Methods and Assessment of Mental Health Problems

Scientific method #

1. Determine an interesting and answerable question
2. Formulate hypothesis regarding the question
3. Develop method for testing the hypothesis
4. Use resulting data to draw a conclusion

For the statistical machinery behind step 4 (null vs. alternative hypotheses, p-values, false-positive control), see hypothesis testing.

Example: humans typically have a fear of public speaking.

1. Question: Is concern for reputation something that is unique to humans?
2. Hypothesis: Young children have concern for reputation, but chimpanzees do not.
3. Method: conduct stealing task and helping task with two conditions: observed (someone watching) and unobserved
4. Conclusion: observed vs unobserved didn’t matter for chimps: they stole/helped at high levels independent of the condition. However, human children stole less and helped more when an observer was present.

Reliability #

Reliability is the degree to which independent measurements of a behavior under study are consistent.

• Interrater reliability: between two raters who observe the same behavior, how consistent are their observations?
• Test-retest reliability: between two independent tests under the same conditions, how similar are the observed behaviors?

Validity #

Validity is the degree to which an experiment measures what it is intended to measure.

• Internal validity: whether effects observed within experiments can be attributed with confidence to the factor the researcher is testing
• External validity: ability to generalize findings beyond the particular setup of the specific experimental design

Internal vs external validity is often a tradeoff: controlled laboratory tests have a lot of internal validity but may not be generalizable; naturalistic studies have a lot of external validity but are harder to control.

Purposes of Assessment #

Diagnosis: identifying the nature of the current problem Prognosis: generating predictions for future behavior under specified conditions Case conceptualization and treatment planning: putting all of the information together as a formulation of a case as a whole

• Similar to research study in that a hypothesis (diagnosis) is formed and evaluated based on evidence Treatment Monitoring and outcome evaluation: ongoing assessment over the course of intervention to see if it’s working

Domains and Constructs of Assessment #

The following domains all provide additional context to clinicians to best design a treatment plan.

The big idea: collect data from multiple informants and multiple contexts using multiple methods.

• This is because the clinical significance of behaviors varies by context. For instance, a child acting out in the classroom would likely be seen as more problematic than a child acting out during a party. This concept is known as developmental expectability.
• Children’s self reports of symptoms may be of limited value due to differences in perception and difficulties in verbal communication.
• Cultural differences may play a factor

Symptoms and Presenting Problems: what behavior can be observed?

Functional Impairments and Adaptive Behaviors: what behavior can we test?

• Gross and fine motor skills
• Independent daily living
• Social skills
• Cognitive and academic skills

Risks and Factors:

• Genetic/family history
• Physical health
• Environment (culture, community, school, family)
• Personality and temperament

Psychological Processes:

• Cognition (thoughts and thought processes)
• Language
• Emotion

Methods of Child Assessment #

Comprehensive clinical interview

• Typically done with parents/caregivers
• Varying levels of structure (set question list vs casual meeting)
• Covers multiple domains: parental concerns/perceptions, developmental history, resources for treatment, parental risk/asset factors
• Pros: large amounts of data with in-depth information
• Cons: can’t be used for interviewing infants and young children (dependent on verbal ability); confounding factor of social desirability (will give answer that makes them look good to the experimenter, rather than what they actually think)

Behavioral rating scales

• Standardized sets for assigning scores in a checklist: can compare cases to large dataset
• Can ask the same set of questions to multiple individuals (parent, teacher, child) multiple times throughout the treatment and compare results

Behavioral Observation

• Direct observation of behaviors in specific settings not controlled by the researcher
• Standardized by observation scales
• Logistical and cost considerations
• Pros: naturalistically relevant (high external validity), applicable to everyday social interactions
• Cons: relevant behaviors are very rare, mostly only descriptive data

Psychological Testing (experiments)

• one on one testing with child using standardized materials and procedures
• Many domains of tests (intelligence, function, attention, language, motor/sensory perception, learning/memory)
• Pros: reveals psychological mechanisms underlying behaviors
• Cons: tests are conducted in unfamiliar setting for child (lacks external validity)

Technologies #

Eye tracking: useful for assessing infants and young children

• Preference paradigm: when given a choice, what do children select? (can use eye tracking)
• Violation of expectation paradigm: can children detect when something unusual has occurred?

Depth-sensor imaging: using body posture to assess thoughts/emotions

• Changes in posture: more elevated = feeling confident/good, lowered = ashamed/bad

Approaches #

Longitudinal design vs Cross-sectional design: examine development of the same child over many years, vs. examine many children at different stages of development at the same time

Cross-cultural approach: see cross cultural psychology Comparative approach: relating human behavior to evolutionary history and the behavior of related species to determine what behaviors are unique to humans Computational approach: combine computational models with behavioral experiments to represent how humans behave mathematically Open Science movement: scientists are encouraged to share their data publicly for use by other researchers, and preregister their hypothesis+methods so they can’t change it after data arrives

Criterion for Good Measures #

Evidence-based assessment (EBA): assessment methods and processes that are based on empirical evidence for reliability, validity, usefulness for prescribed population/purpose

Test Reliability: consistency of test scores within participants

DSM #

Diagnostic and Statistical Manual for Mental Disorders

• Compilation of diagnostic criteria by disorder
• Symptoms have threshold criteria of length/severity to qualify
• Categorical, not dimensional: either someone meets the criteria or they don’t

Benefits:

• Organized/structured
• Allows researchers to locate relevant studies
• Reliable, valid diagnostic measures

Limitations:

• artificial cutoff point
• misses comorbidity
• lack of attention to causes
• culturally rooted (may miss normalized behaviors in other cultures that are abnormal in western/American culture)
• some disagreements

#