I’ve been writing on this blog since 2014, and in that time, I’ve written a lot, and improved substantially as a writer. I wondered if there were any quantitative metrics I could look at to identify trends in my writing in the more than 200,000 words I’ve written on this blog.
For today, I satisfied my curiosity by looking at a few simple aggregate numbers over time in my blog posts. I wrote a simple script in Ink to look at the median lengths of words, sentences, and paragraphs I published over the last seven years.
Trends, or the lack thereof
What surprised me the most was that my average word length in my writing has remained almost completely constant over the last seven years, at around 4.5 to 5 letters long. There’s a very very slight upward trend, but I don’t think it’s significant enough to note.
Apparently, even as my writing voice evolved and my vocabulary grew, the average size of the words I use didn’t change much.
By contrast, there’s a noticeable and significant trend in the length of my sentences over the years – they’ve gotten shorter, from about 22-24 words per sentence to somewhere in the upper teens.
This was a trend I expected to see, because during my first couple of years, I struggled with being overly verbose, and I never paid much attention to how difficult it would be to read what I wrote. As a result, I rarely revised specifically for sentence length. Lately, while I revise and proofread, I specifically look out for sentences that are too long or phrases that are too unwieldy, and the overall trend seems to show that it’s working. If sentence length is any indication of readability, my writing is getting easier to read, and that’s great.
The median paragraph length follows a similar trend, though for a different reason. The length of a median paragraph declined from routinely being 100-150 words long, to around 50 words per paragraph. When I first started writing, I wrote strictly following the “introduction-main points-conclusion” structure of an essay we learn as students in high school. As I found my voice and tried to write more readable essays, I structured my paragraphs less around large, high-level topics and more around specific ideas and arguments that are easier to anchor to visually, and easier to scan when scrolling, especially on a phone.
There are other statistics I think might be interesting to look at, like the distribution of word lengths and sentence complexity in my posts rather than just a single number in aggregate. But I think it would be most insightful to be able to study what I’ve written using more sophisticated NLP techniques that understand grammar and sentiment.
An incomplete list of things I might want to study further include…
- How repetitive I am in my writing, or alternatively, how many unique words I use
- The ratio of active to passive voice in my writing
- A more exact measure of sentence complexity, looking at things like the number of dependent clauses per sentence
- Ambiguity and specificity in my writing over time, i.e. how vague or precise I am
Hopefully, with more time and better tools in my arsenal, I’ll come back to this topic of studying my writing habits quantitatively and add any new insights I find here.
If you enjoyed this piece, you might also enjoy my next post, Finding it.
I share new posts like this on my newsletter. If you liked this post, you should consider joining the list.
Have a comment or response? You can email me.