Extensive reading and reading speed

[pinion]

Long time lurker, first time poster...thought this data might be of interest to some of you. The graph below shows my increasing reading speed over the course of about 15.6 million characters read between December 2018 and July 2020 (so just over a year and a half).

 

 

Some notes:

• Reading time includes time spent looking up unknown words in Pleco's document reader, creating Pleco flashcards, and googling unknown references, plus a little occasional texting. 
• Most of what I read was webnovels, with a few real books thrown in here and there. Other than starting out with a webnovel that I'd heard was easy, I didn't make much of an attempt to filter for difficulty.
• When I started, I'd learned around 1600 characters (recognition only), but I'm a heritage speaker, so my vocabulary was probably somewhat larger than that of a second-language learner who knows an equal number of characters. At this point I'd say I recognize 4000 characters or so.
• I read roughly the first 2.5 million characters either fully out loud or muttered under my breath, and switched to reading silently only when reading out loud began to noticeably slow me down. I still tend to semi-voluntarily mouth the words when reading something unfamiliar or difficult.

 

I hope this is helpful for someone, as my small attempt to give back after all the time I've spent reading the massive amount of accumulated wisdom on these forums.

[roddy]

Thanks! Do you feel like you're still speeding up, or have you reached a level you're happy with?

[i__forget]

How many hours per day did you spend doing this, and when you say heritage speaker do you mean that you could actually speak but not read?

[pinion]

@roddy: Well, the data stops in July 2020 because that's when I switched over to working on traditional characters, but I don't think I'd hit a plateau yet at that point. There's definitely been a significant drop in speed after switching to traditional, but I expect it'll catch up at some point--though maybe not as quickly as I'd like, given that I've recently shifted more of my time/attention to speaking. Ideally I'd like to read as fast as an adult educated in Chinese, but I don't know how achievable that will ultimately be. At any rate, the reason I started studying Chinese in the first place was to read things anyway, so I'm just going to continue to read, and I expect that as I do so my reading speed will continue to improve (albeit increasingly slowly).

 

@i__forget: My goal during this period was three hours a day, which I probably met more often than not (although things definitely fell off after the pandemic started). My speaking when I started was broken at best, but I already had good listening comprehension for everyday conversation, so I certainly wasn't starting from zero.

[Jan Finster]

On 10/11/2020 at 11:52 PM, pinion said:

about 15.6 million characters read between December 2018 and July 2020 (so just over a year and a half).

 

10 million characters a year is an insane amount for someone "learning" Chinese? ?

[大块头]

Congratulations on your progress! You appear to have reached a near-native reading speed. I've been collecting a similar dataset, and I hope to have a graph like yours someday.

1. Do the points in the plot represent your reading rate for random sections of text from the novels you were reading?
2. Would you mind sharing a spreadsheet or csv file of your measurements? When I have time I'd like to try fitting a couple models to these data.
3. Consider sharing this dataset with someone researching extensive reading. As I understand it, all studies in the literature only measure reading rate at two points in time (e.g. the start and end of a semester). The high time resolution of this dataset is unique.

[alantin]

Wow!

An amazing chart! And for me extremely motivating!

 

By your description I’m about where you were when you began your extensive reading on broken speaking, ok listening and about 1600 characters. Now I’m about a month and 200 000 characters in and I measured my reading speed on a couple of different days at around 60 - 80 characters per minute. I’m looking forward to similar  results!

 

did you notice any effects on your speaking and/or listening skills that you can attribute to the extensive reading?

[Sreeni]

On 10/12/2020 at 5:52 AM, pinion said:

The graph below shows my increasing reading speed over the course of about 15.6 million characters read between December 2018 and July 2020 (so just over a year and a half).

@pinion what is your

 

1. vocabulary recognition Count

2. word count for which you know meaning and pronunciation

 

why you want to measure your reading speed rather Vocabulary Count? Is it to know comprehension level? Reading speed?

[alantin]

Can’t speak for @pinionbut I think reading speed correlates with comprehension level and character and vocabulary recognition. It is also a lot easier to measure than these things.

 

it would be interesting to see how these correlate.

 

Mine is currently around 80 chars/minute and I can understand the overall meaning of a passage in the books I read but there are a lot of words that I don’t know already. The already known words are a breeze to read but my speed drops drastically when I encounter unknown words even if I can guess their pronunciation and meaning from the characters.

 

My character recognition is generally between 90-95 %. I don’t know about the word recognition.

[Jan Finster]

15 hours ago, Sreeni said:

why you want to measure your reading speed rather Vocabulary Count?

 

It gives you an idea of your "sight vocabulary" [words you recognise instantly and therefore do not slow you down when reading]. Imron has posted about how to improve your reading speed. It involved re-reading text passages in order to increase your sight vocabulary.

[pinion]

@大块头 Apologies for the extremely belated reply!

On 10/17/2020 at 6:41 AM, 大块头 said:

Do the points in the plot represent your reading rate for random sections of text from the novels you were reading?

They represent the total number of characters in the text divided over the total time spent reading--so, for instance, it took me exactly 9.5 hours to read the 86,214 characters in 活著, giving an average of 151 characters per minute.

 

On 10/17/2020 at 6:41 AM, 大块头 said:

Would you mind sharing a spreadsheet or csv file of your measurements? When I have time I'd like to try fitting a couple models to these data.

Sure! I've attached the file to this post. Your models look very interesting (and much more sophisticated than the linear regression I tried at some point)--I'm excited to see your results.

 

On 10/17/2020 at 6:41 AM, 大块头 said:

Consider sharing this dataset with someone researching extensive reading. As I understand it, all studies in the literature only measure reading rate at two points in time (e.g. the start and end of a semester). The high time resolution of this dataset is unique.

Haha, I'm happy to share my data with anyone interested, but I suspect it won't be very useful to researchers, who are usually trying to figure out how to teach students who aren't obsessive readers...

閱讀.xlsx

[pinion]

@alantin

 

Along with extensive reading I also did extensive listening, albeit in much less systematic fashion, so it's impossible to separate the effects of the two. That said, they were also virtually the only forms of study I did (other than flashcarding characters), and I successfully passed TOCFL 6 (listening/reading) at the end of last year, so...I think it's fair to say that they were extremely beneficial.

 

As for speaking--I didn't practice speaking at all until relatively recently, but after starting I've progressed very very quickly, which I do attribute to the effects of massive input. That said, I was fluent as a child, so I didn't need to practice pronunciation/tones. I will note that when reading, I make a point of subvocalizing every character, which I think helps strengthen the sound/meaning connection and thus improves both listening and speaking.

[pinion]

@Sreeni

 

On 3/18/2021 at 10:20 AM, Sreeni said:

1. vocabulary recognition Count

2. word count for which you know meaning and pronunciation

I do not and have never tracked vocabulary count.

 

On 3/18/2021 at 10:20 AM, Sreeni said:

why you want to measure your reading speed rather Vocabulary Count? Is it to know comprehension level? Reading speed?

@alantin and @Jan Finster are both right--change in reading speed is relatively easy to track over time, and I think it's as good a measure as any of overall reading effectiveness. Also, I read very fast in English and found it very frustrating how slowly I read in Chinese at the start, so it was nice to have concrete evidence of progress.

[Sreeni]

you are currently @151 CPM on average  and that is very good progress considering difficulties in learning Chinese. @pinion ???

[Jan Finster]

18 minutes ago, Sreeni said:

you are currently @151 CPM on average  and that is very good progress

Reading speed is also related to the difficulty of your text. I can read HSK 2 texts at 150 CPM or more, but I may read a complicated philosophy text in my native language at 300 words per hour.  So, I guess any comment on reading speed should be related to the level and type of material you are reading.

[pinion]

@Sreeni 151 CPM was just an example data point from fairly early on. The latest data point on the graph is 261 CPM  

 

@Jan Finster Yes, absolutely! And also I still find with harder books that the second half goes much faster than the first.

[Sreeni]

@pinion @Jan Finster  Thanks, Understood. I was using “ How to read faster and better by Norman Lewis and initially my English reading speed was 170 wpm and it increased to 250-350 WPM. I thought that speed should be fine and I never proceed further to improve. That book has some texts and author mentions the number of words in that text.

 

How are you accurately counting characters per minute for Chinese? Is this approximate count?and  Chinese characters generally vary in strokes from 3-10. How is your experience? What is major issue you encountered with Chinese speed reading and how you resolved or overcome?

[pinion]

@Sreeni

On 3/23/2021 at 11:01 PM, Sreeni said:

How are you accurately counting characters per minute for Chinese? Is this approximate count?

Most of my reading has been using txt files, for which any text editor can give an exact character count, and I timed every reading session for the given data points, so the characters per minute stats are as precise as reasonably possible.

 

On 3/23/2021 at 11:01 PM, Sreeni said:

Chinese characters generally vary in strokes from 3-10. How is your experience?

I haven't found stroke count to make much of a difference for pure recognition. (Writing by hand is of course another matter.)

 

On 3/23/2021 at 11:01 PM, Sreeni said:

What is major issue you encountered with Chinese speed reading and how you resolved or overcome?

I think speed reading issues are really just the same as reading issues in general--character and vocabulary recognition, word boundary parsing, etc. Unfortunately I don't have much in the way of specific strategies to offer; my position in general is that sufficiently massive exposure fixes everything.

[micky056]

How many hours per day did you spend doing this? I am new to novel arena, how much time I have to spent like how long it will take for me to improve my reading speed? Sometimes I just get stuck in a random word and waste's my time. How to avoid these kind of barriers ?

[pinion]

16 hours ago, micky056 said:

How many hours per day did you spend doing this?

I tried to spend three hours a day on reading. I didn't record all the time I spent on this, but what I did record amounts to about 1200 hours.

 

16 hours ago, micky056 said:

Sometimes I just get stuck in a random word and waste's my time. How to avoid these kind of barriers ?

I did the vast majority of my reading in Pleco's document reader, which is a paid add-on that I think is well worth it. Any time I encountered an unfamiliar word, I just tapped on it to see the pop-up dictionary definition. I know others on this forum feel that this makes looking up familiar words too easy to a degree that is counterproductive, but I think the advantages outweigh the disadvantages because it allows you to get through a greater volume of material much more quickly. Your mileage may vary.