A Study of Application and Device Effects Between a WAP Phone and a - - PowerPoint PPT Presentation

Mobile HCI '04 1

School of Informatics

Application & Device Effects

A Study of Application and Device Effects Between a WAP Phone and a Palm PDA

Jiraporn Buranatrived Paul Vickers

Mobile HCI '04 2

Application & Device Effects

School of Informatics

Mobile HCI '04 3

Application & Device Effects

School of Informatics

Background

• Growth of mobile device market
• More companies moving into mobile e-commerce,
• r m-commerce
• Design for mobility has several success factors

– Networking – Security – Application/ device mix – Usability

Mobile HCI '04 4

Application & Device Effects

School of Informatics

Obstacles to consumer adoption

PDAs Phones Obstacle

* source http://www.epaynews.com/statistics/mcommstats.html# 36

12% 10% Never heard of it before 13% 11% Other 16% 16% Don’t understand how it would work 31% 35% Fear of ‘klunky’ user experience 47% 52% Credit card security concerns

Mobile HCI '04 5

Application & Device Effects

School of Informatics

Constraints

• Screen size
• Memory capacity
• Processing power
• Input/output modalities
• Not as simple as squeezing existing applications

into a smaller GUI

Mobile HCI '04 6

Application & Device Effects

School of Informatics

Device heterogeneity

• Applications require multi-platform availability
• Devices have very different characteristics
• Different usability issues for phones, PDAs, pocket

PCs etc.

• Began to study two application types running on

two platforms

– Ticket purchasing & stock broking – Mobile phone and Palm OS PDA

Mobile HCI '04 7

Application & Device Effects

School of Informatics

Development

• Used Java 2 Micro Edition (J2ME) to develop two

applications

– Movie ticket purchasing

• Customer-led activity with server responding to client requests
• Supports localisation

– Stock broking

• In addition to customer-led model has real-time event-driven

aspect: customer reacts to stock price updates from server

• Simulated the wireless communication

– (even the network delays)

Mobile HCI '04 8

Application & Device Effects

School of Informatics

Study

• 16 subjects used both applications on both devices

(4 tasks)

• Performance measured by

– Time – Error rate

• Qualitative measures via a questionnaire
• Subject workload measured by NASA TLX

Mobile HCI '04 9

Application & Device Effects

School of Informatics

Task ordering

• Task 1: Phone broking
• Task 2: PDA broking
• Task 3: Phone ticketing
• Task 4: PDA ticketing

4, 2, 3, 1 B2 2, 4, 1, 3 B1 3, 1, 4, 2 A2 1, 2, 3, 4 A1 Task order Group

Mobile HCI '04 10

Application & Device Effects

School of Informatics

Procedure

• Subjects given a workbook

– Describing how to use the applications – Requirements for the 4 tasks – Task questionnaires – TLX rating sheets

Mobile HCI '04 11

Application & Device Effects

School of Informatics

Data collection

• Performance

– Time taken – Error rate

• Data used to calculate a task correctness score
• TLX used to collect workload data

Mobile HCI '04 12

Application & Device Effects

School of Informatics

Results

• Task duration

– Ticketing takes longer than stock purchasing, so cannot meaningfully compare times – Between devices the times were not significantly different for either application (F= 0.144, p> 0.05) – Times compare favourably with Fitt’s Law predictions except stock broking on the phone which took longer than predicted (t(15)= 2.265, p= < 0.05)

• Network delays, or problem with applying Fitt’s law to wireless

devices?

Mobile HCI '04 13

Application & Device Effects

School of Informatics

Results: task error

• Some difference in scores between both

application and device

– Two-way ANOVA – No difference in scores between application

• r device, and no

device/application interaction effect revealed

92.19 93.75 95.70 94.53 90.00 90.50 91.00 91.50 92.00 92.50 93.00 93.50 94.00 94.50 95.00 95.50 96.00 Mobile Broking Mobile Ticket Mean correctness (%) Mobile Phone PDA

Mobile HCI '04 14

Application & Device Effects

School of Informatics

Satisfaction

• Phone had higher mean

satisfaction rating

(not significant difference)

• Phone ticketing had

highest rating…

• …yet the lowest accuracy score
• Due to device familiarity?

78.19 70.69 74.58 71.39 66.00 68.00 70.00 72.00 74.00 76.00 78.00 80.00 Mobile Broking Mobile Ticket User's Satisfaction (%) Mobile Phone PDA

Mobile HCI '04 15

Application & Device Effects

School of Informatics

Ease of use and response

• Applications had significant ease-of-use score
• differences. Not affected by device
• Applications had significant response-satisfaction

score differences. No device effect.

• Suggests device may have less influence on user

than the application

– Get the app. right and any device will do? – Need more studies to explore this

Mobile HCI '04 16

Application & Device Effects

School of Informatics

Workload

• NASA TLX ratings:

– highly significant

application diff

(p< 0.01) – but not device

• Data entry on

broking slightly harder on phone?

16 16 16 16 N =

TicketPhone TicketPDA BrokingPhone BrokingPDA

100 80 60 40 20

Mean=43.13 Mean=48.44 Mean=29.44 Mean=25.44

* * * *

Mobile HCI '04 17

Application & Device Effects

School of Informatics

Usability in m-commerce

• Phones are common, but no difference found in

task duration, error rate, user satisfaction between devices

• Application differences seemed more important

than platform

• Need to explore more complex tasks to look for

device effects (e.g. list scrolling)

• Alternative representations for different displays?