Master's Thesis

Improving User Comprehension of Unfamiliar Concepts in a One-Time-Use Interface

A common saying is "people fear what they do not understand". The resolution to this problem is to improve the conveyance of information, and by extension, understanding. Yet, in our age of ever-present technology, consumers are presented with greater control of settings in the devices they use in everyday life, giving consumers more complexity to control without aides in understanding how to correctly make use it.

This project sought to determine what aspects of existing wireless router interfaces and configuration processes must improved to make a user more willing and able to use advanced features of their device, such as wireless encryption.

The experiment tested two hypothesis:

• Does providing increased depth of information about functions of the device affect how well a user understands their tasks?
• Does providing increased depth of information about the functions of the device affect the time taken to complete tasks with that device?

The experiment had two variables, with multiple levels:

• Instruction Format - All versions only contained the information relevant to the tasks. Other sections or topics were omitted. The instructions came in three formats:
  • The in-box documentation from the router manufacturer
  • Instructions rewritten by the researcher to use simpler language.
  • A custom-built goal-based router interface mockup incorporating all important settings and the re-written instructions relevant to that goal into a single screen.
• Pre-existing Settings - To see if interacting with pre-existing settings led or assisted users in completing their tasks, the router started with one of two states:
  • Factory Defaults
  • Some settings of the interface already held values, such as a wireless network name.

The experiment had 29 participants completing four tasks with a wireless router:

1. Add a passphrase and encryption to a wireless network
2. Change the administrator password
3. Restrict access to the router using a MAC address
4. Reset the router to factory default settings

After each task, participants rated their understanding of the task's instructions, understanding of the moderator prompt, and understanding of what the task asked of them. Participants also recorded whether they believed they had completed the task successfully. The researcher recorded the actual outcome of the task for comparison.

Two analyses were conducted, the first compared how well preconfiguration status and information format predicted understanding, using a discriminant analysis. The second sought to see if the same variables (when coded) were correlated with lower time on task.

Of all four tasks, only task 3 resulted in a good model of prediction for understanding, with Wilks' λ=.543 and a Χ²(2)=9.78, p<.01 . 84% of cases were classified correctly. There was not enough evidence in this experiment to reject the null hypothesis.

If you would like to read the entirety of the document, it is available below.