Creating Low-Code Loan Applications using the Trisotech Mortgage Feature Set
Satisfaction measurements in a cross-cultural environment: How valid are subjective judgements? Presentation Statistische Woche 2011
1. University of Cologne
Center for Evaluation
Faculty of Management, Economics and Social
Sciences
Dr. Christian Bosau
Satisfaction measurements in a
cross-cultural environment:
How valid are subjective
judgements?
2. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 2
Background of the study
Are we really measuring „true“ satisfation?
Job satisfaction (JS) in organizations:
§ keyword: international employee satisfaction survey
§ the absolute level of JS often has direct consequences for leaders
§ intercultural measurement problems often are NOT considered
Main question: How can we compare the results of JS across national
and cultural borders?
3. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 3
Cross-cultural JS-results
8,4 8,4
8,2
8,1
8,0 7,9 7,9 7,8 7,8 7,7 7,7 7,7 7,7 7,7 7,6 7,6 7,6 7,6 7,6 7,6 7,5 7,5 7,5 7,5
7,3 7,3 7,2
7,1 7,1 7,1 7,1 7,1 7,0 7,0 7,0 6,9 6,9 6,9
6,7 6,7
6,6 6,6 6,6 6,5
6,2
6,1
5,9
5,5
5
6
7
8
9
Switzerland
Malta
Denmark
Norway
Canada
Iceland
Ireland
Austria
USA
Sweden
Belgium
NorthIreland
Mexico
Japan
Argentina
Chile
Finland
Netherlands
Brazil
Luxembourg
Poland
Nigeria
Portugal
UnitedKingdom
Italy
Germany-West
Slovenia
CzechRepublic
Hungary
SouthAfrica-White
Germany-East
Spain
India
China
Slovakia
France
Lithuania
Greece
Croatia
Estonia
Rumania
Latvia
Bulgaria
Korea
Russia
Turkey
Ukraine
Belarus
• results from World-Values-Survey:
• very often: deskriptive intercultural results are published
(Slocum & Topichak, 1972; Lincoln et al, 1981; Griffeth & Hom, 1987; Near & Rechner, 1993; Chiu & Kosinski, 1999; van de Vliert &
Janssen, 2002; Llorente & Macias, 2005)
Unclear: Can we really just interpret those descriptive differences of JS directly?
4. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 4
Problems of cross-cultural research
response styles and culture:
§ different cultures show different response styles, e.g. acquiescence (ARS)
(Hui & Triandis, 1989; Johnson et al., 1997; Chen et al.; 1995; Takahashi et al., 2002; Harzing, 2006)
§ different response style should be understood as different communicational behaviour, not
just methodological bias (Smith, 2004; Smith & Fischer R., 2006)
standardising of measures (see Fischer, R.; 2004):
with-in-subject, group centering, grand mean centering
problems:
§ measures are not independent from each other anymore
§ absolute level of measures is lost
§ interpretation is possible only in relation to standardising value (group or grand mean, etc.)
What can we do about it?
Note: Especially problematic if we want to compare absolute measurements,
like JS-levels of countries/subsidiaries/etc.!
5. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 5
Examples:
- Does well-being increase with higher GDP?
- The more interesting a job is, the higher the
JS is?
- etc.
Methodological classificiation
level oriented structure orientedvs.
Examples:
- How high is JS in different countries?
- In which countries are people more satisfied
with their lives?
- Are Americans more extroverted than
Germans?
- Is the image of my product better in
Germany compared to Spain?
- etc.
construct bias method biasvs. item biasvs.
different understanding of
construct across cultural
boundaries
bad item translation or
culturally inapropriate wording
different response styles
6. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 6
The 3 important constructs
What is the relationship looking like between these constructs?
response style
culture
job satisfaction level
? ?
?
-
individualism
acquiescent response style
Known: Negative relationship between individualism and acquiescent
response style (ARS)
7. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 7
Method - secondary analysis
§ Data from World-Values-Survey & Eurobarometer
§ operationalisation of acquiescence response style (see Harzing, 2006) :
acquiescence-index 5-point-scale: ALL items, having received 5 or 4
MINUS ALL items, having received 2 or 1
§ aggregation to national level
§ correlation of national means (controlled for soziodemographic differences of nations, i.e. age,
gender)
Important question:
How will the JS-measurement be influenced by response styles?
8. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 8
Nation-level-results - 1
results:
§ job satisfaction and ARS are indeed correlated: BUT negatively.
§ i.e., high measures of job satisfaction only occur in countries/cultures with small or no ARS
response tendency norms
Keeping in mind: in all countries respondents are on average satisfied with their jobs
§ Note: if ARS exist, respondents do not report dissatisfaction but only lower satisfaction
§ Fits with the self-construal results of Markus & Kitayama (1991): interdependent/collectivist
people – having higher ARS – are restrained in telling their personal feelings.
-.13 (14)
-.68** (11)
-.59*** (16)
-.71*** (16)
-.68*** (16)
Job Satisfaction
(nation-level-mean from
Eurobarometer)1
-.29 (25)ARS-index from Hofstede (2001)
-.47** (23)ARS-index from Harzing (2006)
-.38*** (47)
ARS-index 3 from WVS
(4-point-scale ‚agree/disagree‘-label)
-.47*** (42)
ARS-index 2 from WVS
(5-point-scale ‚important/unimportant‘-label)
-.47*** (46)
ARS-index 1 from WVS
(5-point-scale ‚agree/disagree‘-label)
Job Satisfaction
(nation-level-mean from
World-Values-Survey)1
1 pearson-correlation coefficient; significance: *** p < .01, ** p < .05, * p < .10; in parentheses: number of countries
9. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 9
Nation-level-results - 2
-.13 (14)
-.68** (11)
-.59*** (16)
-.71*** (16)
-.68*** (16)
Job Satisfaction
(nation-level-mean from
Eurobarometer)1
-.29 (25)ARS-index from Hofstede (2001)
-.47** (23)ARS-index from Harzing (2006)
-.38*** (47)
ARS-index 3 from WVS
(4-point-scale ‚agree/disagree‘-label)
-.47*** (42)
ARS-index 2 from WVS
(5-point-scale ‚important/unimportant‘-label)
-.47*** (46)
ARS-index 1 from WVS
(5-point-scale ‚agree/disagree‘-label)
Job Satisfaction
(nation-level-mean from
World-Values-Survey)1
1 pearson-correlation coefficient; significance: *** p < .01, ** p < .05, * p < .10; in parentheses: number of countries
§ the correlation is higher in european nations (countries of the EU)!
§ Note: european nations have – on average – better working conditions, since economic
wealth is higher (compared wordwidely)
§ we know: job satisfaction measurements are certainly influenced by working conditions as
well – not only communicational norms
10. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 10
Re-test of WVS-data
§ Operationalisation working conditions : GDP (gross domestic product)
§ Median split into good (rich countries) and bad (poor countries) working conditions
Rich countries Poor countries
Job satisfaction
(nation-level-mean from
World-Values-Survey)
Job satisfaction
(nation-level-means from
Word-Values-Survey)
ARS-index 1 from WVS
(5-point-scale ‚agree/disagree‘-label) -.51* (22) -.02 (24)
ARS-index 2 from WVS
(5-point-scale ‚important/unimportant‘-label) -.51*** (22) -.34 (20)
ARS-index 3 from WVS
(4-point-scale ‚agree/disagree‘-label) -.37* (23) -.04 (24)
ARS-index from Harzing (2006) -.61** (13) .08 (10)
ARS-index from Hofstede (2001) -.34 (19) sample to small
§ result: correlation remains important and significant only in rich countries with good working
conditions, almost no correlation within poor countries with bad working conditions
§ possible interpretation: following communicational norms becomes important only if a
sufficient working standard is established
1 pearson-correlation coefficient; significance: *** p < .01, ** p < .05, * p < .10; in parentheses (number of countries)
11. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 11
Re-test of the WVS-data
Very robust effect:
correlation remains the
same by including
economic indicators:
- HDI
- GDP
- Quality-of-Life-Index
Acquiescent response style
World-Values-Survey–jobsatisfactionlevel
Group of countries
rich countries
poor countries
rich countries
poor countries
12. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 12
-
acquiescent response style
individualism
Negative relationship between acquiescent response
style and job satisfaction level (in rich countries)
?
? job satisfaction level
Relationship of culture and job satisfaction:
JS-Level is higher in individualistic countries
(see Bosau, 2008, as well as Hofstede, Judge, etc.)
-
+
New finding
The 3 important constructs
13. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 13
+
-
acquiescent response style
individualism
?
job satisfaction level
-
Hypothesis:
Relationship of individualism and JS is only a spurious correlation
Instead: culture à response style à JS
The 3 important constructs
14. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 14
§ statistical analysis: mediation analysis (Baron & Kenney, 1986)
Test of mediation model
analysis by rich countries of WVS (N=21)
§ interpretation: relationship of culture and JS can be (almost completely) mediated by ARS;
thus: culture à response tendency à JS-level
total effect Individualism (Hofstede) job satisfaction
β = .39* (p=.08)
mediation model
job satisfaction
β = .20 (p=.39)
β = -.49** (p=.03) β = -.39 (p=.11)
Individualism (Hofstede)
ARS
§ same analysis with poor countries showed no total and no mediation effect
§ using several indicators for Individualism, ARS and job satisfaction in 8 different mediation
models we always get the same pattern of results
15. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 15
§ statistical analysis: stepwise multilevel analysis with HLM 6 (Bryk & Raudenbush)
Multi-Level analysis
analysis by rich countries of WVS (N=21)
ARS individualindividual level
national level ARS norm
JS
.072-1.43*Step 2: ARS norm 5.31** (1)
Step 3: ARS individual .07 .078 4.31** (1)
Step 4: ARS individual – random slope 70.60** (1)
§ result: negative relationship with national ARS norm remains significant; on individual level
no clear relationship of ARS and JS
§ same analysis with poor countries showed no effects at all
Step 1: Basic Model, controlled for age & gender
unstand. beta SE
increase in model fit
Chi-Square (df)
Step 5: ARS norm x ARS individual -.04 .889 0.002 (1)
16. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 16
JS and ARS on the individual level
7.10
7.56
8.03
8.50
8.96
Arbeitszufriedenheit
-1.25 -0.75 -0.25 0.25 0.75
Zustimmungstendenz
ARS individual
JSindividual
17. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 17
Conclusion
Results:
§ indeed: JS-measurements are influenced by response styles (ARS)
§ consequence: measurements from different cultures cannot be compared directly
§ national level: negative relationship, i.e. the higher the norm to agree, the more
moderate the JS-measurement
§ individual level: no clear relationship, i.e. in some countries negative and in some
countries positive
§ individualism-JS-relationship can be understood as a spurious correlation
§ instead: cultural values and their socialization leeds to a specific communication
style that in turn influences the JS-measurements
In conclusion: Cross-culturally, we are not measuring „true“ satisfaction. To a
great extent we are getting results that are an expression of
culturally socialized communication norms!
18. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 18
Life satisfaction
All countries together: N-ARS
5-point-scale
‚agree/disagree‘-label
WVS 1
Subjective Well-Being (Diener et al., 1995) -.52*** (40)
Life Satisfaction (Diener et al., 2000) -.36** (30)
Life Satisfaction (Suh & Oishi, 2002) -.50*** (29)
Happiness (Inglehart & Klingemann, 2000) -.42*** (51)
Life Satisfaction (Inglehart & Klingemann, 2000) -.54*** (51)
Ideal Life Satisfaction (Suh & Oishi, 2002) -.29 (29)
N-ARS
5-point-scale
‚agree/disagree‘-label
WVS 1
- Poor countries -
N-ARS
5-point-scale
‚agree/disagree‘-label
WVS 1
- Rich countries -
Subjective Well-Being (Diener et al., 1995) -.08 (11) -.58*** (21)
Life Satisfaction (Diener et al., 2000) -.02 (11) -.20 (11)
Life Satisfaction (Suh & Oishi, 2002) -.05 (11) -.53* (12)
Happiness (Inglehart & Klingemann, 2000) -.02 (22) -.57*** (21)
Life Satisfaction (Inglehart & Klingemann, 2000) -.19 (22) -.61*** (21)
Ideal Life Satisfaction (Suh & Oishi, 2002) -.03 (11) .02 (12)
Split into poor and rich
countries:
1 pearson-correlation coefficient; signifikance: *** p < .01, ** p < .05, *p < .10;
in parentheses: number of countries
1 pearson-correlation coefficient; signifikance: *** p < .01, ** p < .05, *p < .10;
in parentheses: number of countries
19. Bosau – satisfaction measurements in a cross-cultural environment 19. – 22.09.2011 – Statistische Woche – Leipzig Seite 19
contact details:
Dr. Christian Bosau
Dipl.-Psych. & Master of HRM & IR
Center for Evaluation
Faculty of Management, Economics and Social Sciences
University of Cologne
Herbert-Lewin-Str.2
50931 Cologne/Germany
Tel. +49 (0)221 470-4120
christian.bosau@uni-koeln.de
Thanks for your attention