and Claire) in my answer. Not sure if this is necessary for a top score.
• On re-reading my essay, I realize that it's not grammatically perfect. The 3rd sentence should contain "there are." The final sentence of the first paragraph should include "in" the countryside.
• The organizing strategy was to summarize Claire and Kelly's points, but I also tried to "add to the discussion" with personal details (I live in Tokyo, which is dealing with the same problem)
• I also tried to "add to the discussion" by adding to Kelly's points (concerts, sporting events, transportation).
• After summarizing and "adding the the discussion" in paragraph one, I directly answered the question with my own opinion in paragraph two. The only new detail in paragraph two is "creating jobs outside agriculture." All the other paragraph two details are rooted in paragraph one.