How many subjects for Web Usability? For: Practical orientation – Against: Statistical significance

The question, how many subjects a result meaningful and representative is, face many entrepreneurs.

preached at the Web Usability, that a few experts and volunteers are adequately, to improve the usability. Although this is true, there are still skeptics, are only the opinion of a very high number of people could be "true" statements about the usability.

Especially when considering web usability is the choice of methods qualitativ before quantitatively.

It is true, that a small number of results for a statistical analysis of course is not usable. Especially for the Web usability is in most cases not necessary.

[This is of course different for example in the usability of a Passierflugzeugcockpits. Clear, since then life will depend on other people from, intuitive to use as a system for individual pilots.]

A high level of web usability to ensure a user experience. Nutzerfreundlich, effective, and efficiently should be sites. With the demand for this 80 realizing percent, have the practical methods enforced. This is based on a simple efficiency calculation.

For one thing is clear, dass qualitative Tests mit z.B. 500 People for most companies remain unaffordable. Therefore remains the comparison of the online questionnaires (quantitatively) with a continued focus on usability testing (qualitatively).

On the other hand it is also clear, of a potential problem not only 500 Time must be described, to be classified as such.

Online questionnaires are used to answering basic research questions. They have a high range and can also set standards. For usability testing, they represent a high financial cost in the construction. At least, if the test criteria are met. This guarantee, however, in the first place, a statistical analysis.

The goal in practice is primarily potential Problems can quickly find and fix. The statistical significance ( is less important.

Most companies do not claim to predict scientifically, how about some random problems in the total user base will occur exactly.

It is rather in practical qualitative tests Experts use the right tasks and questioning techniques.

There have been about in practice. 8-12 People for a homogeneous target group proved to be useful.

"Bumpy" sample:

Imagine, Watch until a young man and later an older woman, as they stumble at the entrance of a store.

Do you think the shop owner is curious to know, how many people will stumble probably due to a specific cause? Or he tries to find out, why these two are tripped?

First and foremost want to be a turnover of interested shop owners recognize the cause, order to eliminate as soon as possible. (Presumably he would also like both. However, this is again dependent on the effort to resolve the problem)

8-12 People and repeated testing is a gute solution. It will not necessarily 500 People needed.

Or wie many people would watch if you stumble, a exclude a "random" stumbling, accepted, it it were your business?

Leave a Reply

Your email address will not be published. Required fields are marked *

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