Downloadable Files to Replicate Web-Reading Study

This page lists the original study materials from one of the experiments in a research project about how users read on the Web and how authors should write their Web pages. The studies were done in the SunSoft Science Office during the summer of 1997.

The following files are available:

Five different writing styles used to present the same information (note: you can also download all five versions, including all pages and illustrations, as a single 141K ZIP file):
1. Promotional writing style (the control condition in the study)
2. Concise writing style (50% word count of the control condition)
3. Scannable writing style
4. Objective writing style
5. Combined style (concise, scannable, and objective in a single site)
Questionnaire 1, given to the users after they had performed the first two directed tasks
Questionnaire 2, given to the users after they had performed an additional judgment task

These files are placed online in order to allow interested parties to replicate the experiment and/or revise or extend it with different pages and/or tasks. We hereby grant permission for you to download and use any or all of these files under the following conditions:

Any reports from your own studies should acknowledge that the original files were developed by John Morkes and Jakob Nielsen at Sun Microsystems in 1997
Any reports from your own studies should include a reference to our online writings about the project at this URL: http://www.useit.com/papers/webwriting/ (you can obviously include additional references if appropriate)
You can use the files in their original form or you can modify them in any way you please. Any reports from the studies you conduct based on these files should state whether you used the original files or modified ones
You agree that we are not liable for any problems resulting from your use of our materials: it is your own responsibility to verify the suitability of the files for whatever use you plan to make of them

Also, even though this is not a requirement, we would be grateful if you would please send us any papers or reports describing studies done based wholly or partly on our materials. Send to the address listed at the bottom of Jakob Nielsen's bio.

If you want to run your own studies with these files, you should download a complete set and put them on your own server. Please do not burden this server by using it as a resource in an experiment with multiple test users. You can download the entire set of five test sites as a single ZIP file (141K).

A disclaimer: The sample Web pages used in our study concerned travel to Nebraska. Even though we attempted to make the pages realistic, they do not attempt to be real travel information. Please do not base any real-life decisions or vacation planning on the information contained in our pages.

Also, please note that the HTML markup in these pages is not optimal. In our laboratory-based study, we knew exactly what computer and Web browser would be used, and we were not concerned with maintaining our small sample site beyond the duration of the study. Pages intended for commercial sites require more robust markup.

Sample Text from Rewriting Study

As a follow-up to our study on how people read on the Web, we rewrote real pages from a popular website according to our guidelines. The revised pages scored 159% better in overall usability than the original pages.

Please note that these experimental pages should not be relied on as a source of official information about Java: much has changed since the text was written.