Upload
ivan-pretel
View
286
Download
1
Embed Size (px)
DESCRIPTION
Most of the mobile phones have turned into full-connected devices. This provides companies with a perfect channel to interact with their potential clients and employees. The quality of the experience with these applications can directly affect the profits of the company it represents. Focusing on the mobile field and its extremely dynamic context, the quality of the experience can highly fluctuate. Inside this field, several methods and tools have been developed by defining a context of use. However, current methods can only capture it through adding external capture tools (added cameras, human observers...) that can change the experience. The main contribution in this article is a new approach to automatically measure effectiveness through a tiny but powerful mobile tool that can capture interaction metrics and the surrounding context without biasing the measured experience.
Citation preview
1/37
CLIHC 2013
DeustoTech - Deusto Institute of Technology, University of Deusto
http://www.morelab.deusto.es
December 6, 2013
Effectiveness Measurement Framework for Field-Based Experiments Focused on Android
Devices
Ivan Pretel, Ana B. Lago
CLIHC 2013
2/37
CLIHC 2013
Outline
Introduction
Usability, Quality and Effectiveness
Context model focused on Mobile Devices
Context model Capturer
System Validation
Conclusion
3/37
CLIHC 2013
Introduction
Usability, Quality and Effectiveness
Context model focused on Mobile Devices
Context model Capturer
System Validation
Conclusion
4/37
CLIHC 2013
5/37
CLIHC 2013
Introduction (I)
► According to the last Ericsson Mobility Report (November 2013)
►By Q3 2013, total mobile subscriptions were around 6.6 billion. By the end of 2019, they are expected to reach around 9.3 billion.
Introduction
6/37
CLIHC 2013
Introduction (II)
► Testing tools have changed and have been focusing on the web domain
► But, the web domain is focused not only on its own but also on the mobile one
► Mobile device -> More contexts than ever
► Issue
►The quality of mobile applications fluctuates depending on the context in which they are used
►When we try to capture it, we modify it
Introduction
7/37
CLIHC 2013
Introduction (III)
► Approach
► A mobile-based tool to automatically evaluate the effectiveness of interactions and capture metrics of the surrounding context without adding disturbing elements
Introduction
8/37
CLIHC 2013
Introduction
Usability, Quality and Effectiveness
Context model focused on Mobile Devices
Context model Capturer
System Validation
Conclusion
9/37
CLIHC 2013
Usability, Quality and Effectiveness (I)
► Usability
► ISO 9241-11
► “Usability is the extent to which a product can be used by specified users to achieve specified goals with effectiveness, efficiency and satisfaction in a specified context of use” – “Effectiveness is the level of accuracy and completeness with
which users achieve specified goals”
► ISO 9126-4
► “Usability is the capability of the software product to be understood, learned, used and attractive to the user, when is used under specified conditions”
– “Effectiveness is the capability of the software product to enable users to achieve specified goals with accuracy and completeness in a specified context of use”
Usability, Quality and Effectiveness
10/37
CLIHC 2013
Usability, Quality and Effectiveness (II)
► Quality according to ISO 9126 ► ”Represents a property of the software product defined in terms of a set of
interdependent attributes (usability, security, reliability, performance, complexity, readability, reusability) expressed at different levels of detail and also taken into account the particular context of use”
► Can be measured by 3 quality aspects
► Internal Quality
– From an internal view (e.g. spent resources)
►External Quality
– Its behaviour (e.g. number of wrong expected reactions)
►Quality in Use
– “The quality of software that user can perceive when the software is used in an explicit context of use”
– It is measured by 4 main capabilities
» Effectiveness, satisfaction, productivity and safety
Usability, Quality and Effectiveness
11/37
CLIHC 2013
Usability, Quality and Effectiveness (III)
► Effectiveness metrics (ISO/IEC 9126)
► Task Effectiveness
►What proportion of the goals is achieved?
– Measures the quantity of the goals achieved by a user
– 0<=TE<=1
► Task Completion
►What proportion of the tasks is completed?
– Measures the level of success the user achieves performing tasks
– 0<=TC<=1
► Error Frequency
►What is the frequency of errors?
– Measures the number of times that an error is made within a given period
– 0<=EF<=1
Usability, Quality and Effectiveness
12/37
CLIHC 2013
Introduction
Usability, Quality and Effectiveness
Context model focused on Mobile Devices
Context model Capturer
System Validation
Conclusion
13/37
CLIHC 2013
14/37
CLIHC 2013
Context model focused on Mobile Devices (I)
Context model focused on Mobile Devices
► It has been taken into account and has been defined several times
► Physical location
► Weather attributes
► Involved communities & stakeholders
► Goals
► …
► Tries to answer
► Who are the users?
► What are they doing?
► Where?
► Which resources are nearby?
15/37
CLIHC 2013
Context model focused on Mobile Devices (II)
► Nuevo modelo basado en tres pilares
Context model focused on Mobile Devices
► Context of Use model based on 3 main components
16/37
CLIHC 2013
Introduction
Usability, Quality and Effectiveness
Context model focused on Mobile Devices
Context model Capturer
System Validation
Conclusion
17/37
CLIHC 2013
18/37
CLIHC 2013
Context model Capturer - Lab vs Field (I)
Context model Capturer
► Lab-based
►Log information in a highly controlled environment using specific devices and users
– Control: high
– Context variables: limited
– Objectivity: low
► Field-based
►Log using added agents and invasive evaluation methods
– Control: very low
– Context variables: unlimited
– Objectivity: depends on the added agents
► Capturing tools
19/37
CLIHC 2013
Context model Capturer - Lab vs Field (II)
► If you do not want to influence the context, you have to remove all the biasing elements which can influence the captured data and use only the elements which is composing it
Capture using only the mobile device
► This tool should capture the context model via the built-in mobile sensors and logging interaction events
Context model Capturer
20/37
CLIHC 2013
Context model Capturer - Purposed System
Context model Capturer
► The purposed system is formed by a tiny Android library and a server to store and log the performed interactions
► 1. The app-to-test (ATT) should be integrated with the library
► 2. This library automatically captures context and interaction information and stores it in a local database
► 3. When the device has internet connection and its owner wants, all the information is uploaded
21/37
CLIHC 2013
Context model Capturer – Library (I)
► Interaction attributes
► Using one command ►log(event_type, interface, object)
Context model Capturer
22/37
CLIHC 2013
► Static attributes
► Captured during the registration
Context model Capturer – Library (II)
Context model Capturer
► Automatic ► Manual
23/37
CLIHC 2013
► Device Sensors
► Android API ► External Services
Context model Capturer – Library (III)
Context model Capturer
► Dynamic attributes
► Captured during the interaction (all automatically)
24/37
CLIHC 2013
Introduction
Usability, Quality and Effectiveness
Context model focused on Mobile Devices
Context model Capturer
System Validation
Conclusion
25/37
CLIHC 2013
System Validation – Experiment (I)
► 4 subjects
► A tiny memory game
► 1 day registering interactions
► Contexts
► At home [H]
► Public Transport [PT]
► Walking Down the street [C]
► At work (office) [T]
System Validation
26/37
CLIHC 2013
System Validation – Experiment (I)
► Steps ► Download the Android app
– From http://qualityinuseserver.appspot.com
► Sign up on the testing platform through the app
► Play the game
► Upload info to server
System Validation
27/37
CLIHC 2013
System Validation – Experiment (I)
► Focusing on the effectiveness measurement
► Task: To end up with all of the cards flipped face up in less than 15 turns
►More than 15 turns: task is not completed
► 8 pairs of cards
► It means the best round is made by only 8 interactions and 0 errors
System Validation
28/37
CLIHC 2013
System Validation
► Effectiveness metrics ► Task Effectiveness (TE)
►TE = 1 – ΣAi * error weight
►The weight of an error is 1/7 = 0.143
– Max number of errors is 6 with 8 correct interactions
– If you make 7 errors
» (1- (7*1/7) = 0)
► Task Completion (TC)
►TC = won rounds / played rounds
► Error Frequency (EF)
►EF = Number of errors / Total turns
System Validation
29/37
CLIHC 2013
System Validation – Results
► The outdoor contexts (PT and S) affect subjects and lead them to be less effective than in the indoor contexts (H and W).
► TE indicates that the S context is the context where subjects have the worst results
► The most efficient context is W, although EF exposes that is where subjects make more mistakes
System Validation
30/37
CLIHC 2013
Introduction
Usability, Quality and Effectiveness
Context model focused on Mobile Devices
Context model Capturer
System Validation
Conclusion
31/37
CLIHC 2013
Conclusion
► The effectiveness and the new context model, which can bias the usability and the quality of mobile applications, can be automatically captured and measured by the own mobile device minimizing the interaction bias
Conclusion
32/37
CLIHC 2013
Conclusion
► Found issues
► Security –android.permission.INTERNET android.permission.ACCESS_FINE_LOCATION
–android.permission.ACCESS_NETWORK_STATE android.permission.ACCESS_WIFI_STATE
–android.permission.MODIFY_AUDIO_SETTINGS android.permission.RECORD_AUDIO
–android.permission.WRITE_EXTERNAL_STORAGE
► Weather variables capture depends on the Internet connection
► Developers have to specify where are the errors
► Next steps
► Solve the found issues
► Increase the number of users
► Calculate correlations
Conclusion
33/37
CLIHC 2013
Thank you for your attention
¡Pura vida!
34/37
CLIHC 2013
DeustoTech - Deusto Institute of Technology, University of Deusto
http://www.morelab.deusto.es
Effectiveness Measurement Framework for Field-Based Experiments Focused on Android Devices
Iván Pretel
35/37
CLIHC 2013
All rights of images are reserved by the original owners*, the rest of the content is licensed under a
Creative Commons by-sa 3.0 license.
* • http://www.fotopedia.com/items/flickr-2830319467 • http://www.fotopedia.com/items/flickr-4609827016 • http://en.wikipedia.org/wiki/File:Skyline_oklahoma_city.JPG • http://www.flickr.com/photos/grimsanto/6190478735/ • http://commons.wikimedia.org/wiki/File%3AAstronomy_Amateur_3_V2.jpg • http://commons.wikimedia.org/wiki/File%3AAldrin_Apollo_11_cropped.jpg