When deciding about research participation, a person

1
Does Race Alter Your Attitude Toward Research?
Jonathan Rubright BA1,5,6,8, Kristin Harkins
BA1,5,6, Paige Brookstein BA1,5,6, Jason
Karlawish MD1,2,3,4,5,6,7 University of
Pennsylvania, Departments of Medicine1, Medical
Ethics2 Center for Clinical Epidemiology and
Biostatistics3, Center for Bioethics4, Institute
on Aging5, Alzheimers Disease Center6, Leonard
David Institute of Health Economics7, Graduate
School of Education8 The research reported on
this poster was supported by R01-MH071643. The
investigators retained full independence in the
conduct of this research.
BACKGROUND

METHOD

• When deciding about research participation, a
  person
• balances altruism against potential harms.
  Attitudes about
• research may alter this balance.
• Research attitudes may be lower among African
  Americans.
• Studies have not directly measured factors
  that may
• modify association between race and research
  attitude.

• 542 persons agegt65 interviewed in person (see
  table 2).
• Research Attitudes Questionnaire and covariates
  collected.
• A simple and multivariate regression model run
  to examine
• impact of race on scores.

OBJECTIVE
To compare attitudes toward research between
black and non-black elderly, and identify
covariates which may mediate any relationship.
plt0.05.
RESULTS
INSTRUMENT

• Simple linear model shows black race decreases
  estimated
• scale scores by 2.3 points.
• Adding in covariates drops this beta to -1.8,
  which is
• statistically smaller (seemingly unrelated
  estimation chi2
• plt0.05).
• Magnitude of change in Black ß1 is 22.
• R-square changes from 0.07 to 0.12 in
  multivariate model.

• Kim 11 item Research Attitudes Questionnaire
  (see table 1).
• Likert scale options strongly disagree to
  strongly agree.
• Scores range from 11 to 55, higher scores
  indicate
• higher value placed on research.

SIMPLE LINEAR MODEL
?(research score) ß0 ß1(Black)
SUMMARY

• Controlling for covariates reduces the
  statistical impact
• of race on research attitude scores.
• Though statistically significant, differences in
  research
• attitudes between blacks and non-blacks may
  not be
• large enough to be meaningful.
• Future studies should examine gender, education,
  and
• socioeconomic status for their influence on
  decisions
• about whether to participate in research.

plt0.05.
MULTIVARIATE MODEL
?(research score) ß0 ß1(Black) ß2(Age)
ß3(Female) ß4(Education)
ß5(Finance) ß6(Medical job)