Assessing writing skills – a response to “What Does It Mean to Know How to Write?”

Tom Johnson has sparked a lively debate with his blog post What Does It Mean to Write?. In the post, he wrote “It seems that writing is a spectrum skill”, providing a chart to demonstrate this:

In the post’s discussion thread, a consensus seems to have been reached that you cannot define writing skills and types of documents as a spectrum on a single line.

An alternative approach to assessing writing skills

I suggest the “writing spectrum” could be described more effectively, by using a radar or polar chart.

By using two or more axes, we can then start to differentiate between the different skills needed for a number of writing roles. For example, we could create a diagram of skills needed to create (a) persuasive, “selling”, marketing-type documents (b) educational, “telling”, technical-type documents (c) creative writing and (d) general business communication:

If we assess people’s writing skills against the same criterion, it’s likely we can get an idea as to which profession would best match their abilities.

What should go on the axes?

The key question is, what should be measured? Some initial thoughts are:

  • Structured and organised v. unstructured and disorganised
  • Emotional v. unemotional
  • Clear and understandable v. vague
  • Succinct v. flowery
  • Short document v. Long document

Another issue to bear in mind, is that the axes do not necessarily have to be positive/negative:

Potential measures could be: expression, adequacy of content, cohesion of information, compositional organisation and mechanical (grammatical) accuracy.

What do you think should go on axes?


Tom Johnson

I like it. The graphs you use make more sense, and I like your bulleted list of measurements. I can see how this topic might fit into recruiting of writers and placing them in professions that match their skillsets. Thanks for picking up on this thread and taking it to the next level.

Three Components of Writing Skill?

[…] of increasing writing skill. Lively discussion followed and Ellis Pratt responded with a blog post Assessing writing skills – a response to “What Does It Mean to Know How to Write?” which proposed that writing skills might be plotted on a two dimensional grid. Ever one to jump on […]


I think we’re almost saying the same thing with our posts. My suggestion of using a radar diagram means you can plot up to 7 or so different factors (= similar to your 3 dimensions).

I looked at some of the measures used elsewhere to measure writing ability (for example, some research on English carried out by a university in Nigeria as well as the GMAT test) and they tended to be:

Expression or style (not a key factor in technical writing),
Adequacy of content
Cohesion of information
Mechanical (grammatical) accuracy.

I believe that’s very similar to your three: composition, grammar and domain.

Two of my colleagues have a background teaching English as a foreign language (TEFL), and I’ll pick their brains when they return to the office next week.

I think it’s important to point out that we’re looking at writing skills only, and a Technical Writer’s skill set is more than that: they also need skills in planning, tools knowledge, domain knowledge etc. These would need to be assessed/mapped on a separate graph.

jay maechtlen

Good article. The radar plot is great for this, because each axis starts from zero at the middle. (could someone have a “negative” skill?)
Heh – it also shows us how well-rounded we are….

jay maechtlen

Well – in looking at your bullet points – most of them can be refactored into 0-10 scales
Structured and organised = 10
unstructured and disorganised = 0
“Emotional v. unemotional ” > ability to use proper emotional content/cues 0-10
Clear and understandable = 10
vague = 0
Succinct v. flowery > selection of proper level of brevity for target audience
Short document v. Long document > not a meaningfull measurement
* length of doc is a whole ‘nother set of discussions!



I started by looking to build on Tom’s writing spectrum, and whether a “one or the other”, positive/negative approach might work. I then concluded a radar graph might be a better approach. This approach allows you to assess a range of skills, rather than one or two. It also means you can rate people on a scale of 0 to 10, for example. So in the most latest suggestion, there is no negative scale.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.