Clinical Psychology Research Methods

• Research is fundamental to clinical psychology, influencing diagnosis, treatment, and our understanding of human behavior.
• Psychology is unique among mental health fields because of its strong research foundation.
• A PhD in psychology signifies an academic and research focus, unlike practice-based degrees like MD, JD, and MSW.
• Even with the PsyD model that emphasizes practice, clinical psychology still requires research for both scientific and practical reasons.

Experiments

• Experiments are a primary method for applying the scientific method.
• Determining if relaxation or aerobic exercise can effectively reduce test anxiety can be done through research.
• A hypothesis can be formed that both relaxation and aerobic exercise reduce anxiety compared to a control group.
• One may hypothesize that relaxation will be more effective than aerobic exercise due to its proven anxiety-reducing benefits.
• Hypotheses are made using existing knowledge regarding interventions.
• A hypothesis is required as the first step to designing an experiment

Variables

• Independent variables are manipulated by the researcher and create the study structure.
• Treatment condition (relaxation, aerobic exercise, control) is an independent variable.
• Dependent variables are expected to change due to the influence of independent variables.
• Scores on a test-anxiety scale are an example of a dependant variable.
• Research studies assess the effect of independent variables on dependent variables.
• A study must control all factors influencing the dependent variable, except the independent variable.

Minimizing Experimental Error

• Experimental error occurs when factors beyond the independent variable affect the dependent variable.
• Experimenter expectancy effects are biases that can unintentionally influence results.
• These biases can manifest as differing behaviors toward subjects in different conditions.
• Using a research assistant unaware of the hypotheses reduces bias.
• Standardizing procedures (scripted instructions and uniform attire) ensures consistency.
• Random assignment of subjects to conditions evenly distributes the influence of uncontrollable or unknown variables.
• Randomization aids negation of confounding variables.

Reliability and Validity in Research

• Researchers must use reliable and valid assessment measures.
• Reliability is the consistency of a measurement procedure.
• Assessments should yield similar results regardless of when or by whom they are administered.
• Validity is the extent to which an instrument measures its intended construct.
• A test measuring anxiety should accurately measure anxiety rather than depression or self-esteem.
• Measures in research must have adequate reliability and validity.
• Errors can occur in instrument choice, data collection, scoring, and analysis.
• Objective, precise, reliable, and valid scoring methods must be used to minimize error.
• Checking and rechecking coded data, data entry, and data analysis is essential to spot mistakes.

Maximizing Validity

• Internal validity means the independent variable accounts for results on the dependent variable in an experiment.

• Extraneous influences on the dependent variable threaten internal validity.

• Researchers ensure internal validity through well-constructed, sound methodologies.

• History, maturation, testing, instrumentation, statistical regression, selection bias, and experimental mortality threaten internal validity.

• History refers to events outside the experiment impacting results.

• Maturation includes changes within subjects (e.g., aging, fatigue) influencing experimental results.

• Instrumentation refers to the impact of tests used in the study.

• Statistical regression is the tendency of extreme scores to move toward the mean.

• Selection bias is a problematic selection procedure for research subjects.

• Experimental mortality is attrition or subject drop out in an experiment.

• External validity is the generalizability of research results, which is greater when the experiment mirrors real-world situations.

• Researchers must assess threats to external validity, including testing, reactivity, multiple-treatment interference, and selection biases.

• Testing may sensitize subjects and alter their responses.

• Reactivity is a subject's potential response to participating in an experiment that changes their behavior.

• Multiple-treatment interference prevents isolation of a specific condition or factor due to exposure to several treatments.

• Interaction of selection biases means some subjects may be uniquely responsive to the experimental condition.

• An important threat to external validity includes the sampled population.

• If only middle-class Caucasian college students are sampled, results can't be generalized.

• Researchers must ensure the sampled group adequately represents the overall population.

Experimental Designs

• Various means exist to experiment and get different advantages and disadvantages.
• Using the right experimental design with the right research question with careful construction to maximize validity is key.

True Experimental Designs

• Randomization procedures with experimental and control conditions must be performed to demonstrate cause-and-effect relationships.

• Efforts must be made to minimize potential error to maximize validity.

• Randomization ensures all research subjects have an equal chance of placement in experimental and control groups.

• Randomization helps ensure outcome measures are due to the manipulated independent variable.

• True experiments face unique challenges in clinical psychology.

• Random assignment of humans to conditions is often impossible or unethical (e.g., sexual abuse on depression).

• Assigning patients to a no-treatment control group is also often unethical (e.g., suicidal patients).

• Certain disorders are rare, making it hard to get enough subjects.

• Co-morbidity makes it difficult to find people with only the specific disorder under investigation.

Quasi-Experimental Designs

• A quasi-experimental design can be chosen if random assignment is not possible due to limitations.
• Treatment-outcome studies may use patients already being treated at a clinic.
• Control subjects may include persons not being treated at the clinic or those on a wait-list for treatment.
• Experimenters cannot always decide who receives treatment, making randomization impossible.
• Important differences exist between treated and untreated children that cannot be controlled.
• Practical limitations require caution when interpreting research findings, and direct cause-and-effect relationships cannot be inferred.

Between Group Designs

• Between group designs use two or more groups of subjects, each receiving a different intervention.

• The independent variable is manipulated so different groups receive different experiences.

• One group gets relaxation, one gets aerobic exercise, and a control group reads magazines as a test-taking anxiety example.

• Subjects are ideally randomly assigned to treatment and control conditions.

• Experimenters assume all other potential factors that could influence the study results will be evenly distributed between the groups.

• Matching procedures ensure potential factors do not differentially affect the experimental and control groups.

• Matching ensures gender and age are similar across groups as well as distribute the severity of anxiety evenly among the groups.

• Pretest-posttest control group designs include two or more groups, where one receives treatment and the other doesn't.

• A test-anxiety questionnaire would be used before and after treatment.

• Control completes the questionnaire at the same time as the experimental group.

• Pretest administration might sensitize subjects or influence their treatment response.

• A posttest-only control group can be added to control for pretest sensitivity.

• Factorial designs study two or more factors in a given study.

• Two independent variables of interest can be examined at the same time.

• 4 groups would be tested for; male African American therapist, female African American therapist, male Caucasian therapist, female Caucasian therapist can be used as an example.

• The advantage of factorial studies is the opportunity to examine interactions between factors.

Within Group Designs

• Within group designs examine the influence of the independent variable on the same subjects over time.

• Subjects are not assigned to different groups but experience the same research procedure.

• The same patient is examined at different points, such as during baseline, treatment, and follow-up periods.

• Each subject serves as their own control.

• Subjects at risk are asked to participate in a hypertension study to examine their blood pressure and heart rates while performing different intellectual and cognitive tasks.

• Blood pressures before,during and after a series of lab experiences were taken.

• Changes within subjects allowed experimenters to study stress and physiological reactivity among hypertensive high risk persons.

• Experimenters must be careful when choosing designs within groups as ordering or sequencing effects may cause issues.

• Ordering effects are the influence of the order in which treatment is presented to subjects.

• The experimental/treatment condition switches or "crosses over" during the experiment in the cross over design.

• Treatments are generally the same for each group but the order is altered.

• To control for order presentation the crossover design counterbalances the treatments.

• Presenting laboratory stressors can be counterbalanced. Half the subjects would experience one stressor before the second and the other half reversed.

• Crossover designs can become very complicated when three or more treatments are required and are called multiple-treatment counterbalanced designs.

• Six counterbalanced presentations are needed if three treatments/experimental conditions are used.

• Suppose three different treatment strategies existed; behavioral contracting, group therapy, and interpersonal psychotherapy.

• All subjects should receive all three treatments but in a different randomized order.

• Controlling the order of presentation is needed as treatment effects might be influenced by the order experienced (first, second, or third).

• Difficulty can arise in assessing subjects because a carryover can affect the previous treatment.

Mixed Group Designs

• Mixed group designs have elements of both within and between group designs.
• Experiments are constructed so different groups receive different experimental experiences (between group) while responses are assessed over time (within group).
• In the hypertension study control groups of subjects who were not at risk for hypertension were also used.
• The presence or absence of hypertension risk was a between group variable, while experimental tasks were within group variables.
• Mixed group designs more complex but commonly used in clinical psychology.

Analogue Designs

• Analogue designed studies use procedures, subjects, and measures that approximate a real-life clinical situation.

• They are typically conducted in a lab where experimental conditions can be better controlled.

• Subjects may or may not be suffering from the disorders.

• Procedures may or may not be identical to those used.

• Therapists may or may not be licensed mental health providers.

• A test anxiety experiment was conducted to see if brief treatment might be useful during a simulated IQ test.

• The subjects were merely students who identified themselves as test anxious and who scored high instead of actual students who sought help for treatment.

• The treatment conditions of relaxation training and aerobic exercise were conducted in the laboratory rather than in a more natural setting.

• Analogue studies may use trained research assistants to conduct “treatment conditions” and may involve subjects who do not meet specific diagnostic criteria for a particular disorder.

• Advantages include maximizing internal validity more effectively than studies conducted in a more natural environment.

• Disadvantages include threats to external validity, since generalizability of research findings may be compromised.

Case Studies

• A case study is an in-depth investigation, description, and observation of a single person, group, or situation.
• It was the primary technique used by Sigmund Freud to describe cases of hysteria and other problems.
• Case studies are not experiments because they lack dependent variables, experimenter manipulation of independent variables, and randomized assignment to treatment and control conditions.
• Case studies provide an intensive observation of a person and phenomenon that allows for theory development.
• Case studies are useful for examining a new, rare, or during the early descriptive stages of a research program.
• Theories developed through case study methods can be tested later with more intensive research designs.

Single Subject Designs

• Single case designs blend case study and experimental techniques.

• It offers rigor of experimental methods, while allowing clinical relevance because they are used with only one patient.

• Clinicians can study and treat individual patients in their practice.

• It avoids the problems associated with variability among a large number of subjects.

• These designs use time-series methodologies, which require a series of measures conducted on the same person over a period of time.

• Rather than using a separate control group or subject, the patient acts as his or her own control during the baseline phase with data on the phenomenon of interest without any active treatment or intervention.

• The ABAB design is one of the most commonly used single subject designs.

• ABAB refers to alternating between baseline/no treatment and treatment phases.

• The ABAB is both a single subject design and a within subjects design.

• It is also referred to as an intrasubject replication design.

• An initial baseline period (A) is done without any intervention.

• Followed by a period of treatment intervention (B).

• Followed by a return to no treatment (A).

• Followed by a second treatment intervention (B).

• Treatment-no treatment phases can alternate numerous times based on research needs.

• As an example, a child with ADHD might frequently leave his seat in a classroom setting.

• The teacher utilizes social reinforcement and praise when the child is sitting in his or her seat.

• The teacher can then record the number of times the child leaves the seat without reinforcement.

• Then the teacher should record the amount of times the child leaves the chair while offering social praise as reinforcement.

• Collected data (number of times child left classroom seat without permission) can be tabulated for each baseline for analysis of the intervention.

Multiple Baseline Designs

• Baseline data are collected for all the behaviors of interest.
• Then treatment is provided for one target behavior while baseline data are still collected on the other behaviors.
• Treatment intervention can then target a second behavior while continuing or removing treatment for the first behavior.
• A child with attention deficit disorder might not complete homework assignments and might also get into fights with other children.
• Reinforcements such as earning points for a gift might be used to target improvement in getting homework assignments completed.
• Reinforcement for the homework assignments is supplemented with social praise for playing cooperatively.
• If target behaviors do not improve, intervention strategies can be developed and implemented later on.

CORRELATIONAL METHODS

• Correlational designs examine the degree of association between two or more variables.

• While correlational designs do not allow cause-and-effect conclusions they can tell how closely two or more variables correspond.

• Correlations can be either positive or negative.

• A positive correlation refers to two or more variables that move in the same direction.

• For example, the more depressed someone feels, the more hopeless he or she may feel.

• A negative correlation refers to two or more variables that move in opposite directions.

• The more television a student watches, the less time he or she devotes to studying.

• None of these correlations or associations between variables is perfect.

• Some couples who argue frequently may report high satisfaction, while some students who watch a great deal of television may still study.

• A correlation coefficient expresses the degree of association between variables.

• The scores range from -1.00 to +1.00.

• Scores close to − 1.00 reflect a near perfect negative correlation.

• All the students who watched a lot of television studied infrequently while all of the students who did not watch television studied frequently.

• Scores close to +1.00 reflect a nearly perfect positive correlation.

• Those who tend to score high on depression are also likely to score high on hopelessness, while those who score low on depression are also likely to score low on hopelessness.

• There is no correlation between two variables if the correlation is close to .00.

• Association between variables does not mean that one necessarily caused the other.

• Correlation does not imply causality.

• For example, unmotivated students may watch television and spend very little time studying.

EPIDEMIOLOGICAL METHODS

• Clinical psychologists use epidemiological methods, which examines the incidence/distribution of a variable.
• It describes the prevalence/incidence of an issue of concern, answering questions such as # with alcohol problems?
• Data is collected from various sources including census, hospital records, and survey approaches.

Cross-Sectional and Longitudinal Designs

• Cross-sectional designs provide a "snapshot" view of behavior at a given moment in time.

• Because they are simple and less expensive, cross-sectional designs are the more popularly employed.

• As an example, test-anxious subjects were provided in a laboratory to see if treatment might be useful in reducing anxiety during one laboratory session.

• Longitudinal designs generally collect data over a long period of time

• Examining test-anxious subjects throughout the course of high school and college is an example.

• This exmaple takes approximately eight years to complete.

Treatment Outcome Research

• Treatment outcome research has become critical due to health care changes.
• Changes in the health care delivery has demanded treatment outcome research to show psychological and other treatments do work and justify funds.
• Different strategies are employed to conduct research dependent on interest.
• Strategies include treatment package strategy, dismantling treatment strategy, constructive treatment strategy, parametric treatment strategy, comparative treatment strategy, client and therapist variation strategy, and process research strategy.