Measuring reading time by characters is approaching the problem in the wrong way, perhaps.
Many readers read by "sight words": snapshotting an entire word in their mind and recognizing the whole of it as they read. In the USA, the Dolch and Fry sight word lists are taught, each with the most common 1000 words in the English language. This is why a word with 9 characters like "yesterday" (a sight word) will be read faster than a word with 6 characters like "abjure" (a non-sight word).
So I would prefer to measure words per minute (WPM) where sight words and uncommon words are counted alike as 1 word. And over a body of text, the law of large numbers works for you and you get a clear picture of how long the text takes to read.
My simple answer for selecting an average WPM, is to look at what smarter people than me have already done.
From Read time on Medium:
Read time is based on the average reading speed of an adult (roughly 265 WPM). We take the total word count of a post and translate it into minutes, with an adjustment made for images. For posts in Chinese, Japanese and Korean, it's a function of number of characters (500 characters/min) with an adjustment made for images.
If you must use characters to determine the size of words, know that the average amount of characters per word will vary depending on the audience the text is intended for. According to studies by Wylie Communications:
- The New York Times averages 4.9 characters per word.
- The Wall Street Journal averages 4.8 characters per word.
- The BBC averages 4.7 characters per word.
The Google books corpus averages 4.79 characters per word.
The problem with the most popular answer
This is to say that the calculation done (and the reasoning behind) the most popular answer above is incorrect. The author states:
The same Wikipedia article mentions that words-per-minute is not technically true, since a "word" is counted every 5 characters (including spaces and punctuation): For example, the phrase "I run" counts as one word, but "rhinoceros" and "let's talk" both count as two.
But this section of the Wikipedia article is under the "Alphanumeric entry" section, not the reading section! They are talking about how to measure keystrokes for data entry.
Is there a difference between 4.7 and 5 average characters per word? In some applications, yes. Wolf Garbe, founder of SeekStorm, optimized a spell checker by using the lower number:
The average word length in English language is 4.7 characters.(http://norvig.com/mayzner.html). If you use maximum edit distance=5 for spelling correction then the algorithm would return (too) many spelling correction suggestions, which have nothing in common with the input word.
Failures of previous studies
The Wikipedia article on Words per minute > Reading and comprehension is not a great resource for understanding the WPM of silent reading by a reader on a screen. Rather they discuss only 3 kinds of reading:
- Reading a text aloud, in which English-language readers scored 228±30 WPM.
- Proofreading text closely on paper, in which English-language readers scored 200 WPM.
- Proofreading text closely on old pre-1992 computer monitors, in which English-language readers scored 180 WPM.
Meta-analyses like "How many words do we read per minute? A review and meta-analysis of reading rate" often include numbers irrelevant to our use case. E.g. on paper, on old or low contrast monitors, reading aloud, proofreading, etc. Thus, they also "get it wrong" for us, arriving at 238 WPM.
Use words per minute instead of characters per minute. I would use 265 (as concluded by Medium, a modern application that has put in the research) and give time for images and scrolling behavior. I would lower the number if the text was more difficult or technical (though first I would try to simplify my text).
If you must convert words per minute to characters per minute, use a number like 4.7 average characters per word. Or better yet, measure your own average characters per word in examples of your own text. (And consider using median characters per word instead of average if you have a lot of text with a few outliers.)