| Chapter Outline (See related pages)
- Characteristics of the Scientific Method
- Skeptical Attitude
- Skepticism is necessary because behavior is complex
- Many factors are needed to explain behavior
- It is difficult to identify these factors
- Skepticism is necessary because science is a human endeavor
- Scientists are not passive observers
- Many decisions made in science are subjective
- Objectives
- Four (4) objectives are met through the scientific method
- Description is defining and classifying events and their relationships
- Descriptions must have reliability
- Descriptions must have validity
- Prediction is based on description
- Control involves development of treatment and preventive strategies
- Understanding of behavior consists of identify cause(s)
- Covariation of events refers to two events varying
together
- Time-order relationship where cause must occur
before effect
- Elimination of plausible alternative causes rules
out other likely causes
- Internal and External Validity
- Confounding often arises when eliminating plausible
alternative explanations
- Studies free of confounding have internal validity
- External validity refers to degree to which results
can be generalized
- Generalizability is ability of a finding to be applied
to different populations, settings, and conditions
- Representativeness determines generalizability and
refers to degree to which sample matches population
- Random sampling achieves representativeness; a random sample
is where everyone in population has equal chance of being included
- Many studies use samples of convenience
- To replicate a study tests its generalizability
- Scientific Procedures
- The Hypothesis
- A hypothesis is a tentative explanation for behavior
- Intuition and hunches often play roles in developing a hypothesis
- Clinicians develop hypotheses by observing their patients'
reactions to treatment
- Operational Definitions
- Hypotheses must be falsifiable, that is stated in a
way that it can be proven untrue
- Terms in hypothesis must have operational definitions
with regard to operations that can be measured or observed
- Methods of Control
- Independent variable is factor that is manipulated
to measure its effects
- Dependent variable is factor that may be affected and
whose changes will be measured
- Internal validity when control techniques are properly
used
- Independent variable is manipulated at two levels: experimental
group and control group
- Other variables are held constant to prevent confounding
- Balancing is used to control factors that cannot be
held constant; random assignment is the most important
balancing technique
- Minimizing the effects of expectations
- Demand characteristics are expectations that the
subject has
- Experimenter effects are biases from experimenter
expectations
- Placebo and placebo subject groups are procedures
to control for demand characteristics and experimenter effect
- Double-blind procedure is when subject and experimenter
are unaware of which treatment is given
- Statistical inference is used to conclude that results
are real and not due to chance
- Assumption of null hypothesis
- Use of probability theory to determine odds of obtaining
results if null hypothesis were correct
- Statistically significant leads to rejection of null
hypothesis
- Problems with statistical inference
- Statistical significance does not necessarily mean finding
is important
- Statistical significance does not necessarily mean clinical
significance
- Research Designs
- Correlational Research
- Correlational research examines relationships
- In case-control designs, cases (those with mental disorder)
are compared with controls (those who do not)
- Correlation does not show causal relationships
- Relationships between variables may be due to third-variable
problem
- Matching may solve third-variable problem
- Longitudinal studies involve examining same subjects over
period of time
- High-risk designs involve examining subjects who have
high likelihood of developing a disorder
- Genetic high-risk designs include subjects who are genetically
predisposed to disorder
- Behavioral high-risk designs include subjects who have some
behavioral characteristic that makes them at risk
- Caution is needed when inferring causation
- Epidemiological Studies
- Epidemiology studies frequency and distribution of disorders
within specific populations
- Incidence is number of new cases within given time
period
- Prevalence is percentage of population that has disorder
- Duration is average length of disorder
- Epidemiological studies are prone to pitfalls such as unrepresentative
samples
- Experimental Designs
- Clinical Trials
- Clinical trials test effectiveness of treatments
- Empirically supported treatments are those that pass the test
of controlled experimental research but may not be clinically
significant
- Analogue Experiments
- Analogue experiments are experiments that serve as
models of real life
- A high degree of internal validity is possible due to types
of control necessary to identify causal relationships
- Allows testing of variables that otherwise could not be manipulated
- The more ethical an analogue experiment is, the less analogous
it is likely to be
- Animal models can be used to address problems of analogue
experiments
- Animal models allow variables to be held constant
- Some procedures are too intrusive to be used with humans
- Long-term consequences of treatment and disorder can
be assessed
- Threats to external validity exist in analogue experiments
- The Single-Case Experiment
- Multiple groups of subject pose problems
- Single-case experiment focuses on one person
- First stage is baseline: behavior before treatment is assessed
- Treatment is introduced; comparison is made
- Common design is ABAB: baseline, treatment, baseline, treatment
- Multiple-baseline design does not involve an interruption of treatment
- Limitation of single-case design is weak external validity
|
|