PSYU3330 Measurement & Research

Non-Scientific Methods (5)

TIARE: Tenacity, Intuition, Authority, Rationalism, Empiricism

What would high precision look like on a bell curve?

High precision → Tall bell curve

Low precision → Wider bell curve

Correlational Design

Relationships (strength & direction), describing events, what’s happening?

What would low accuracy look like on a bell curve?

Low accuracy → Darts spread out far from the bull's eye, indicating a bell curve centered away from the target. Same degree of precision (tall, skinny) but pushed to the side.

Experimental Design

Cause and effect, manipulation + control, explaining events, why’s it happening?

What kind of error does a random error lead to?

Low precision

What kind of error does a systematic error lead to?

Low accuracy

Which values are considered “good reliability”?

Values above .8

Correlation values range from

0 (completely unreliable) to 1 (perfectly reliable)

What does higher rxx indicate?

Lower measurement error

Independent Variable

Manipulated/changed to see the effect it has on DV

Dependent Variable

Measured

Describe the two experimental designs

Between-subjects: diff groups assigned to 1 condition each

Within-subjects: same participants do all conditions/levels of IV

What was the main issue with the Nuremberg Code (1947)?

Not legally binding, and leaves decisions up to the researcher

Animal Research Ethics (3)

1. Replacement: use alternatives to animals where possible

2. Reduction: use minimum number of animals necessary to get meaningful results

3. Refinement: minimise suffering as much as possible

Describe the types of interviews (3)

1. Structured: highly standardised, often w close-ended qs (market research surveys)

2. Semi-structured: flexible, open-ended qs guided by an interview schedule, allowing for exploration of participant responses

3. Unstructured: very open, participant-led, no specific qs in mind, aiming to immerse in the participant’s experience

Pillars of Research (4)

POEM

1. Purpose

2. Ontology (nature of reality)

3. Epistemology (position of researcher)

4. Methodology

Combining qualitative and quantitative methods increases [reliability/validity] of findings

reliability

Describe the 4 mixed methods

1. qual-QUANT: Qual first (exploratory) → quant (to test findings)

2. quant-QUAL: Quant first (to identify participants) → qual (for depth)

3. QUANT-qual: Primary quant → secondary qual follow-up

4. QUAL-quant: Primary qual → secondary quant follow-up

Evaluation Principles (5)

FURCT

1. Fit for purpose

2. Useful

3. Robust, ethical and culturally appropriate

4. Credible

5. Transparent where appropriate

List the process of program evaluation

1. Establish an Expert Advisory Group

2. Inputs (resources, staff, funds)

3. Outputs (participants, activities, workshops)

4. Outcomes & Impact (short, medium, long-term changes)

5. Project Manage

6. Disseminate Findings

Critical Appraisal Skills Program (CASP)

Assessment of Quality

For internal validity

Random patient assignment, generalisability, benefits vs harms/costs

Cochrane’s Risk of Bias Tool

Assessment of Quality

To assess potential biases in studies (specifically randomised controlled trials)

Consolidated Standards of Reporting Trials Checklist (CONSORT)

Article Reporting Guidelines

NHMRC Evidence Hierarchy

Assessment of strength of evidence for specific interventions

Ranks types of evidence from highest to lowest quality

• Highest: systematic review of randomised controlled trials

• Lower: quasi-experimental studies, comparative studies, case series, expert opinion, anecdotal evidence

Hit

Signal present, and observer says “yes” = correctly detects signal

Miss

Signal present, but observer says said “no” = fails to detect signal

False Alarm

No signal (only noise), but observer says “yes” = incorrectly detects signal

Correct Rejection

No signal, observer says “no” = correctly rejects idea of signal

What does a larger d-prime (d’) indicate?

Better sensitivity

What does a Liberal Criterion mean?

Lower threshold (more “yes” responses, leading to more hits but also more false alarms)

What does a Conservative Criterion mean?

Higher threshold (fewer “yes” responses, leading to fewer false alarms but also more misses)

What can Signal Detection Theory be applied to?

• Lie detection

• Radiologists inspecting MRI scans for cancerous tumours (distinguishing tumour from dense tissue or imaging artifacts)

• Digital electronics (distinguishing binary 0s and 1s amidst electrical noise