Results

In testing their hypothesis that touch screen devices were superior to all other input devices in the cockpit environment, Stanton, Harvey, Plant and Bolton (2013) studied how these devices compared to each other in terms of performance and also how they induced perceptions of usability, discomfort and workload levels using both objective and subjective analysis. The input devices used for the study were a trackball, rotary control, touch pad and touch screen for two separate tasks i.e. Drop Down Menu Task (DDMT) and Action Menu Based Task (AMBT). The research also had a second hypothesis that auditory inputs positively aided results but this has been excluded in the analysis. This is due to the fact that the auditory inputs did not have significant effects on the majority of tasks and also because of the fact that the article did not account for the possibility of ‘practise effect’ confounding the results of the quantitative analysis.

Table 1 partly supports the hypothesis that touch screens induce better performances in terms of task times. All results were extremely significant at p<0.001.

Table 1. Task times (seconds) for various input devices for the separate tasks
Input Device	Task Time - DDMT		Task Time - AMBT
	Mean	StdDev	Mean	StdDev
Trackball	5.35	1.85	21.75	7.25
Rotary Controller	5.8	1.6	24.5	5
Touch Pad	5.45	1.35	26.25	13.25
Touch Screen	3.1	1.5	11.75	2.75

Table 2 shows the errors scores for both DDMT and AMBT with lower mean values indicating low error scores. The errors were negligible for the DDMT results and the statistical analysis showed no significance. For the AMBT, the rotary controller however offered the lowest error rate which does not support the hypothesis. All results for the AMBT tasks were extremely significant at p<0.001.

Table 2. Errors scores for various input devices for the separate tasks
Input Device	Errors - DDMT		Errors - AMBT
	Mean	*StdDev	Mean	*StdDev
Trackball	0.1	0.31	1.95	2.95
Rotary Controller	0.1	0.31	0.95	1.05
Touch Pad	0.055	0.23	6.3	12.2
Touch Screen	0.055	0.23	2.25	1.75

*Note: The StdDev values being higher than the mean in most situations were mainly due to negative values accorded to correct actions and positive values to errors. The extremely high StdDev value for the touch pad under the AMBT was also due to one particular participant making numerous errors.

Table 3 shows the median ratings, used popularly for subjective ratings, for the individual devices. The touch screen was perceived as highly functional as compared to the rest of the devices but also induced perceptions that it required higher workload and also contributed considerably to discomfort levels. This was mainly due to the fact that the touch screen was a direct input device requiring the participants to lean forward and hold their body and arm postures to manipulate the inputs on the computer screen. To a certain degree, the high scores in workload and discomfort rationalises the results in Table 2 where the touch screen did not support the hypothesis of being capable of encouraging superior performance in terms of error scores. However, the rotary controller was also not the most popular but did produce significantly lower error rates than the other devices in the AMBT.

Table 3. Median ratings of perceptions of various input devices
Input Device	Workload	Usability	Discomfort
	Range: 0 - 120	Range: 0 - 100	Range: 0 - 630
	Median	Median	Median
Trackball	54	74	6
Rotary Controller	54.5	70	20
Touch Pad	62.5	63.75	7
Touch Screen	65	82.5	120

[[/tab]]
[[tab Methods]]

Methods

Research approach

A mixture of both quantitative and qualitative research was used to determine how well participants performed in tasks and to understand their perceptions of the usability, workload and discomfort levels when using the different input devices.

Sample

A convenient sample of twenty participants, comprising 4 females and 16 males, were chosen from GE Aviation in the UK. They were aged between 21 and 57 years of age, with a mean age of 31 years (SD = 9.7). Four of the participants had some flying experience although none had professional licenses. All the participants were right-handed and their prior experience with the use of the input devices were varied with 70% of them having used the touch screen regularly while only 1 participant using the trackball and rotary control regularly.

Design and Procedure

A repeated measures design was employed for the quantitative part of the research and counterbalancing was introduced by varying the order of presentation for the different participants, with the exception of the auditory feedback which is not included in this analysis. For the qualitative assessment on workload, usability and discomfort, various questionnaires were administered to the participants for their opinion on the handling of the individual devices. Some of these questionnaires measured separate aspects e.g. ratings for hand and body discomfort while the workload measured ratings for six separate aspects ranging from physical and mental scores to even perceptions of frustration. The individual scores are ignored in this analysis but instead the overall scores for each attribute are presented.

A flight deck simulator was used but the DDMT and AMBT tasks did not require the participants to simulate any aircraft related manoeuvres. Instead the DDMT required the participants to navigate through a software menu using the different devices to replicate the actions of pilots in a modern flight management setting. The AMBT also used the computer but was more challenging in that five initial tasks required five more follow-on tasks to complete the whole AMBT task.

Variables

Disregarding the independent variable (IV) of the auditory feedback, Stanton et al. (2013) used two separate groups of IVs, specifically the input devices and the type of tests i.e. DDMT and AMBT. Eight conditions (4 devices x 2 task types) were then randomly administered to all participants to ascertain five dependent variables in task time, errors, workload, usability and discomfort. Prior experience in the usage of the devices may have confounded the results but the article briefly mentioned the practise sessions on the DDMT and AMBT tasks using all devices for all the participants, without specifying details like duration and desired proficiency over time.

Data analysis

Factorial repeated measures ANOVA (parametric statistics dealing with interval data) was used to analyse data on time and errors.
Friedman’s ANOVA (non-parametric statistics dealing with ordinal data) was used to analyse the subjective ratings on workload, usability and discomfort.

Generalization potential

Given the small sample size and for the fact that the participants were not pilots who could appreciate the association of the tests and perceptions with cockpit tasks, there may not be enough generalization potential. However, this study gives a basic understanding of different performance and perception levels of different devices which may then be expanded to study all aspects of work in the cockpit and how different devices can aid the individual aspects.

[[/tab]]
[[tab Reference]]

References

1. STANTON Neville A, HARVEY Catherine, PLANT Katherine L & BOLTON Luke (2013). To twist, roll, stroke or poke? A study of input devices for menu navigation in the cockpit. Ergonomics (ISSN 1366-5847) , 2013, volume 56, number 4, pages 590-611.

+++ Notes +++

2. Most of mean and median figures were translated from a line graph. Hence, there could be minor errors in the translation to a few decimal points.

3. The orginal article offers more information on the input devices, task types as well as the results for the quantitative and qualitative analysis.