Measures of
Acculturation
Nazleen Bharmal, MD, PhD
RCMAR/CHIME and Project
EXPORT Methods Seminar
August 2013
Acculturation
• Process by which individuals exposed
regularly to another culture adopt the attitudes,
values, customs, beliefs, and behaviors of the
foreign culture
• Unidirectional vs. Bidirectional
• Health Outcomes: depression, obesity, DM,
CVD, cancer
• Health impact depends on how acculturation is
measured
Measures of Acculturation
• Acculturation Scales:
– Cultural practices, values, identification
– Hispanics: ARSMA, SASH, BAS
– Asians: SL-ASIA
• Proxy Measures:
– Language proficiency
– Country of nativity
– Temporal: length of residence, age at immigration
California Asian Indian
Tobacco Survey
• Telephone survey of 3228 Asian Indian
adults
• RR=81% (adult), 54% (house x adult)
• Conducted in several South Asian
languages
Acculturation Domains
Domain
Question & Responses
Language use
1. English as primary language? Y/N
2. How often native language spoken at home? Likert scale
3. Language of the interview? English, Gujarati, Hindi, Punjabi
Media
behavior
1. How often do you read Indian newspapers, magazines, books,
or content on the internet? Likert scale
Social customs
1. How often do you eat Indian food? Likert scale
2. How open would you be to your son/daughter marrying outside
of cultural group? Likert scale
3. Do you observe traditional holidays impt in your culture? Likert
Social contacts
1. How often keep in contact with family and friends in India? Likert
Ethnic identity
1. Self-reported ethnic identity measured on a 5-response scale:
Full Indian, Indian first-American second, Equal blend, American
first-Indian second, Full American
Generational
status
1. Generational status? 1st or 2nd+ generation
2. Born in a developed country? Y/N
Temporal Measures
• Duration of residence in the U.S.
• Percentage of time in the U.S.
• Age at immigration
Relationship Between Ethnic Identity
and Number of Years in the U.S.
Acculturation Items
% sample Mean yrs 95% CI
Full Indian
24
7.6
(7.1, 8.1)
Indian 1st, American 2nd
16
12.3
(11.3, 13.2)
Equal Indian, American
50
14.4
(13.8, 15.0)
American 1st, Indian 2nd
6
17.1
(15.0, 19.2)
Full American
6
18.7
(16.3, 21.2)
F-test p-value<0.001; r = 0.32
Associations of acculturation items
and number of years in the US
Evaluation of Dimensionality of
Acculturation Items
Evaluation of Dimensionality of
Acculturation Items
• Exploratory Factor Analysis: criteria for
determining number of factors
– PCA eigenvalues > 1
– Scree test plot: graphical display of
components and corresponding eigenvalues
– Parallel analysis: based on a Monte Carol
simulation, compares observed eigenvalues
with uncorrelated normal variables
EFA Examples
0
.5
1
1.5
2
Scree plot of eigenvalues after factor
10
1
8
.5
6
Number
0
4
-.5
2
Eigenvalues
0
1.5
-.5
2
Parallel Analysis
0
2
4
6
8
Factor
Factor Analysis
Parallel Analysis
10
Factor Rotation
• Rotate plausible number of factors
– Look for simple structure
– Orthogonal v. Oblique (permits factor
correlation)
• STATA Command:
estat common
estat structure
rotate, promax factors (#)
griegan, yline (1)
estat smc
Promax obliquely rotated
2-factor solution
Factor 1 Factor 2
Frequency of contact with family & friends in India
0.7880
-0.2501
Frequency read Indian media
0.6329
-0.0152
Generational status/Born in developed country
0.5939
-0.0114
Frequency native language spoken at home
0.5278
0.3916
Frequency eat Indian food
0.5016
0.2014
Self-assessed cultural identity
0.4821
0.0725
Language of interview
-0.1503
0.7570
How open to child marrying outside cultural group
-0.0703
0.7177
English as primary language
Observe traditional holidays in your culture
0.2464
0.0958
0.5688
0.5216
Factor 1 – Engagement in Indian cultural behaviors; Factor 2 – Social behaviors in US
Bold – factor loading > 0.40
Confirmatory Factor Analysis
• Specify specific model and evaluate fit
• Model fit evaluation:
– Chi-squared tests
– Goodness of fit index
– RMSEA (root mean square error of
approximation)
– CFI (comparative fit index – relative to null)
Parameter estimates for CFA model
ITEM
Null
Factor 1 Factor 2
Factor 1: Indian cultural behaviors
Frequency native language spoken at home
Frequency eat Indian food
Frequency read Indian media
Generational status/Nativity
Self-assessed cultural identity
Frequency contact w/ family & friends in India
Factor 2: Social behaviors in U.S.
English as primary language
How open to child marrying outside group
Observe the traditional holidays in culture
Language of interview
Factor 1
Factor 2
0.71
0.45
0.43
0.39
0.35
0.38
0.53
0.36
0.37
0.34
1.00
0.60
0.57
0.52
0.50
0.49
-
1.00
0.72
0.72
0.55
Factor 1 Factor 2
1.00
0.08
1.00
Confirmatory Factor Analysis
• 1-Factor Model:
– RMSEA (0-1) = 0.10
– CFI (0-1) = 0.65
• 2-Factor Model:
– RMSEA = 0.08
– CFI = 0.79
RMSEA:
Smaller values
indicate better
model fit ≤ 0.06
CFI: Larger values
indicate better
model fit ≥ 0.90
• 2-Factor Model with correlated residuals:
– RMSEA = 0.07
– CFI = 0.89
Scoring of
Acculturation scale
• Transformed each of items to have a 0100 possible range
– e.g., frequency of native language had 5
response categories (1-5) that became: 0, 25,
50, 75, 100
• Higher score means greater level of
acculturation to U.S.
Acculturation Scale
Characteristics
• 11-items averaged together
• Mean = 39 (range: 0-100; SD 17)
• Skewness = 0.45, Kurtosis = 3.05
• Internal consistency reliability = 0.73
• Product-moment correlations with:
– r = 0.37 with years lived in the US
– r =0.45 with percentage of lifetime in the US
– r = -0.34 with age at immigration
Mean acculturation score by yrs
in U.S. among immigrants
Mean Acculturation Score
100
90
80
70
60
50
40
30
45.3
32.5
34.6
36.4
0-5
6-10
11-15
20
10
0
Years lived in the US (years)
15+
Acknowledgements
• Ron Hays – for methods mentoring
• William McCarthy – for content mentoring
• NRSA, AHA, STAR fellowships
• Validity of Temporal Measures as Proxies
for Measuring Acculturation in Asian Indian
Survey Respondents. J Immigr Minor
Health. 2013
Descargar

FACTOR ANALYSIS: Acculturation Scale