Performance errors that affect motor impaired users on input with keyboard and mouse have been studied for a long time. In particular, Trewin and Pain have conducted an empirical study and have identified several keyboard and mouse errors due to motor disabilities. For example, a motor impaired user tend to press a key too long that causes the key to repeat. This is called long key press error. However, whether these errors established by Trewin and Pain also affect mobile users is still unkown. Therefore, this study reuses Trewin and Pain’s methodology with mobile users to investigate whether they experience same problems with motor impaired users.

Research Questions

With this study we aim to answer the following questions:

  1. Does long key press error affect mobile users?
  2. Does simultaneous key error affect mobile users?
  3. Does additional key error affect mobile users?
  4. Does bounce error affect mobile users?
  5. Does pointing error affect mobile users?
  6. Does clicking error affect mobile users?
  7. Does dragging error affect mobile users?

Participants

Instead of motor impaired users, we invite able-bodied mobile users (smart phones or PDAs). There are no limitations on gender, age and previous experiences. Participants are recruited through personal contacts and advertisements. All participants are unpaid volunteers.

Apparatus

In Trewin and Pain’s experiment, participants used ClarisWorks word processor. As this software is not available on mobile devices, we cannot use that software in our study. Therefore, as an alternative we use a MediaWiki editing page served from our Web server. Participants access the page from a HP PDA, which is provided by us.

In Trewin and Pain’s experiment, they used an InputLogger software to record a user’s key presses and mouse events. However, this software cannot be used on mobile devices. Therefore, we use the UsaProxy which sits between the web server and a client browser and records key presses and mouse events.

Tasks

In this study, participants are asked to complete the following five tasks:

Typing task
In this task, participants are asked to enter a given text passage into a text area on the experiment webpage.
Pointing task
This follows the typing task, and participants are presented another webpage where he/she needs to perform pointing, clicking and dragging using a PDA stylus.
Editing task
In this task, participants need to follow instructions and edit a given text passage, using typing and pointing skills tested in previous tasks.
Repeat pointing task
This is a repeat of the pointing task.
Repeat typing task
This is a repeat of the typing task.
After finishing all tasks, participants are asked to give feedback and comments about the completed tasks.

Data

Following data are collected during the experiment:

  1. A log file of input events produced by the UsaProxy.
  2. Video recordings of hand movements from three different angles (left, right, front) of the participants performing the tasks.
  3. Observations made by the experimenter.
  4. Demographic information about the participants.

Data Analysis

We adopt Trewin and Pain’s analysis methodology. The overall analysis covers: the ease at which the devices can be used, measured through participants’ feedbacks; and the errors occurring through difficulties in using the devices, which are recorded by the log file and videos, and examined in terms of frequency of occurrence and the timing information associated with them. Errors are manually annotated in the log files. These annotations indicate the type of error, differentiating between classes of performance error, and placing all other errors into a single class.

Ethics Application forms

This experiment has been approved by the Senate Committee on the Ethics of Research on Human Beings (study reference number 07188). To get this ethical approval, we have submitted the following documents to the Senate Committee on the Ethics of Research on Human Beings:

Results

Further information about this study and the results can be found in our technical report called “Rerunning Trewin and Pain’s experiment with mobile users”.

Press release about this study also gives a broad overview of our findings. This study was also on the news: