Outline Lecture 3 Jan 2103 Chapter 3 Research

Outline: Lecture 3 – Jan. 21/03 [Chapter 3 - Research] • • • The Scientific Approach The Research Process Common Research Topics Forms of Measurement Defining Research Terminology General Approaches to Research Designs Research Methods Qualitative Research Ethical Issues

The Scientific Approach • Theory - propositions that explain phenomena. • Hypothesis – an educated guess; a tentative assumption • Purpose of Scientific Approach: to describe phenomena and offer explanations for them. • Reasons for skepticism of research: » experts will often disagree » research findings are often in conflict » research will sometimes lead to different recommendations » research conclusions are often qualified

The Research Process

Common Research Topics 1. Nature & Distribution of Childhood Disorders 2. Risks, Causes, and Correlates 3. Moderating and Mediating Variables – See Figure 3. 3 in text 4. Outcomes associated with childhood problems 5. Interventions – Treatment efficacy – Treatment effectiveness

Forms of Measurement 1. Reporting – includes unstructured clinical interviews, highly structured interviews, and questionnaires – inaccuracies may occur because of inability to recall events, selective recall or bias, and intentional distortions – requires a certain level of verbal ability, therefore often not considered reliable with children under age 7 or 8

Forms of Measurement [cont’d] 2. Psychophysiology and Neuro-imaging – physiological responses recorded include: heart rate, blood pressure, respiration, pupil dilation, and electrical skin conductance – limitations of physiological measures include: inconsistency, high level of inference often involved, and susceptibility to extraneous influences – electroencephalogram (EEG) measures electrical activity of the brain – neuroimaging procedures examine the structure and/or function of the brain

Forms of Measurement [cont’d] 3. Observation – may range from unstructured, naturalistic observations to highly structured – compared to naturalistic observations, structured observations are often cost-effective, allow for focused attention, are useful for studying infrequent behavior, and allow for greater control over the situation – major limitation is that one cannot be certain the observations are a representative sample of behavior

Identifying the Sample • A careful definition of the sample is critical for comparability of findings across studies • Must consider comorbidities among the sample • Random selection is rare in child psychopathology studies; often need to use a sample of convenience

Defining the Variables of Interest • A variable must be defined so that it can: 1. Be measured consistently over time. 2. Be measured consistently across researchers. 3. Provide an accurate measure of the behaviour of interest. » Operational definition: a description of a construct in objective, measurable terms.

Research Terminology • Standardization: the process by which a set of norms is specified for a measurement procedure so that it can be used consistently across different assessments • Reliability: the consistency of a measure, either across raters or time • Types of reliability: • inter-rater agreement - having multiple raters rate the same item, and then comparing their data. • test-retest reliability - whether the same results can be achieved when a measure is used more than once. • internal consistency - how well the questions within a measure relate to each other.

Research Terminology • Validity: the extent to which the method actually measures the construct of interest • Types of validity: – face validity - the appearance of validity. – convergent validity - a relationship between two measures of the same topic. – discriminant validity - lack of a relationship between two measures of different topics. – criterion-related validity - predictive ability either in the present (concurrent validity) or in the future (predictive validity).

Validity • Internal validity: – the extent to which change in the dependent variable is due to the manipulation of the independent variable. – may be threatened by maturation, effects of testing, subject selection biases • External validity: – the extent to which findings of a particular experiment can be generalized to other subjects under other circumstances. – may be threatened by subject reactivity to participation, the setting, and the time measurements are made

General Approaches to Research • Experimental vs Non-experimental Research – In true experiments, researchers have maximum control over the independent variable, subjects are randomly assigned, and possible sources of bias are controlled – can determine cause and effect – Correlation studies only examine relationships among variables- causality cannot be determined – Natural experiments involve comparisons between conditions that already exist

General Approaches to Research [cont’d] • Prospective vs Retrospective Research – Real-time prospective designs: sample is followed longitudinally over time [follow-forward method]; time consuming and susceptible to sample attrition Retrospective designs: sample is asked for information relating to an earlier time period [followback method]; highly susceptible to recall bias and distortion • Analogue Research – evaluation of a specific variable under conditions that only resemble the situation to which one wishes to generalize

Research Designs • Between-Group Comparison Studies – involves comparisons between experimental and control groups • Cross-Sectional Studies – individuals at different ages or stages of development are studied at the same point in time – often efficient and less susceptible to attrition and practice effects – do not allow for inferences regarding change and are susceptible to cohort effects

Research Designs • Longitudinal Studies – same individuals are studied over time Strengths: can see how behaviour changes over time helps eliminate 3 rd variable factor Weakness: attrition expensive sociocultural changes over time are measures valid at all ages subjects become test-wise Accelerated Longitudinal Studies [Sequential] combination of longitudinal and cross-sectional approaches – support for developmental change over entire age range – control for societal change

Non-experimental Methods • Case Studies: intensive and usually anecdotal observations and analyses of an individual child – Strengths: rich in detail – aids intervention planning Weakness: investigator bias [reliability issue] not generalizable [validity issue] 2. Observational Studies: Naturalistic or Controlled – – Strengths: good for generating hypotheses Weakness: inter-rater reliability observer bias subject reactivity

Non-experimental Methods 3. Correlational Research - where the relationship or degree of association between variables is examined. – Correlation Coefficient (Pearson r) - describes the strength and direction of a relationship between two variables; ranges between -1 and +1. – Strengths: useful in generating hypotheses and when ethical considerations preclude manipulations – Weakness: cannot rule out third unknown variable statistical vs clinical significance

Correlation - Examples

Experimental Methods 1. True Experimental Design – Random assignment to groups – independent variable [IV] – dependent variable [DV] 2. Quasi-experimental & Mixed – Existing groups [e. g. age] – manipulation of IV Strengths: can determine cause & effect Weakness: cannot be used when ethical considerations preclude manipulation weak in external validity

3. Single-Case Experimental Designs – often used to evaluate the impact of treatments – involves repeated assessment of behavior over time, replication of treatment effects within the same subject, and the subject serving as own control – common examples are the A-B-A-B (reversal) design and the multiple-baseline design – weaknesses include possible interactions between treatment and subject characteristics, limited generality of findings, and subjectivity and inconsistency of visual inspection of the data

Single Subject Designs A-B-A-B (reversal)

Single Subject Designs

Qualitative Research • Purpose is to describe, interpret, and understand the phenomenon of interest in the context in which it is experienced • Although intensive and intimate, may be biased by researcher’s values and preferences, and findings cannot be generalized to other individuals

Ethical Issues 1. Informed Consent and Assent – informed consent must be obtained from parents – child’s assent must be obtained when child is around age 7 or older 2. Voluntary Participation – participation in research must be voluntary – may be compromised by subtle pressure and coercion

Ethical Issues 3. Confidentiality and Anonymity – disclosed information must be kept confidential – individuals must be advised at the outset of any exceptions to confidentiality – disclosures of abuse common problem in child research 4. Non-harmful procedures – no research procedures may be used that may harm a child physically or psychologically