Respondents' Race, Sex, and Language (probably) Predict Item Nonresponse on Sexual Identity and Behavior Questions Evidence from Three Cycles of the California Health Interview Survey
Similar to Respondents' Race, Sex, and Language (probably) Predict Item Nonresponse on Sexual Identity and Behavior Questions Evidence from Three Cycles of the California Health Interview Survey
Similar to Respondents' Race, Sex, and Language (probably) Predict Item Nonresponse on Sexual Identity and Behavior Questions Evidence from Three Cycles of the California Health Interview Survey (20)
Pteris : features, anatomy, morphology and lifecycle
Respondents' Race, Sex, and Language (probably) Predict Item Nonresponse on Sexual Identity and Behavior Questions Evidence from Three Cycles of the California Health Interview Survey
1. Respondents' Race, Sex, and
Language (probably) Predict Item
Nonresponse on Sexual Identity
and Behavior Questions
Evidence from Three Cycles of the
California Health Interview Survey
Matt Jans
David Grant
Annie Lee
2. Sexuality and Sensitive Questions
Private topic
Violates social norms to discuss with stranger
Disclosure risk – Sexual Orientation
For closeted respondents in certain contexts
Only sensitive for certain responses
Cultural and language complexities
Conversation/privacy norms
Equivalent (or lack of) concepts
3. What CHIS Brings to the Discussion
Sexual orientation questions for past 11 years
Record “Don’t Know” and “Refused” responses
So does BRFSS
Interviews in 5 Languages (2 Chinese Dialects)
English, Spanish, Cantonese, Mandarin, Korean, and
Vietnamese
Captures detailed ethnicity data
8 Asian ethnicities, 7 Latino ethnicities, Multi-group
identification for both
Other detailed language and country of origin info
4. CHIS/Data Overview
Multi-frame Cell RDD, Landline RDD, Surname
All interviews by phone
Data collected in 2-year survey cycles, now
continuous
Data on adults, teens, kids (adults analyzed here)
Data reported here includes 2-year cycles from
2005, 2007, and 2009
Exploratory
5. Sexual Orientation and Behavior
Questions
Located with other items on general health,
disability, and sexual health
Sexual orientation Immediately follows past 12 mos.
sexual behavior (including partner sex)
Not asked of adults over 70 years, proxy
respondents, kids, and teens
6. Non-Heterosexual Orientation
Prevalence and Nonresponse
3.78% in total in 2009
1.80% Gay/Lesbian (447,535)
1.36% Bisexual (339,608)
0.62% Nonsexual/Celibate/Other (153,844)
About 2% Item nonresponse
If all (or even most) NR comes from these groups,
proportion of those groups in population is
greatly under-estimated
7. “Is that partner male or female/In the past 12 months,
have your partners been male, female, or both male
and female?”
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
6.00%
% Not Answering
(DK, Ref & Not Asc.)
8. “Do you think of yourself as straight or heterosexual,
as gay {, lesbian} or homosexual, or bisexual?”
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
6.00%
7.00%
% Not Answering
(DK, Ref & Not Asc.)
Weighted
9. Interview Language and Sex Matter
0.00%
2.00%
4.00%
6.00%
8.00%
10.00%
12.00%
n=114,599
Male
Item NR %
Female
Item NR %
% Not Answering
(DK, Ref & Not Asc.)
10. Don’t Know & Refusal Reveal More
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
6.00%
7.00%
English Spanish Cantonese
Male DK
Male Ref
Female DK
Female Ref
11. Asian Ethnicities Not Identical
0.00%
2.00%
4.00%
6.00%
8.00%
10.00%
12.00%
14.00%
16.00%
Male
Item NR % (weighted)
Female
Item NR % (weighted)
% Not Answering
(DK, Ref & Not Asc.)
Satisficers?
(small n)
12. Mexican, Central American and
Other Latino Subgroups
0.00%
0.50%
1.00%
1.50%
2.00%
2.50%
3.00%
3.50%
4.00%
4.50%
Male
Item NR %
Female
Item NR %
% Not Answering
(DK, Ref & Not Asc.)
13. Language, Ethnicity, and Culture
Native language v. English
Comprehension and terminology
Cultural concepts and norms around sex and
sexual orientation
Need explore linguistic aspect of culture more
Exploit our other “enculturation” measures
14. Next Steps in Analysis
Is NR common?
Sensitive and nonsensitive items
Is missingness random?
Little’s test for MAR
What variables predict it?
More detailed “enculturation” measures
DK v. Refused
Interviewer effects