Module 3 - Research Methods Flashcards
What are the 3 components of Evidence-based practice?
Empirical Science: best tool for reliable & valid knowledge
Duty of care: base treatments on evidence
Scientific literacy: essential to understand research
Theory
Hypothesis or explanation
Must prove true or false
Empiricism
measurable
Observable
Can be directly or indirectly measured
Falsifiable
Theory can be rejecting or supporting
The Scientific Method Process (3)
Theory, hypothesis is tested empirically (measured)
Results -> falsify / support hypothesis
Reject/ or tentatively accept theory
2 main aims of research
Description
Explanation
Description: 2 research types
Descriptive research
Correlational research
Explanation: 1 research type
Experimental research
Descriptive research features (3)
Detailed
Systematic
Ethical
Case Studies (2)
Descriptive research
In-depth observation of a small number of cases
Uses & Advantages
of Case Studies (3)
Describes processes in individual cases
Studies complex phenomena
Provides data for framing hypothesis
Limitations
of Case Studies (4)
Generalisability of the population
Not replicable
Researcher bias
Cannot establish causation
Naturalistic Observation
In-depth observation of a phenomena as it occurs in nature
Uses & Advantages
of Naturalistic Observation (3)
Reveals phenomena as they exist outside laboratory
Allows study of complex phenomena not easily reproduced experimentally
Provides data for framing hypothesis
Limitations
Of Naturalistic Observation (4)
Generalisability of the population
Observer effects: presence of observer may affect behaviour of participants
Not replicable
Cannot establish causation
Survey
Ask people questions about their attitudes and behaviours
Uses & advantages
of surveys (2)
Reveals attitudes and behaviours or a large sample of people
Allows quantification of attitudes or behaviours
Limitations
Of surveys (2)
Self report bias
Cannot establish causation
Correlational Research (2)
Examines the relationship of two or more variables
Used to predict one another
Uses & advantages
Of correlational research (2)
Reveals relationship between variables as they exist outside of the laboratory
Allows quantification of relationships between variables
Limitations
Of correlational research
Cannot establish causation
Correlation Coefficient (r) shows:
Direction (positive or negative)
Magnitude
0-0.1 = no trend
0.1 - 0.3 = weak trend
0.3 - 0.5 = moderate
0.5+ = strong
Relationship between 2 variables
Correlation Coefficient (r)
0
No relationship
Correlation Coefficient (r)
1
Perfect, Positive, linear relationship
Correlation Coefficient (r)
-1
Perfect, Negative, linear relationship
Correlation Coefficient (r)
-0.3
Moderate, negative relationship
Correlation Coefficient (r)
-0.1
Weak, negative relationship
Scatterplot features
(Correlation research)
2 variables plotted against 2 axes
Strength of correlation (magnitude) depends on how closely the dots sit to the hypothetical line
Correlation Coefficient r=+1
/
Perfect, positive
Correlation Coefficient r=-1
\
Perfect, negative
Correlation Coefficient r=-0.3
\ scattered dots
Weak, Negative
Correlation Coefficient r=-0.4
\
Moderate, negative
Correlation Coefficient r=0.5
/
Strong, positive
Correlation Coefficient r=0
No relationship
Third Variable Problem
Limitation of Correlational research
Correlation ≠ Causation
Eg. Just because there is an increase with ice cream sales with shark attacks does not mean ice cream consumption causes shark attacks
There is a possibility of a third variable
Experimental research features (3)
Manipulation of variables (Independent Variables)
Measure outcomes (dependent variables)
establish causal relationship
Uses & Advantages
Of Experimental Research
Demonstrates causal relationships
Replicable
Maximum control over relevant variables
Limitations
Of experimental research
Generalisability outside of lab
Difficulty in testing complex phenomena
Randomized Control Trial (RCT) (4)
(Experimental Research)
Gold standard in experimental research
Random allocation of control and experimental groups
Variables (age, gender) are randomly spread across groups
Double blind design reduces bias
Double blind experiment:
Participant and doctors are blind
Single blind
Participants are blind
Reliability
How consistent is the result going to be?
Retest reliability
A measure is reliable if it yields similar results after retest
Internal consistency
A measure is internally consistent if several ways of asking the same question yields similar results
Interrater reliability
If 2 different interviewers rate the same person, they should give the person a similar score
Validity
How relevant is the study?
Internal Validity
Have all of the likely possible confounding variables been identified, controlled, or eliminated?
External Validity
Has the research been conducted in such a way, that the results will generalize to the ‘real world’?
