Download slides - Qc.edu

IMPROVING SURVEYS of
SEXUAL BEHAVIOR
USA National STD and Behavior
Measurement Experiment
(NSBME)
Charles Turner
Alia Al-Tayyib
Susan Rogers
Maria Villarroel
Anthony Roman
James Chromy
Phillip Cooley
Funded by NIMH and NICHD
(R01-MH56319 & R01-HD31067)
OBJECTIVE
• Importance of Self-Report
• Final Installment of NSBME
• “PIs know best”
SELF-REPORT
• Almost everything we know about sexual
behaviors
• Population Surveys
• Evaluation of Interventions
• Clinical Medicine
• Privacy, SAQs, and , Audio-CASI
• Developed by Phil Cooley in 1994-95.
• Human interviewers recruit respondents and
introduce survey
• Computer-recorded questions played and
respondents used telephone keypad to respond
• First demonstration of impact of T-ACASI:
1996 ASC meetings
T-ACASI Advantages
• Entirely private administration
• Completely standardized – everyone
hears exactly same question wording,
intonation
• Eases multilingual interviewing
NSBME: National STD and
Behavior Measurement Experiment
Two probability samples of English speaking
adults ages 18-45 living in households with
telephones
– National stratum, N=1,543
– Baltimore stratum, N=744
Telephone numbers randomly assigned to T-IAQ
or T-ACASI condition
NSBME
FINDINGS TO DATE
Increased illicit but not licit drug use
Addictions, 2004
Increased same-gender sex --particularly in gay-unfriendly regions
Public Opinion Quarterly, 2006
Increased reporting of STD history
STDs, 2008
Increased reporting of unpopular social
attitudes
Public Opinion Quarterly, 2009
Heterosexual Sex
Forthcoming
International Journal of
Epidemiology
Summer / Fall, 2009
No Sexual Experience
OR= 2.1, P = 0.001
6
5
4
T-ACASI
3
T-IAQ
2
1
0
No Sexual Experience
No Main Sex Partner in Past Year
OR= 1.9, P = 0.001
New Sex Partners, 12 months
p < 0.001
Extra-Relationship Sex during
Committed Relationship
OR= 1.6, p = .001
14
12
10
8
T-ACASI
6
T-IAQ
4
2
0
Extra-Relationship Sex
Never Oral Sex
OR= 0.5, p < 0.001
OR=0.5, p < 0.001
25
000
20
15
T-ACASI
10
T-IAQ
5
0
Never Gave
Never Received
Ever Heterosexual Anal Sex
OR= 1.7, p < 0.001
Never Sex while Menstruating
OR= 0.7, p < 0.001
50
40
30
T-ACASI
T-IAQ
20
10
0
NEVER
Used Condom Every Time
Past Year
OR = 0.5, p < 0.001
Difficulty having
Satisfying Sex
OR = 1.4, p = 0.011
OR = 1.3, p = 0.097
“PIs Know Best”
“Very Easy” to become Sexually Aroused
OR = 1.8
p < 0.001
OR = 2.8
p < 0.001
OR = 1.5
p = 0.003
Always Asked or Told about
Past Sex Partners
OR = 1.4, p = 0.058
OR = 1.5, p = 0.028
CONCLUSION
T-ACASI Reduces Self-Report Bias
What People Do
vs.
What People Say They Do
Lessons in Fallibility of PI Assumptions