UNDERSTANDING PSYCHOPATHOLOGY

1
UNDERSTANDING PSYCHOPATHOLOGY

• Research Methods in Psychopathology

2
OVERVIEW

• The key components of the scientific method
• Major designs exemplified
• Potential sources of bias

3
WHY CONDUCT A STUDY?

• Describe a phenomenon
• Generate hypotheses
• Test hypotheses

4
THE SCIENTIFIC PROCESS

• Hypotheses specify the expected relationship
  between independent and dependent variables
• All variables are operationalized
• A design is selected for testing the hypotheses
• The study is executed
• Conclusions are drawn about the hypotheses

5
STUDY DESIGNS

• Passive-observational studies (correlational
  studies)
• Experiments

6
SCIENTIFIC METHOD

• Generalizability
• Reliability
• Validity

7
PASSIVE-OBSERVATIONAL STUDIESEXAMPLES

• Cross-sectional study
• Longitudinal study

8
CORRELATIONS STUDIES e.g., POPULATION SURVEYS

• Generalizability is high if
• Representative sample
• High participation rate
• Randomness of non-participation
• Measurements are of comparable reliability and
  validity across sub-populations (e.g., gender
  ethnicity age). (Marsella, 1997 linguistic,
  conceptual, scale, and normative equivalence)

9
POPULATION SURVEYS EXAMPLES

• Epidemiological Catchment Area Study (ECA)
  (Robins Regier, 1991) -- early 1980s
• 20,000 respondents in five communities
• included institutionalized people (e.g.,
  hospitals, prisons)
• National Comorbidity Survey (NCS) (Kessler et
  al., 1994) -- early 1990s
• national sample of 8098 non-institutionalized
  males and females (ages 15-54)

10
12-MONTH PREVALENCE OF MENTAL DISORDRES IN THE
U.S.
11
NCS 12-MONTH PREVALENCE OF MENTAL DISORDERS IN
MEN AND WOMEN
12
NCS ETHNICITY RESULTS COMPARED TO WHITES...

• Asian American Native American -- ??
• Black Americans
• lower rates for mood, substance use disorders
• no disorder where rates were higher among blacks
  than whites
• Hispanics
• higher rates for current mood disorders
• no disorder where rates were lower among
  Hispanics than non-Hispanic whites

13
POTENTIAL SOURCES OF BIAS

• Representative population sample
• Response rate 82.6
• no differences between responders and
  non-responders on gender or age
• non-responders were found to have significantly
  higher rates of disorder (based on a parallel
  study offering financial incentives to a
  subsample of nonresponders).
• did participation vary by ethnicity?

14
REPRESENTATION OF ETHNOC MINORITY GROUPS IN THE
NCS?

• Black Americans (12 of U.S. population) are over
  represented among
• people who are homeless (40 of homeless)
• people who are incarcerated (almost 50 of state
  and Federal prison inmates)

15
REPRESENTATION OF ETHNOC MINORITY GROUPS IN THE
NCS?

• Asian Americans/Pacific Islanders (4 of U.S.
  population)
• Extremely diverse group 43 different ethnic
  subgroups
• Over 100 languages and dialects
• 35 live in households where there is limited
  English proficiency

16
REPRESENTATION OF ETHNOC MINORITY GROUPS IN THE
NCS?

• Latinos/Hispanic Americans (11 of U.S.
  population)
• Quite diverse group
• 40 have limited English proficiency
• Over represented among prison inmates

17
REPRESENTATION OF ETHNOC MINORITY GROUPS IN THE
NCS?

• Native Americans/Alaska Natives (1-1.5 of U.S.
  population)
• Diverse group over 560 recognized tribes, over
  200 indigenous languages
• Over represented among
• the homeless (8 of NA/AN are homeless compared
  to 2 of U.S. population)
• prison inmates (4)

18
CORRELATIONAL STUDIES

• Suitable for answering descriptive questions
• Also are used for testing hypotheses, but
• Cannot test causal hypotheses

19
EXPERIMENTS

• Key feature control over the independent and
  third (confounding) variables