12. Research Design and Statistics: A Foundation for Clinical Science

Question 1

Empiricism and Determinism

Answer 1

Science is based on certain philosophical assumptions including empiricism (knowledge based on sensory experience or objective observations) and determinism (events have causes, nothing happens without a cause)

Question 2

Inductive vs Deductive Method

Answer 2

Inductive: experiment-first-and-explain-later approach

Deductive: explain-first-and-verify-later approach; deductive reasoning first proposes a theory and then verifies it

Question 3

Null vs Alternative Hypothesis

Answer 3

Null: two variables are not related

Alternative: variables are related; cause-and-effect

Question 4

Reliability and Validity: Definition and Types

Answer 4

*Important for good scientific measurement

See Chapter 11’s Flashcards

Question 5

Experimental Research: Independent vs Dependent Variables

Answer 5

Independent: directly manipulated
Dependent: affected by independent variable

Independent/Dependent Examples:

• Stressful situation/amount of stuttering
• Amount of b/g noise/amount of diff. hearing speech

Question 6

Types of Single Subject Design (2)

Answer 6

*AB design is similar to a case study; baselines are established (A) and treatment is offered (B), and the progress is summarized in a case study
*Two types: ABA and ABAB
A=baselines, B=treatment, A=treatment withdrawl, B=reinstatement of treatment

• Multiple-Baseline design: effects of treatment are demonstrated by showing that untreated skills did not change and only the treated skills did
• 3 types: across subjects, settings, and behaviors

Question 7

Experimental vs Descriptive Research

Answer 7

Experimental: permits us to actively manipulate conditions; cause-and-effect

Descriptive: helps us understand what already exists

Question 8

Descriptive Research: Classification vs Criterion Variables

Answer 8

These variables in some kinds of descriptive research

• Classification variable: analogous to the independent variable in experimental research
• Criterion variable: analogous to the dependent variable in experimental research

Classification/Criterion Example:
Group status (dementia vs not)/Performance measures

Question 9

Descriptive Research: Types (6)

Answer 9

• Ex post facto research AKA causal-comparative studies (retrospective or case studies)
• Survey research (questionnaires, interviews)
• Comparative research AKA standard-group comparisons
• Developmental research AKA normative research (independent variable: maturation; 3 methods: longitudinal, cross-sectional, and semilongitudinal)
• Correlational research (correlation does not imply causation; direction: pos/neg and strength: high/low of relationship; Pearson r)
• Ethnographic research (observation and description of naturally occurring phenomena)

Question 10

Evaluation of Research: Internal vs External Validity

Answer 10

Internal: Degree to which data in study reflect a true cause-effect relationship; no confounding variables

External: Generalizability; “To which people and situations can these results be generalized?”; degree to which internally valid results can be generalized

Question 11

Threats to Internal Validity (8)

Answer 11

• Instrumentation (mechanical/electrical probs, inappropriate testing materials, judges may reduce validity and make observation/scoring mistakes)
• History (life events; e.g., recent surgery)
• Statistical regression (a behavior that goes from an extreme high/low point to an average level; e.g., hoarse voice may change due to statistical regression vs tx)
• Maturation
• Attrition (losing participants as experiment progresses)
• Testing (change in dependent variable because it has been measured more than once, e.g., administration of pre- and posttests)
• Subject selection biases
• Interaction of factors

Question 12

Threats to External Validity (3)

Answer 12

• Hawthorne effect (extent to which results are affected by participants’ knowledge that they are taking part in an experiment or that they are being treated differently than usual)
• Multiple-treatment interference (incl. “order effect”)
• Reactive or interactive effects of pretesting (e.g., may increase awareness prior to tx)

Question 13

Statistical Techniques for Organizing/Summarizing Data (3)

Answer 13

• Measures of central tendency: mean, median, mode
• Measures of variability: refers to dispersion in data set; e.g., range, interquartile range, standard deviation
• Measures of association/correlation