Kolja Sam Pluemer

Paper summary - complexity in webpages (Harper, Michailidou & Stevens, 2009)

Summary, takeaways and learnings of the academical UX paper Toward a definition of visual complexity as an implicit measure of cognitive load


The last paper I summarized cited a paper by Harper, Michailidou and Stevens: Toward a definition of visual complexity as an implicit measure of cognitive load is an interesting approach to analyzing complexity in the web.

They basically got a bunch of webpages and a bunch of participants and ran some methods to find out what makes a given web page (not) complex. In a first step, they used Card Sorting and Triadic Elicitation. Never heard of Triadic Elicitation. Quite neat.

I wanted to link the Wikipedia article for Triadi…TE here, but it doesn’t exist. Looks like a to do. To explain, you let the user pick two out of three items that seem similar and thus different from the third, and let her explain why this be so.

A badly drawn comic Parasaurolophus sorting cards with Triadic Elicitation

Anywho. They did arrive at some criteria which makes a web page “simple”, “neutral” or “complex”. Methodically it sounds sound, although it is of course 2009 data and all that. These are my takeaways:

Simple Webpages

  • have one subject only
  • have less than 40 links (I am pretty sure the cutoff is somewhere else nowadays)
  • use neutral colors
  • have less of everything
  • require no scrolling (side note: Today “above the fold”, and thus “below the fold” is all the rage. When did this start? Is this useful?)
  • have no or one menu
  • use light backgrounds

Neutral Webpages

  • have much text, but it only flows in one direction
  • are a bit longer than simple ones
  • do not have prominent ads
  • use a few primary colors

Complex Webpages

  • require a lot of context switching
  • have long menus
  • have a lot of subject variety
  • use lots of different colors
  • have at least one form field (e.g. search)

Some of these are quite historic, some of these are quite timeless. And a good reminder. I liked the paper; two caveats: Age and quite small sample size.

Thanks for reading! This post is part of my series of reading and summarizing papers, mostly relating to UX. I use a casual tone, because that’s the most fun to me. That means my interpretation of a given paper may be off. Or incomplete. Or plain wrong. Always think for yourself, and for the love of God, don’t cite this in an academic context. Use the original article instead. Cheers!