21.07.2013 Views

User Interface Design and Ergonomics - National Open University of ...

User Interface Design and Ergonomics - National Open University of ...

User Interface Design and Ergonomics - National Open University of ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

you use. But then you need some way to compare the numbers you get for one design<br />

with the numbers from the others.<br />

The simplest approach to use is called a BETWEEN-GROUPS EXPERIMENT. You use<br />

two groups <strong>of</strong> test users, one <strong>of</strong> which uses version A <strong>of</strong> the system <strong>and</strong> the other version<br />

B. What you want to know is whether the typical value for version A is likely to differ<br />

from the typical value for version B, <strong>and</strong> by how much. Here's a cookbook procedure for<br />

this.<br />

Using parts <strong>of</strong> the cookbook method above, compute the means for the two groups<br />

separately. Also compute their st<strong>and</strong>ard deviations. Call the results ma, mb, sa, sb. You'll<br />

also need to have na <strong>and</strong> nb, the number <strong>of</strong> test users in each group (usually you'll try to<br />

make these the same, but they don't have to be.)<br />

Combine sa <strong>and</strong> sb to get an estimate <strong>of</strong> how variable the whole scene is, by computing<br />

s = sqrt( ( na*(sa**2) + nb*(sb**2) ) / (na + nb - 2) )<br />

("*" represents multiplication; "sa**2" means "sa squared").<br />

Compute a combined st<strong>and</strong>ard error:<br />

se = s * sqrt(1/na + 1/nb)<br />

Your range <strong>of</strong> typical values for the difference between version A <strong>and</strong> version B is now:<br />

ma - mb plus-or-minus 2*se<br />

Another approach you might consider is a WITHIN-GROUPS EXPERIMENT. Here you<br />

use only one group <strong>of</strong> test users <strong>and</strong> you get each <strong>of</strong> them to use both versions <strong>of</strong> the<br />

system. This brings with it some headaches. You obviously can't use the same tasks for<br />

the two versions, since doing a task the second time would be different from doing it the<br />

first time, <strong>and</strong> you have to worry about who uses which system first, because there might<br />

be some advantage or disadvantage in being the system someone tries first. There are<br />

ways around these problems, but they aren't simple. They work best for very simple tasks<br />

about which there are not much to learn. You might want to use this approach if you were<br />

comparing two low-level interaction techniques, for example. You can learn more about<br />

the within-groups approach from any st<strong>and</strong>ard text on experimental psychology.<br />

3.1.7 DETAILS OF SETTING UP A USABILITY STUDY<br />

The description <strong>of</strong> user testing given up to this point should be all the background you<br />

need during the early phases <strong>of</strong> a task-centered design project. When you are actually<br />

ready to evaluate a version <strong>of</strong> the design with users, you will have to consider some <strong>of</strong> the<br />

finer details <strong>of</strong> setting up <strong>and</strong> running the tests. This section, which you may want to skip<br />

on the first reading <strong>of</strong> the chapter, will help with many <strong>of</strong> those details.<br />

186

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!