Please enable JavaScript.
Coggle requires JavaScript to display documents.
Great PSY1010 (Experimental Methods 8. (Validity in Experiments (Test…
Great PSY1010
Experimental Methods 8.
Topics of Lecture
- Understand Humans and their Behavior
- to find Causal (Cause-Effect) Relations
- Variation as a Prerequisite to Measure Differences
- What is Experimental Design
- Validity in an Experiment
-
-
-
Validity in Experiments
Test Validity
- Content Validity
- Criterion Validity (Concurrent, Predictive)
- Construct Validity (Convergent, Divergent)
Inferential Validity
- Internal Validity
- External Validity
Internal
Validity
Definition
Formal
- whether Researchers have Drawn the Correct Conclusions on the Conditions of Cause-Effect in the Experiment
- Inferences have Internal Validity if a Causal Relation btw 2 Variables is Demonstrated
- the extent to which a Causal Conclusion based on a study is Warranted, which is determined by the degree to which a study Minimizes Systematic Error (or 'Bias')
- 1 more item...
Personal
- when we Control all Extraneous Variables and the Only Variable Influencing the Results is the One being Manipulated (Indipendent V.)
and Not other Extraneous Variables
- the Degree to which the Experiment supports Clear Causal Conclusions
- 1 more item...
Threats
Between-Subjects Exp.
-
Selection Bias
- 1 more item...
-
Within Subjects Exp.
Carry Over
Effect
- 2 more items...
-
-
Statistical Regression
- 1 more item...
-
-
-
N=1 Design 5.
Topics of Lecture:
- what is Experimental Design
- what Characterizes N=1 Design
- 3 main types of N=1 Design
Validity
Internal V.
Formal
the extent to which a Causal Conclusion based on a study is Warranted, which is determined by the degree to which a study Minimizes Systematic Error (or 'Bias')
External V.
Formal
the degree to which a Causal Conclusion is Warranted to Generalize Results to Other Contexts (People or Situations)
-
N=1
Features
- just 1 Subject
- a type of Experimental Design
- Not the Same as Case Studies
- the Subject is his own Control
Historical Examples
- Thorndike Cat's trap
- Pavolv's Dogs
- Ebbinghaus
Use Today
- Pilot Study
- as a Follow-Up to a Study made w Groups
- to Test Individual-fit intervention in Therapy research
3 Main Types:
Baseline Design
2 Phases
-
-
AB, ABA, ABAB design
- A: Baseline Phase
- B: Intervantion Phase
-
Baseline Behavior
Ideal
Stable: Flat, not up and down
-
Challenges
-
Too High, too Low (ex. too depressed)
to track modificationstext
-
-
Multiple Baseline Design
- More than 1 Indipendent Variable
- More than 1 Behavior that can be
Influenced by Manipulation
Multifactor Design
- Manipulation w Diff. Indipendent Variables
- look at the Effect of a Combination of Ind. Var.
-
-
Correlation 6.
-
-
Correlation
Definition
Formal
- a Measure of how 2 Variables
Change in Relation to Each other
- Covariance: how Similarly 2 Variables Vary
- Reciprocal Dependence, Relation
Personal
- a Measure of how 2 Variables Vary Compared to Each other
- Strenght of a Relation btw 2 Variables
Pros
- Quantifying and Measuring Relations
- Naturalistic (in the Real World)
- Replicable
Cons
- Difficult to Isolate Causes (No Control/Manipulation)
- Causality Problems
- Cause Direction? (X->Y / Y->X)
- Third Variables
- Often Retrospective: first find Correlation
and Then ask: Why? La scienza a posteriori ....
-
-
Goals / Application
Measure Smth w
Observation, Interview, Questionnaire
-
Covariance
Definition
- 2 Variables that Vary Together
- how 2 Variables Change when Compared to One Another
-
-
Correlation (r)
-
Values
Conventionally:
- 0.00 - 0.30 Small or No C.
- 0.30 - 0.50 Low
- 0.50 - 0.70 Moderate
- 0.70 - 0.90 High
- 0.90 - 1.00 Very High
-
-
Describe Data 2.
Variable
Concept
must Vary
2 main variants:
Quantitative variables
-
must Vary
- not in Time
- but in its Values
-
-
-
Measurements levels:
Scales
NOI:
- Null Point
- Order
- Interval
NOIR
Ratio
- Yes, clear Order
- Yes, constant Interval
- Yes, absolute null point
Interval
- Yes, clear Order
- Yes, constant Interval
- No absolute null point
Ordinal
Ex.:
- Likert scale: bad, quite bad,
decent, good, very good
- Yes, clear Order
- No constant Interval
- No absolute null point
Nominal
Ex.:
- colors
- philosophical traditions
- No obvious Order
- No constant Intervals
- No absolute null point
-
-
Population and Sample 4.
Population
-
Sampling Frame
-
Sample
-
Formal
- A relatively small part of the population
- A Subset of the Population
- Equals or minor the memebers in the Samplig Frame
-
-
-
Correlation (2) 7.
-
Regression
-
-
Linear Regression
Formula
- Y-hat: Predicted Score
- X: the Score we Use to Predict
- b: Regression Coefficient (Slope)
- a: Regression Constant (Y-Intercept, Cross Value on Y axis)
-
-
Definition
Formal
- Regression uses 1 Independent Variable to Explain or Predict the Outcome of the Dependent Variable Y
- a Statistical Analysis that uses 1 Indipendent Variable to Predict values of Y
Ex. Given IQ scores (X),
want to Predict University Grades (Y)
Regression Line
- 1 more item...
2 sources of Variance
- 2 more items...
-
-
Test and Measurement 3.
Psychometrics
Definition
Formal
Measurement of Psychological variables, through:
- Testing, measurement, assessment, etc.
Personal
-
Problems:
- No Psychological Construct is Directly measurable
- Difficult to Define psych. constructs
Ex.
- Emotions, Anxiety, Neuroticism,
Psychological
Tests
Definition
Personal
we put ppl. in a Standard Situation, Measure
their Behavior, and look at the Differences
-
Formal
- Objective and Standardized Measure of a Sample of Behavior (Set of Test Scores)
- a Standard Situation to which Different ppl are exposed, where the Focus is on Individual Differences in Performance and Reactions
Operationalization
Definition
Kid
you have an Abstract concept you canNot Measure, so you look at those things you Can Measure that you think are Caused by that concept
Ex. Anxiety:
- physiological arousal, self-report of
tension and apprehension, etc.
Manifest (x) / Latent (L) Variables
Personal
- there's an Abstract Psych. Construct you canNot Directly measure, so you look at those things you Can Directly Measure that you think are a Manifestation of that Abstract Construct
- Indirectly Measuring an Abstract concept by
its supposed Directly Observable Manifestations
Formal
- Defining an Abstract concept by Directly Measuring the supposed Manifestations of that concept
- Defining a Latent Variable by the Manifest Variables that are supposedly caused by it
- Defining a Latent Variable by the Operations undertaken to Measure it
What is a Good Test?
-
Reliability
Definition
Formal
- if it produces Same Results given the Same Condition (hypothetically: same Situation and Individual)
- It is the characteristic of a Set of Test Scores that relates to the amount of Random Error from the Measurement Process that might be embedded in the scores
- a reliable test gives Predictable results Free from Random Errors
Personal
- how Accurately it Measures the Same Smth (Not necessarily what we Intend to measure) under same Conditions
- how Reproducible are some Results given similar Conditions
Kid
how much Similar Results we get in Similar Conditions (even
if maybe we are Measuring smth else than we want to)
Reliability
- 3 more items...
Validity
Definition
Formal
- The extent to which a Test Measures what it is Intended to Measure
- the degree to which Evidence and Theory Support the Interpretations of Test scores (entailed by the Intended uses of tests)
- it's a Property of the Interpretation of a Test
-
Personal
how Accurately a Test represents the Abstract Psych. Construct or the Differences we Intended to measure
Kid
if we are Measuring what we Wanted to Measure,
and Not Smth Else instead (other Variables or Errors)
-
-
-
-
T-scores
Decided by Researchers
the Scale is Transformed so that:
- Mean = 50
- Standard Deviation = 10
-
-