Windows 95 Menu Styles:
Windows Explorer vs. My Computer

Experiment

Introduction and Hypothesis

This experiment evaluates the three major navigation methods provided by the Windows 95 operating system – Windows Explorer, My Computer with a single window, and My Computer with cascading windows. We hypothesize that My Computer with a single window will yield the quickest results, and also rank as the most preferable in the subjective ratings.

In this 1x3 experiment, the independent variable is the navigation method, while the dependent variables are time and the subjective ratings.  There are three treatments, one for each of the independent variables.

Windows Explorer displays a tree hierarchy for all the drives, directories and subdirectories in the system. Although all the drives and directories can be seen at once in the left frame, only the contents of the selected one can be viewed at any given time, displayed in the right frame.  Plus signs (+) next to a drive or directory in the tree indicate there are subdirectories inside, while minus signs (-) indicate the presence of only files or symbolic links (shortcuts) or nothing at all. Double clicking on an item with a plus sign expands the tree to show its subdirectories. When the breadth of the tree has grown too wide to be viewable in its entirety in the left frame, the scrollbars appear. Double clicking on a drive or directory in either frame will display its contents in the right frame. Dragging an item from either frame to a drive or directory in either frame will move it there. Highlighting the item to be moved in the open directory, selecting 'Cut' or 'Copy' from 'Edit' on the menu bar, opening the source directory and then selecting 'Paste' from 'Edit' is another method. Right clicking on an item will also bring up a menu with cut, copy and paste options directly at the cursor's location. In Windows 95, creating a symbolic link is referred to creating a shortcut. In menus, any time the 'Paste' option appears, 'Paste Shortcut' appears as well.

My Computer with a single window does not show the hierarchical tree, but instead starts with the all-encompassing My Computer.  Double clicking on each directory will display its contents, but still only the contents of the open one can be viewed at any given time. Dragging works, but the entire tree is not viewable. Cutting, copying and pasting is the only way to move something to a parent directory. This can be accomplished with 'Edit' on the menu bar or the right click menu.

My Computer with cascading windows begins again with My Computer, but each time a directory is double clicked on, a new window automatically pops up.  Though this allows multiple directories to be displayed, it can quickly clutter the screen and make navigation confusing. Both copying, cutting and pasting, as well as dragging an item from one window to another are viable methods of moving items. For the former method, both 'Edit' on the menu bar and right clicking are available options.

In our preliminary assessment, we tested two subjects. Each subject ran all three of our treatments: Windows Explorer, My Computer with cascading windows, and My Computer with a single window. Each subject was given a maximum of five minutes to complete one treatment, but the actual times lasted no more than three minutes. Both subjects were able to finish the entire experiment in less than ten minutes. Below is a chart of the times:  

 

	Windows Explorer	My Computer Single Windows	My Computer Cascading Windows
Subject	Time (seconds)	Time (seconds)	Time (seconds)
1	106	93	115
2	119	107	134

One subject was very familiar with Windows 95, and instead of using the 'Edit' on the menu bar to cut copy and paste, used the shortcut keys (CTRL-X, CTRL-C, and CTRL-V), making the task considerably easier. We then decided that in order to make the tests more uniform, trials would be conducted with only a mouse, instead of both a keyboard and mouse so that shortcut keys would not be available. Since the times were so close, we also decided to increase both the number of tasks and the depth of navigation.

25 subjects were tested for this experiment, but one subject's results were disqualified (see Discussion).  All subjects were undergraduate and graduate students at the University of Maryland at College Park. All the subjects had used Windows 95 before, with skill level ranging from beginner to expert. They all knew what cutting, copying, and pasting meant, as well as how to use a mouse to drag, double click, and right click.

Materials

For this experiment we used a Pentium 200 MHz computer with 32 MB of RAM running the Windows 95 operating system from Microsoft.  The computer had a standard 2-button mouse, a standard PC-compatible keyboard and a 15-inch monitor operating at 800x600 resolution.

Three different task lists were used.  Each task list required the subject to navigate through a directory structure of at least 5 levels. Tasks include searching for a particular file, copying it into another directory, or moving in into another directory.  To ensure subjects were not getting faster by repeating the same task over, there were three different task lists, each coupled randomly to one of the treatments.  A typical task list can be found in Appendix A: Task List.

Since some users have had previous experience with Windows 95, they may have already developed a preference for one form of navigation over the others.  In addition, though one navigation style may prove more timely, it may not be as aesthetically pleasing.  For example, some users may not mind having 10 windows open on the screen at once, but others may find it too cluttered.  Therefore, a subjective rating was requested after the experiment.  The survey can be viewed in Appendix C: Survey.
 

The experiments were conducted in the OACS (Office of Academic Computing Services) at the Lefrak Building at the University of Maryland, College Park. When the subjects arrived, they were given brief instructions not to use the keyboard and how to read a directory path (c:\temp). The subjects were told they were being timed. Subjects were not allowed to see their task lists until timing began.

Each subject was given one of three randomly selected task lists to begin with.  There were three different task lists, but each had the same number of tasks and same number of directory levels to traverse.  The subject was then asked to select a number between 1 and 3, each of which corresponded to one of the treatments.  This was done in an effort to randomize the experiment.  Basically, each task list (1,2,3) was used and each treatment was run (A,B,C). Subject X may have run task list 1 with treatment A, 2 with B, and 3 with C, while subject Y may have run task list 2 with A, 3 with B, and 1 with C. No subject used the same task list twice.

The time it took for the subject to complete the task list was recorded. When the subject completed all three treatments, he/she was given the survey to fill out. The survey was untimed.
 

 Two experimenters were present at each test. One explained the instructions and performed brief demonstrations.  The other experimenter presented the tasks and recorded times. While the subjects filled out the questionnaire, the experimenters gave them an unlimited amount of time.

Department of Computer Sciences Direct questions and comments to the student editorial team