Aud 712 Research Methods

Front 1

Science

Back 1

Latin antecedent from the word "scire" (to know). One means of obtaining knowledge. Science best described as a process.

Front 2

Belief

Back 2

Conviction of truth of some statement or the reality of some being or phenomenon especially when based on examination of evidence; beliefs are potentially testable. "If you don't brush your teeth after every mean, you will get more cavities than if you do brush them."

Front 3

Opinion

Back 3

A view, judgement, or appraisal formed in the mind about a particular matter. "You should brush your teeth after every meal."

Front 4

Advice

Back 4

Recommendation regarding a decision or course of conduct. "Brush your teeth after every meal."

Front 5

Intuition

Back 5

The power or faculty of obtaining your knowledge or cognition without rational thought or inference. A hunch. Bypassing ordinary ways of reasoning.

Front 6

Method of tenacity

Back 6

Holding onto something. Beliefs that are held because this has always been know to be true. Beliefs may exist contrary to evidence. Repetition of these beliefs reinforce them.

Front 7

Method of authority

Back 7

Holding a belief because someone told me it is true.

Front 8

Rationalism

Back 8

Knowledge gained through reasoning based on assumptions. If the assumptions are wrong the reasoning will not be correct.

Front 9

Empiricism

Back 9

Knowledge based on observation and sensory evidence. "Seeing is believing." Lead to the scientific method.

Front 10

Inductive reasoning

Back 10

Issac Newton (late 16th C).
Start with observations. Observations lead to formulation of inferences and theories to explain observations.

Front 11

Deductive reasoning

Back 11

Francis Bacon (early 16th C).
Starts with hypothesis or idea - then deduces what we would expect to find based on the idea. General premise is used to explain observations. Eg. Premise: Gravity makes things fall. Observation: the apple hit me on the head. This was due to gravity.

Front 12

Basic components of science

Back 12

Data/observations: collect information.
Hypothesis: tentative explanation of observations.
Theory: explanation that has been thoroughly tested. Explanatory- explains WHY data and patterns exist. Phenomenological- describe process (phenomena) without explanation. Eg. Laws: "every action has an equal and opposite reaction."

Front 13

Scientific Process

Back 13

Observations --> Hypothesis --> Predictions --> Experiment --> Data --> Formulate/revise theory --> Observations

Front 14

Scientific process: shaping principles

Back 14

Non-empirical factors involved in selecting a "good theory."
Examples: Law of parsimony (simplest explanation). Empirical adequacy (explains the greatest amount of data). Other factors: self consistency, ties in with other theories.

Front 15

Audiologist and clinical practice

Back 15

Old-school clinician: disconnect between clinic and research. "What I do in clinic doesn't have anything to do with research." Primary method of acquiring knowledge. No research collaborations.
New: Clinical scientist. Identify new research needs. Critical evaluation of research articles. Application of research to practice. Scientific contributor.

Front 16

Evidence Based Practice

Back 16

Framework of constructs and methods that originated in clinical medicine. "Evidence-based medicine is the integration of best research evidence with clinical expertise and patient values." "The conscientious, explicit and judicious use of current best evidence in making decisions about the care of individual patients. This involves the integration of clinical expertise with the best available external evidence."

Front 17

Goals of EBP

Back 17

Clinical expertise.
Best current evidence.
Client values and choices.

Front 18

EBP Rationales

Back 18

Substantiate what we do. Establish and maintain professional credibility. Convince their patients and 3rd party payers that our service options provide benefits that justify the cost. Satisfy accrediting agencies (JCAOH). Avoid mismatch between best evidence and actual practice patterns.

Front 19

EBP is a process

Back 19

Constant process of inquiry: "Why am I doing things this way?" "Is there evidence that can guide me to a better outcome?"

Front 20

EBP Steps

Back 20

1. Formulate a question
2. Search for evidence
3. Critical appraisal of the evidence
4. Implementation (evidence, expertise, pt values)
5. Follow-up

Front 21

EBP Step 1

Back 21

Asking questions.
PICO
Patient - define patient characteristics of interest.
Intervention - define intervention (treatment).
Comparison - define comparison condition (optional).
Outcome - define outcome measure (must be specific).

Front 22

Behavioral data

Back 22

Test scores (e.g. % correct speech recognition).
Ratings (e.g. quality of sound, satisfaction ect.)
Preference (e.g. prefer A over B).
Categorizing behavior - frequency of responses.
Response time.

Front 23

Physiological data

Back 23

Amplitudes, latencies/duration, histograms, threshold (physiological).

Front 24

Other types of EBP questions

Back 24

Diagnosis - test selection and interpretation.
Harm or etiology - harmful effects of a Tx? How can these be avoided?
Prognosis - what is the patient's likely course of disease, or how to screen for or reduce risk.
Prevention - how can the patient's risk factors be adjusted to help reduce the risk of disease?
Qualitative - helps to understand clinical phenomena with emphasis on understanding the experiences and values of patients.
Diagnosis - e.g. in patients with suspected depression what is the accuracy of a two-question case-finding instrument for depression compared with six previous validated instruments.

Front 25

Two types of quantitative research

Back 25

Descriptive: subject characteristics, relationships, no manipulation.
Experimental: manipulation of variables.

Front 26

Broad categories of research

Back 26

Qualitative and quantitative

Front 27

Variables: categorized according to how they are used

Back 27

Research variables: independent and dependent.
Measurement variables (may be independent or dependent): qualitative (categorical) or quantitative (numerical).
Extraneous variables: exert an unwanted influence.

Front 28

Research Variables

Back 28

Dependent - variable being measured.
Independent - active variable (experimental): manipulated, must have two or more levels (e.g. HA type). Attribute variable (experimental or descriptive): not manipulated.

Front 29

4 types of descriptive research

Back 29

Comparative research
Naturalistic research
Survey research
Correlation research

Front 30

Comparative research

Back 30

Involves "comparison of subject characteristics, cultures or political systems." Two disadvantages: Difficulty attributing causation. Categorical comparisons - lose information if categories are based on underlying numerical info (severity vs. PTA).

Front 31

Naturalistic research

Back 31

Involves the "observations of behavior and/or events that occur in a natural setting." Includes developmental research - very common in language acquisition studies. Can be qualitative or quantitative.

Front 32

The role of naturalistic observation

Back 32

Descriptive provides data base that can lead to subsequent, more highly controlled research. Two problems threaten soundness of observations: 1) delimiting choice of behaviors to observe. 2) Reactivity. Controlling threats: Unobtrusive observations and measurements (e.g. evidence of behavior - graffiti, tracks.)

Front 33

Advantages and disadvantages of naturalistic observation

Back 33

Advantage: can help define the problem and raise questions for future controlled studies.
Disadvantages: does not allow us to assess relationships among events. More difficult to reproduce. Difficulty maintaining a descriptive rather than interpretive level of observation.

Front 34

Developmental research

Back 34

Measures changes in behavior or subject's characteristics over time. May be conducted in naturalistic or lab setting. Three types of developmental plans: 1) cross-sectional, 2) longitudinal, 3) semi-longitudinal.

Front 35

Correlational research

Back 35

To study the relationships among two or more variables. Two main questions: 1) How close is the relationship between the variables (and what is the nature of the relationship)? Answered by the correlational coefficient. Explain positive and negative correlation. 2) How well can performance on one variable be predicted from knowledge of another variable? (regression analysis).

Front 36

Advantages and disadvantages of correlational research

Back 36

Advantage: Can make predictions if relationship is strong. Can't do this with comparative research.
Disadvantages: Does not imply causation. Hidden variables may exert an influence.

Front 37

Survey research

Back 37

Assembling information about characteristics, practices and opinions. Broad categories: attitude surveys, demographic/lifestyle surveys.

Front 38

Case study research

Back 38

Designed to examine specific individual in depth. To illustrate important principles that might be overlooked in group data. To highlight unusual cases. To explore new therapy technique.

Front 39

Retrospective research

Back 39

Compiling data from past clinical records. Potential problems: Reliaibility issues - how were data collected and by whom? Data may be missing or difficult to quantify. Data may be collected differently by different people.

Front 40

Two sources of variance

Back 40

Error variance and systematic variance.

Front 41

Error variance

Back 41

Variability in measures caused by unsystematic random fluctuations (due to chance). Factors affecting error variance: 1) inconsistency of measurement, 2) inconsistency of subject responses, 3) inconsistency of test conditions (other than those planned).

Front 42

Systematic variance

Back 42

Variability in measures caused either by variables under study or some other extraneous variable.

Front 43

In selecting a particular measurement for an experiment, these factors should be considered

Back 43

1) Measurement utility
2) Reliability
3) Validity

Front 44

Measurement utility

Back 44

Test sensitivity - test is sensitive when it frequently identifies the presence of a disorder when it is truly there.
Test specificity - test is specific when it rarely identifies a person as having a disorder when the person truly does NOT have it (low false alarm rate).

Front 45

Reliability

Back 45

AKA: reproducibility.
Precision of measurement which can be assessed by examining the consistency or stability of a test or measure. The degree that test scores are free from errors of measurement.

Front 46

4 types of reliability

Back 46

1) Test-retest
2) Internal consistency
3) Equivalent forms
4) Observer

Front 47

Test-retest reliability

Back 47

"Consistency/stability of a test measure over time." Assessed by determining how well scores are correlated over time - high correlations = good reliability. Scores that are different on two occasions - does not necessarily mean poor reliability. What does low test-retest reliability indicate?

Front 48

Internal consistency

Back 48

Consistency of items within a test. Assessed by evaluating the split-half reliability: scores on first half of test compared to the second half or odd items compared to even items.

Front 49

Equivalent forms

Back 49

AKA: alternate forms reliability.
Is performance equivalent across two or more forms of the same test?

Front 50

Observer

Back 50

Intra-observer reliability: consistency of observer judgements within an observer.
Inter-observer reliability: consistency of judgements across observers.

Front 51

Validity

Back 51

The degree to which a test measures what it purports to test.
A valid test must be reliability but a reliable test is not necessarily valid.

Front 52

Content validity

Back 52

How well does the test sample the targeted characteristics? Ex. Interested in measuring child's use of irregular verbs. Do the questions elicit the use of irregular verbs?

Front 53

Criterion validity

Back 53

How well does the test correlate with a known indicator of the behavior or characteristic it is supposed to measure?
Predictive validity - How well does a test predict a particular outcome?
Concurrent validity - When a test and outside validating measure are done at the same time.

Front 54

Construct validity

Back 54

How well does the test reflect a theoretical construct (or explanation) of the characteristic/behavior being measured?
Convergence evidence - Is the new test supported by old measures?
Discriminant evidence - Does test have correlations you would not expect?

Front 55

Functional validity

Back 55

How the tests are used and interpreted.

Front 56

Experimental research

Back 56

To purposes of experimental research design: 1) to answer the research question(s), 2) to minimize contamination of results by extraneous variable. Aims of experimental research: 1) manipulate independent variable(s), 2) hold other potential independent variables constant.

Front 57

Descriptive research

Back 57

Tends to be messier, more difficult to control extraneous variables.

Front 58

Internal validity

Back 58

The degree to which the design has accomplished what it was intended to accomplish. Experimental research: Need to be reasonably certain that change in dependent variables can be attributed to manipulation of independent variable(s) & not by extraneous variable(s). The fewer the alternative explanations, the greater the internal validity.

Front 59

External validity

Back 59

Concerns the question of generalizability. To what populations, settings, measurement variables can this effect be generalized?

Front 60

Within-subject designs

Back 60

A single group of subjects tested under more than one condition.

Front 61

Between-subject designs

Back 61

"Group design"
More than one group of subjects tested under one condition.

Front 62

Mixed design

Back 62

"Group design"
More than one group tested under different conditions.

Front 63

Threat to internal validity: unmatched groups

Back 63

Only applicable with 2+ groups.
Subject groups should be equal on important factors. For single-group designs it will just add noise to the data and the design will be less sensitive.

Front 64

Threat to internal validity: attrition

Back 64

AKA "experimental mortality"
Differential loss of subjects in different groups used in study. Problem in the design requires carefully matched groups. Not a problem if different groups are not being compared. Can be serious when pairs of subjects are matched across groups.

Front 65

Threat to internal validity: interactions among groups

Back 65

Compenatory Rivalry: rivalry between control and experimental group. Experimental group may get "extra." May affect behavior of controls (competitiveness). Resentful demoralization: may cause loss of motivation. Opposite effect of compensatory rivalry. Solution: avoid exposing groups to one another if possible.