• No results found

Chapter 11 Discussion

4.5 Presentation of results

• All error bars represent one standard deviation • All statistical tests are two-tailed

Chapter 5

Experiment One

5.1 Introduction

Currently, firefighters use paper floor-plans to gain an understanding of the layout of buildings during incidents. This floor-plan information is available for larger

premises. Interviews conducted in Chapter 3 suggest that often this information is not referred to. Information is often is too complicated to rapidly inform navigation sometimes required at the start of an incident. This experiment compares paper floor-plan information with three other methods of presenting the same floor-plan information on a mobile device. The methods of presentation are general enough to be employed on a variety of novel devices but for the purposes of the experiment, a PDA is used to display the floor-plan data. The spatial information was presented in four different ways, summarised in Table 5.1.

Table 5.1 Summary of the different methods of presentation

Condition Paper Plan Plan Overview Allocentric Segment Egocentric Segment Overview of area or segment:

Overview Overview Segment Segment

In the paper plan condition, participants are presented with an entire plan of the area. In the remaining three of the methods: plan overview, allocentric segment and egocentric segment, involved presenting the spatial information electronically on a PDA. Egocentric information is presented to participants track-up on the first

presentation of the map. Initial presentation of allocentric information was in a fixed, north-up orientation, participants then rotated the device to their preferred

orientation. The limited screen size (77mm × 58mm) of the PDA means that it is not possible to display the full extent of the plan on the PDA screen. Two methods were

used in order to communicate the information to participants. In one method, the entire map was available to participants to move around within the screen (plan overview condition). The second type of presentation divided the map into route segments; participants could not choose which parts of the map to use to navigate in these conditions since the information displayed was under the control of the device. In the overview conditions, participants could view any part of the plan as they wished Zimmer (2004) suggests that humans are able to navigate from, and recall good overall spatial information from segments of maps. In their study, spatial knowledge gained from consecutive presentation of map segments was as effective when informing navigation as overall information being presented. This finding is relevant to the present study since presenting less information by using small segments is an effective way of overcoming limited screen sizes, especially if the formation of spatial knowledge is not impaired when using this method of

presentation Burnett and Lee (2005) found that the formation of a cognitive map from fragmented information found in in-car navigation systems is impaired.

However, since the focus of this experiment is navigation rather than memory of the route this is an effective way of presenting the information. Unlike the previous method, participants will not have access to the full plan, rather a static plan that refers to the immediate environment. Aretz and Wickens (1992) found that track-up, or egocentric presentation of map information reduced workload when navigating. Seager and Stanton-Fraser (2007) also replicated this finding when a map showing a small segment of the environment, as in this experiment, is automatically rotated. in order to test these findings in the current context both initial egocentric and allocentric presentation of the segments will be used in the experiment. Participants were required to navigate a route of 120.7m around the Sir Clive

of navigation support. The floor plan and route are shown in Figure 5.1. The route was restricted to a single storey. Three representations were displayed on the PDA. Figure 5.1 Plan of Sir Clive Granger building with route shown as a continuous line

(Not to scale)

5.2 Experimental Hypotheses

• Different representations will result in differences in route completion time • Different representations will result in differences in workload experienced • Different representations will generate different experiences of navigation • There will be relationships between route completion time and workload

experienced

• There will be relationships between route completion time and subjective measures

5.3 Design

Four experimental conditions were developed to test the experimental hypotheses. The study is a full, between groups design. The conditions are summarised in Figure 5.2. The first method of presentation was the paper plan. The second, electronic, method of presentation was to present the entire plan in an image viewer (Figure 5.3). Participants were required to use the PDA stylus to move parts of the floor plan into view on the screen depending on where they were in the route. In the third and fourth methods, the floor plan was split into parts and these segments were

presented to participants, in sequence, on the screen at the appropriate location (Figure 5.4). The segment method of presentation was split into two further groups. In the allocentric segment condition, the plan maintained the same initial orientation throughout the route. In the egocentric segment condition, the segment was rotated relative to the position of the user, equivalent to rotating a map automatically. Figure 5.2 Experimental conditions

Condition Paper Plan Plan Overview Allocentric Segment Egocentric Segment Overview of area or segment:

Overview Overview Segment Segment Orientation Assistance: Interaction: Any available to participant. Rotating of plan allowed Moving image around the screen using a finger. Rotation of device allowed None available: Rotating of device allowed None available: Rotating of device allowed

Figure 5.3 Participant using the PDA in the plan overview condition

Figure 5.4 Participant using the PDA in a segment condition

Figure 5.5 Participants using the navigation support in each of the four conditions

5.3.1 Dependent variables

Core dependent variables used in this experiment are shown in Table 5.2. Full details of these dependent variables can be found in the experimental methodology (Chapter 4, p100).

Table 5.2 Core dependent variables used in experiment one

Dependent Variable Measurement Performance Route completion time (seconds) Individual Differences

i. Sense-of-Direction Scale & ii. Object Perspective taking Test

Workload NASA-TLX (Scales only)

Experience Integrated Navigation Questionnaire (INQ)

Behaviour Behaviour checklist

5.3.2 Power analysis

Forty-eight participants were used in the study. This corresponds to twelve participants per experimental condition. No literature that could contribute to a realistic calculation of power or inform selection of one of Cohen’s standard effect sizes is available. Indeed, this experiment will produce data that will inform power calculations in future experiments.

5.4 Method

5.4.1 Participants

Forty-eight participants completed the study. Participants were drawn from staff and students at the University of Nottingham. Data from four participants was excluded from the analysis since these participants did not appear to understand the

and reported normal or corrected to normal vision. Table 5.3 and Table 5.4 show full details of participant’s age and familiarity with the experimental environment.

Table 5.3 Age, in years, of participants

Age (SD)

Condition Paper Plan Plan Overview Allocentric Segment Egocentric Segment Male 22.4 (2.4) 26.3 (8.7) 29.1 (12.2) 25.2 (6.4) Female 25.6 (8.8) 21.8 (6.7) 28.6 (8.6) 24.6 (5.7) Total 24.3 (6.8) 24.1 (7.8) 28.9 (10.4) 24.9 (5.8)

Table 5.4 Familiarity with experimental environment

Familiarity (SD)

Condition Paper Plan Plan Overview Allocentric Segment Egocentric Segment Male 1.8 (0.4) 2 (0.0) 1.4 (0.8) 1.5 (0.5) Female 1.1 (0.4) 1.5 (0.8) 1.4 (0.5) 1.5 (0.8) Total 1.4 (0.5) 1.8 (0.6) 1.4 (0.7) 1.5 (0.7)

5.4.2 Materials

For each condition, navigation support was developed in line with the experimental treatments.

Presentation of plan data

Paper Plan

A laminated plan was given to participants at the start of the route.

Plan Overview

A digital image of the plan was displayed, full screen on a Dell Axim™ X51v PDA. Resco® image-viewer software was used to display the image. The full image is too large to be displayed on the screen so participants interacted with the device by moving the image using their finger on the touch-sensitive screen.

Plan Segments

Version 1 of the software is used to present the plan segments. Full details of the software are discussed in Chapter 4. A screenshot of the type of display

experienced by participants is shown in. Figure 5.6 The division of the area into plan segments is shown in Figure 5.8.

Figure 5.6 Screenshot of the type of display used

Route Selection

A route in the Sir Clive Granger Building, University of Nottingham was selected for the experiment. The route was selected since it contains several 90O turns and as such is not straightforward to navigate or recall. Also, the route mainly consists of corridors which contain few distinguishing features or distractions. On a practical level, the route also avoids areas which become crowded by students such as the cafeteria or the lecture theatres. Figure 5.8 shows how these segments ‘mesh’ together if they were viewed as a whole. Figure 5.7 shows examples of the route used in the study.

Figure 5.7 Examples of the route navigated

Figure 5.8 Composite of plan segments arranged in order

5.4.3 Procedure

Figure 5.9 shows a flow diagram of the procedure used in experiment one. The description of specific parts of the procedure are discussed below.

Participants gave written, informed consent in a room away from the route taken in the experiment. Participants then were required to submit personal data and

complete the complete the Santa-Barbara Sense of direction questionnaire and the Object Perspective-Taking Test.

Participants completing trials in the PDA conditions then had the opportunity to use the PDA and practice with the interface they would be using with stimuli comparable to, but not actually used in the experiment.

When participants indicated that they were comfortable interacting with the PDA, they were led to the starting point of the experiment. The route taken to the starting point was deliberately selected so that no part of the experimental route would be either traversed or visible to the participant.

The participant was then handed the PDA or paper plan and instructed to start when ready. Timing was started by the experimenter. At specific points on the route, the experimenter would instruct the participant to advance the map by tapping the ‘Next’ button. This instruction was also given in the paper plan condition to maintain

experimental control, but in this condition no action was taken by the participant. The experimenter counted instances of observed behaviour (see Section 4.2.7, p114) on a form during the experiment.

When participants completed the route the timing was stopped by the experimenter and participants were led into another room, adjacent to the completion point. Participants then completed the NASA-TLX workload questionnaire and the Navigation questionnaire. Whilst participants were engaged in these tasks the experimenter updated behavioural notes and noted anything unusual about the trial. Participants then indicated their level of familiarity with maps and map reading and their employment. Participants were then debriefed and arrangements for payment were put in place.