Lecture 2: Introduction to Research (PSYC2009)

Why do we need statistical methods?

1. To count, chart, summarise, and otherwise describe what we’ve measured.

2. To understand and deal with risk.

3. To distinguish real patterns from random ones.

4. To enable us to make inferences from a sample of people.

What are the ideals of the scientific approach?

1. Universalism

2. Organised scepticism

3. Communalism

4. Transparency

5. Disinterestedness

6. Honesty

Universalism

Evidence-based claims are judged on the merit of that evidence; are based on the evidence itself (and nothing else).

Organised scepticism

All evidence-based findings are provisional/temporary and are bound to change.

Communalism (transparency)

All scientific findings (and processes) should be made transparent.

Disinterestedness

Researchers should approach their research question (and receive research questions) impartially; they should be emotionally detached from their work.This means avoiding personal biases or conflicts of interest, ensuring that findings are based solely on the evidence.

Honesty

Don’t lie, deceive, or conceal.

Continuum of scientific misconduct

Aspects of ethics in research

• Confidentiality, anonymity, and privacy.

• Informed consent.

• Involuntary participation.

• Deception and debriefing.

• Mental and physical stress and discomfort.

• Choosing where, and what to publish.

Confidentiality

Not publishing data from participants/clients.

Anonymity

Keeping participant’s identity protected in publications.

Right to privacy

People’s expectations that personal lives will not be impacted by research participation.

Why are some people ‘anti-deception’?

When you deceive participants one time, you ruin them as participants. Every study after that, they’ll likely expect to be deceived again and question what’s actually happening.

What’s a purpose of deception?

To allow for true results. If you tell participants what you’re studying, their behaviour likely changes.

What must you do to minimise/prevent mental and physical stress and discomfort?

• Debrief

• Follow-up

• Provide contact information

Bottom drawer filing

The fact that typically only fancy findings get published, and boring stuff where nothing happens don’t.

Predatory journals

Journals you have to pay lot’s of money to publish

Why is it considered worse to publish in a predatory journal than to not have published at all?

It involves no peer review process, and as long as you pay, regardless of the paper’s content, it gets published. People commonly perceive such paper’s to involve poor judgement and poor science.

What are four types of research?

1. Experimental

2. Quasi-experimental

3. Non-experimental (survey/correlational)

4. Case study

Definition of Experimental Research

A systematic attempt to manipulate theoretically relevant variables and to examine the effect of these manipulations on the outcome variable. (Haslam & McGarty)

Purpose of experimental research

Cause and effect

What does experimental research involve?

Active intervention and the manipulation of variables. This allows for one to make casual attributions/explanations. Also involves experimental and control groups.

Experimental group

The participants who are subjected to the treatment.

Control group

The participants who are not subjected to the treatment.

Independent variables (IV)

The variable you vary.

Dependent variable (DV)

The variable you’re measuring (i.e., the variable you think will change due to the manipulation of the IV).

Internal validity

The amount of control that you have over your experiments, and the degree to which you are sure that your X is causing the change in Y.

External validity

How true something is in the real world. Is always balanced with internal validity.

Confounds

Variables that look like your IV, but isn’t your IV.

Between-subjects

Different people exposed to different levels of the IV

Within-subjects

Levels of the IV differ within the same participants.

Examples of validity threats

Practice and fatigue effects

Advantages of experimental research

• Causal inference

  • Control

Disadvantages of experimental research

• Practicality

• Reactivity effects

  • Validity (ecological and internal)

Quasi-Experimental research

• Assignments to groups on the basis of pre-existing differences on the independent variable.

• No manipulation

• No random assignment

Disadvantages to quasi-experimental research

• Hard to infer causality

• Influence of extraneous variables
  

Advantages of quasi-experimental research

• Good for studying variables that can’t be manipulated

• Good ecological validity

Non-experimental research

“Survey” or “Correlational”

Advantages of non-experimental research

• Ease of admin (faster, cheaper, more convenient for participants)

  • Many variables

Disadvantages of non-experimental research

• Generalisability

• Causal inferences (correlation is not causation)

When is case study research used?

When the phenomenon under study is rare.

Advantages of case study research

Can attain an in-depth exploration of theoretical ideas

Disadvantages of case study research

Conclusive statements are hard to make.