The Impact of Language
on Website Use
and User Satisfaction
Anett Kralisch &
Veit Köppen
Institute of Information Systems
Motivation:
the Tower of Babel on the Internet
ENGLISH PORTUGUESE URDU FRENCH THAI QUECHUA 2 Institut für WirtschaftsinformatikMotivation:
the Tower of Babel on the Internet
• > 6000 languages spoken worldwide4
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
Motivation:
the Tower of Babel on the Internet
• > 6000 languages spoken worldwide
• > 6000 languages spoken by potential Internet users
• website translation / adaptations decisions often have significant financial ramificatiants
Motivation:
the Tower of Babel on the Internet
• > 6000 languages spoken worldwide• > 6000 languages spoken by potential Internet users
• website translation / adaptations decisions often have significant financial ramifications
• language as a barrier to accessing information / participation on the Internet (digital divide)
6
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
Motivation:
the Tower of Babel on the Internet
• > 6000 languages spoken worldwide
• > 6000 languages spoken by potential Internet users
• website translation / adaptations decisions often have significant financial ramifications
• language as a barrier to accessing information / participation on the Internet (digital divide)
Motivation:
the Tower of Babel on the Internet
• > 6000 languages spoken worldwide• > 6000 languages spoken by potential Internet users
• website translation / adaptations are often important investment decisions
• language as barrier to information access / citizen participation on the Internet (digital divide)
How does language affect website use ?
8
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
Motivation:
the Tower of Babel on the Internet
• > 6000 languages spoken worldwide
• > 6000 languages spoken by potential Internet users
• website translation / adaptations are often important investment decisions
• language as barrier to information access / citizen participation on the Internet (digital divide)
How does language affect website use ?
Which languages should be considered for translation ?
Agenda
1. The model
2. The variables
4. Preliminary Results
3. Data & Measures
5. Outlook
10
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
Agenda
1. The model
2. The variables
4. Preliminary Results
3. Data & Measures
5. Outlook
1. The model
Internet
Language Offer
Language
Proficiency Level
A Website‘s
perceived
net value
Website Use
.
Satisfaction
.
+
-+
+
12
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
Agenda
1. The model
2. The variables
4. Preliminary Results
3. Data & Measures
5. Outlook
2. The variables
Internet
Language Offer
Language
Proficiency Level
A Website‘s
perceived
net value
Website Use
.
Satisfaction
.
+
-+
+
14
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
2. The variables: net value
language-related
NET VALUE
of a website = of information gathering
VALUE
-
COSTS
2. The variables: antecedents
Internet
Language Offer
Language
Proficiency Level
A Website‘s
perceived
net value
Website Use
.
Satisfaction
.
+
-+
+
16
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
2. The variables: antecedents
2. The variables: antecedents
NET VALUE = VALUE - COSTS
18
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
2. The variables: antecedents
COGNITIVE EFFORT
• Revised-Hierarchy-Model (Dufour & Kroll 1995)
• higher cognitive effort required
for non-native language processing / at a lower language-proficiency level
NET VALUE = VALUE - COSTS
2. The variables: antecedents
COGNITIVE EFFORT
• Revised-Hierarchy-Model (Dufour & Kroll 1995) • higher cognitive effort
for non-native language processing / at a lower language-proficiency level
NET VALUE = VALUE - COSTS
TIME
• e.g. Hahne 2001
• slower language processing required in a non-native language / at a lower language-proficiency level
20
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
2. The variables: antecedents
COGNITIVE EFFORT
• Revised-Hierarchy-Model (Dufour & Kroll 1995) • higher cognitive effort
for non-native language processing / at a lower language-proficiency level
H1: A user‘s perception of a website‘s language-related net value increases with his/her language proficiency.
NET VALUE = VALUE - COSTS
TIME
• e.g. Hahne 2001
• slower language processing
in a non-native language / at a lower language-proficiency level
2. The variables: antecedents
NET VALUE = VALUE - COSTS
Internet Language Offer
• How much information is offered in the user‘s native language on the Internet?
Higher value
if few websites
are offered in that language
Lower value
if many websites
22
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
2. The variables: antecedents
H2: A user‘s perception of a website‘s language-related net value
decreases with an increase in the perceived amount of information offered on the Internet in the site‘s language.
NET VALUE = VALUE - COSTS
Internet Language Offer
• How much information is offered on the Internet in the user‘s native language?
Higher value
if few websites
are offered in that language
Lower value
if many websites
2. The variables
Internet
Language Offer
Language
Proficiency Level
A Website‘s
perceived
net value
Website Use
.
Satisfaction
.
+
-+
+
24
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
2. The variables: outcomes
Information Foraging Theory (Pirolli, Card 1995): • Link is followed if: costs < expected value.
2. The variables: outcomes
Information Foraging Theory (Pirolli, Card 1995): • Link is followed if: costs < expected value.
26
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
2. The variables: outcomes
Information Foraging Theory (Pirolli, Card 1995): • Link is followed if: costs < expected value.
Website Use:
• How many pages are accessed?
(User decides about following the link before the page is visited.)
2. The variables: outcomes
Information Foraging Theory (Pirolli, Card 1995): : • Link is followed if: costs < expected value.
Website Use:
• How many pages are accessed?
(User decides about following the link before the page is visited.)
• Which pages are accessed?
(User decides about the length of a page visit while the page is visited.)
28
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
2. The variables: outcomes
Information Foraging Theory (Pirolli, Card 1995): • Link is followed if: costs < expected value.
H3: The number of pages accessed by a user on a site increases with the perceived language-related net value of that site.
Website Use:
• How many pages are accessed?
(User decides about following the link before the page is visited.)
• Which pages are accessed?
(User decides about the lenght of a page visit while the page is visited.)
2. The variables: outcomes
Information Foraging Theory (Pirolli, Card 1995): • Link is followed if: costs < expected value.
H3: The number of pages accessed by a user on a site increases with the perceived language-related net value of that site.
Website Use:
• How many pages are accessed?
(User decides about following the link before the page is visited.)
• Which pages are accessed?
(User decides about the lenght of a page visit while the page is visited.)
30
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
2. The variables: satisfaction
Technology Acceptance Model (Davis 1993): • Perceived usefulness positive attitude
2. The variables: satisfaction
Technology Acceptance Model (Davis 1993): • Perceived usefulness positive attitude
32
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
2. The variables: outcomes
Technology Acceptance Model (Davis 1993): • Perceived usefulness positive attitude
H5: User satisfaction increases with perceived net value.
Agenda
1. The model
2. The variables
4. Preliminary Results
3. Data & Measures
34
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
3. Data & Measures
1. Questionnaire: all perceptual & attitudinal variables
• Multilingual questionnaire
• Likert or semantic differential scales
3. Data & Measures
1. Questionnaire: all perceptual & attitudinal variables • Multilingual questionnaire
• Likert or semantic differential scales:
2. Logfiles: extent of website use
• Linked to the questionnaire trough session-IDs
• IP-address ( geographical information)
• requested page, time of access, etc.
36
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
3. Data & Measures
1. Questionnaire: all perceptual & attitudinal variables • Multilingual questionnaire
• Likert or semantic differential scales: 2. Logfiles: extent of website use
• Linked to the questionnaire trough session-IDs • IP-address ( geographical information)
• requested page, time of access, etc.
3. Webpage analysis: simplified “depth“ of processing
• Self-developed software measures web page complexity
3. Data & Measures
1. Questionnaire: all perceptual & attitudinal variables • Multilingual questionnaire
• Likert or semantic differential scales: 2. Logfiles: extent of website use
• Linked to the questionnaire trough session-IDs • IP-address ( geographical information)
• requested page, time of access, etc.
1. Internet statistics: cross-check with self-reports
• Information about language offer on the Web
• e.g. Cyber Census Report; Mikami & Suzuki 2004; DMOZ.org
World Statistics
3. Webpage analysis: simplified “depth“ of processing • Self-developed software measures web page complexity
WEBPAGE
OTHER
SOURCES
38
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
3. Data & Measures
1. Questionnaire: all perceptual & attitudinal variables • Multilingual questionnaire
• Likert or semantic differential scales: 2. Logfiles: extent of website use
• Linked to the questionnaire trough session-IDs • IP-address ( geographical information)
• requested page, time of access, etc.
1. Internet statistics: cross-check with self-reports • Information about language offer on the Web
• e.g. Cyber Census Report; Mikami & Suzuki 2004; DMOZ.org World Statistics
3. Webpage analysis: simplified “depth“ of processing • Self-developed software measures web page complexity
2. Language Status: cross-check with self-reports
• Which official language(s) in which region?
• Self-developed database
WEBPAGE
OTHER
SOURCES
Agenda
1. The model
2. The variables
4. Preliminary Results
3. Data & Measures
5. Outlook
40
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
4. Preliminary Results
• data from a pretest-questionnaire • simplified data analysis
• patients only (due to significant correlations between language and patients/physicians) • N=83
4. Preliminary Results
• data from a pretest-questionnaire • simplified data analysis
• patients only (due to significant correlations between language and patients/physicians) • N=83
Preliminary results: no significant impact
Website Use
.
Language Proficiency Level
H3 H1 A Website‘s perceived net value H5 H2
4. Preliminary Results
42
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
Tendencies:
Internet Language Offer Language Proficiency Level
A Website‘s
perceived net value
Website Use
.
Satisfaction .
β = 0.182 (p=0.163)
4. Preliminary Results
Internet Language Offer Language Proficiency Level
A Website‘s
perceived net value
Website Use
.
Satisfaction .
β = 0.146 (p=0.462)
β = 0.162 (p=0.417) Tendencies:
44
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
4. Preliminary Results
Internet Language Offer Language Proficiency Level
A Website‘s
perceived net value
Website Use
.
Satisfaction .
β = 0.286 (p=0.151) Tendencies:
4. Preliminary Results
Internet Language Offer Language Proficiency Level
A Website‘s
perceived net value
Website Use
.
Satisfaction .+
+
+
+
-Tendencies:
46
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
4. Preliminary Results
Internet Language Offer Language Proficiency Level
A Website‘s
perceived net value
Website Use
.
Satisfaction .+
+
+
+
-Website culturally adapted to the major language groups?
Agenda
1. The model
2. The variables
4. Preliminary Results
3. Data & Measures
48
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005
Outlook
• Final questionnaire with revised questions is currently posted on the website. • Logfiles are registered at the same time.
• Validation of Internet statistics and language status data for predicting website use and satisfaction.
Outlook
• Final questionnaire with revised questions is currently posted on the website. • Logfiles are registered at the same time.
• Validation of Internet statistics and language status data for predicting website use and satisfaction.
Future questions to answer:
• The role of language associated values (e.g. Dmoch, 1997)
e.g. French – elegance; English - modernity
• The role of the number of languages on one website. • Personal language aspects (language fear / motivation).
Institut für Wirtschaftsinformatik
Humboldt-Universität zu Berlin Kralisch & Köppen 2005