Chapter 3 Research Methods Used to Study Child

Chapter 3 Research Methods Used to Study Child Behavior Disorders

Science and Common Sense • Science is an extension of common sense that emphasizes systematic questioning

Research Questions • Descriptive Research Questions ask about the nature of a phenomenon • Difference research questions ask if there is a difference between phenomena • Relationship research questions (correlational questions) asks to what degree phenomena are related

Research Paradigms: Distinctions, Descriptions, and Illustrations

Methodology Distinctions • Experimental and non-experimental methods are mirror images of one another in some ways – Experimental – investigator manipulates the treatment given to a participant in an environment that is carefully controlled – Non-experimental – investigators observe, analyze and describe phenomena as they exist

Quantitative and Qualitative Methods • Quantitative – conducted in controlled a way that results can be reduced to numerical presentation • Qualitative – emphasizes reluctance to intervene in naturally occurring phenomena

Quantitative vs Qualitative • Quantitative – Usually occurs in more controlled environment, so researcher can be more confident in ascribing results to treatment variable – Controlled environment more artificial and can make participants differently than in their normal setting

Quantitative and Qualitative • Qualitative – Usually occurs in more natural and less controlled environment, which are less likely to influence participant behavior – Relative absence of control can contribute to unreliable data

Quantitative Methods • Characterized by manipulation of the treatment or condition under study – Factor manipulated is independent (experimental) variable – Measure, means by which treatment effect is determined is dependent variable

Quantitative: continued • Preference for precise hypothesis state at outset • Preference for precise definitions state at the outset • Data reduced to numerical scores • Much attention to assessing and improving reliability of scores obtained from instruments

Quantitative: continued • Assessment of validity through variety of design procedures with reliance on statistical indices • Preference for random techniques for obtaining meaningful samples • Preference for precise descriptions of procedures

Quantitative: continued • Preference for specific design control for procedural bias • Preference for statistical summary of results • Preference for breaking down of complex phenomena into specific parts for analysis • Willingness to manipulate aspects, situations, or conditions

Qualitative Methods • Nonexperimental involves observation, analysis and description of phenomena rather than manipulation of treatment variables

Qualitative: continued • Observation – data collection strategy that can be used in several research methods – Vary in degree of involvement of researcher • Participant observation – observer takes part in activities and environment being observed • Nonparticipant observation – observer is not or is only minimally involved in observed setting

Qualitative: continued • Case Studies – in-depth examination of the behavior of an individual or small social unit – Provide important depth, complexity and quantity of information – Information may not be generalizable – Vulnerable to investigator bias

Qualitative: continued • Preference for hypotheses that emerge as study develops Preference for definitions made in context or as study progresses • Preference for narrative description • Preference for assuming that reliability of inferences is adequate • Assessment of validity through cross-checking of sources of information (triangulation)

Qualitative: continued • Preference for expert information samples • Preference for narrative/literary descriptions of procedures • Preference for logical analysis in controlling or accounting for extraneous variables (describing what else seems to be going on) • Primary reliance on researcher to detect and minimize procedural bias

Qualitative: continued • Preference for narrative summary of results • Preference for holistic description of complex phenomena (describing the whole picture) • Unwillingness to tamper with naturally occurring phenomena

Quasi-Experimental Designs • Investigations in which participants cannot be randomly assigned to groups – Participants history cannot be controlled

Studying Development • Longitudinal design – selects a sample of participants for an extended period of time, repeating assessment intermittently • Cross-sectional design – simultaneously samples different groups of participants at several age levels and compares dependent variables across the age groups

Longitudinal Designs: limitations • Participants development may be altered by the repeated assessments – “test-wise” • Participant attrition – participants move away, refuse to continue or die • Measures originally used may become dated in very long studies

Cross-sectional Designs: limitations • Different cohorts or groups are being compared • Age range can be so great that sociocultural or historical changes have been substantial

Time-Series Designs • Investigations in which an independent variable is manipulated across two or more phases and the dependent variable is monitored at each phase • Used to assess the effect of treatment on the behavior of a small number of participants • Clinicians can treat individuals while systematically colleting information • Many measurements are collected over time

Time Series Designs: limitations • Measurements are taken on small number of individuals (or even one) – Results may not be generalizable – Types of participants are frequently atypical

Group Experiments • More participants are included than in timeseries • Small number of measurements • Types of group designs – Repeated measures design (pretest and posttest comparison) – Multifactor designs – include two or more experimental variables

Meta-Analysis • Facilitates the analysis and interpretation of multiple results from different studies • Allows researchers to statistically analyze and synthesize the findings of many previous empirical studies – Uncover consensus – Can present integrated picture of overall results

Fundamentals of Research Design

• Control – eliminating the systematic influence of all variables except the one being studied – All factors must be equivalent for both groups except the independent variable – Procedures must be equivalent for both groups

Common Design Mistakes • Internal validity – technical soundness of an investigation in terms of control • External validity – experiments generalizability, how well can results be applied to other participants, settings and treatments

Common Design Mistakes • Placebo effects – changes in participants’ behavior or performance that occur simply because they are in an experiment and not because of a particular treatment or intervention

Avoiding Design Pitfalls • Random Sampling – Researchers use a selection process in which each individual in the population has an equal chance of being chosen to participate in a study – It is assumed that since each person has an equal chance of selection, that the characteristics of the participant sample will represent those of the entire population

Avoiding Design Pitfalls • Experimental Matching – Procedures by which a researcher forces group equivalence in terms of characteristics thought to be important for the particular study being conducted

Ethical Issues in Conditioning Research • Ethical issues critical because psychology involves people working with other people and controlling aspects of their lives

Protecting welfare of participants • Harm can be physical or psychological stress, or social embarrassment • Cannot completely eliminate stress of any kind for participants – Potential harm weighed against potential benefits of research

Informed Consent • Capacity – persons ability and legal authority to consent to participate in a research project – Children may not be old enough to give legal consent or cannot give consent because of mental disorders

• Information – Investigators must provide complete, and clearly communicated information regarding the study – Participants with behavior disorders may have difficulty understanding information no matter how well it is communicated

• Voluntariness – Participants must participate of their own free will without coercion – Participants may feel coercion simply because the researcher may represent a power figure for them