Results

Online time is a measure of how long a subject spent online, using the MLSS system, in working on the task. This was measured in minutes using the MLSS system's built-in log file system, and was the total time spent online working on the task.

As clearly shown in the above table (see Table 1), there were significant differences between subjects of differing skill levels and interaction modes. Both of these were highly significant at p < .01. Command language interaction by novices clearly took the longest time (216.37), followed by command language by experienced users (137.78). However, there was little difference between the results across skill levels, for users of the wizard (67.09 for novice and 67.45 for experienced).

Table 1: GLM Analysis of Performance/Online Time: Means by Condition (Minutes)

Symbol	Condition	Novice	Experienced	All Conditions
CMD	COMMAND	216.37	137.78	177.55
WIZ	WIZARD	67.09	67.45	67.27
	All Subjects	141.70	103.12	122.41
GLM Results			Significance
Model	F =99.90	p = .000	[a]
SKILL	F = 18.97	p = .000	[a]
MODE	F = 154.96	p = .000	[a]
SKILL [b] MODE	F=19.33	p = .000	[a]
** = Significant at p< .05
[a]Significant at p< .01 [b]Significant at p < .1

System satisfaction is a measure of how satisfied subjects were with the system they used, which in this case was the MLSS software. System satisfaction was measured using a composite system satisfaction scale. The questionnaires and scales used in this research were generally consistent with those used in past NJIT research, which include both research projects and dissertation research (Benbunan, 1997; Dufner, 1995; Fjermestad, 1995; Ocker, 1995).

The results, shown in Table 2, indicate there were significant differences between subjects of differing interaction modes (p < .01), but there were no significant differences between skill levels (p > .1). There was a significant difference when the interaction of skill level and interaction mode was considered (p < .01). In general, wizard users showed a higher level of system satisfaction, and while novice users preferred the wizard, experienced users preferred the command language. It appears, however there is a greater difference for the novice users, as compared with the experienced users.

Table 2: GLM Analysis of System Satisfaction: Means by Condition

Symbol	Condition	Novice	Experienced	All Conditions
CMD	COMMAND	24.24	18.85	21.55
WIZ	WIZARD	15.34	19.67	17.51
	All Subjects	19.72	19.26	19.53
GLM Results			Significance
Model	F = 238.20	p = .000	[a]
SKILL	F = .497	p = .481
MODE	F = 29.01	p= .000	[a]
SKILL [b] MODE	F= 41.88	p= .000	[a]
** = Significant at p< .05
Note: Smaller values indicate higher system satisfaction
[a]Significant at p< .01 [b]Significant at p < .1