Study design - Chapter 4*

Correlational studies

• Observation of naturally occurring variation

• No manipulation

• Often less invasive

• Cannot establish causation

Experimental studies*

• Manipulation of variables

• Conducted in natural or

Independent variable*

• What is causing your behaviour

• A biotic or abiotic factor that is manipulated to test for an effect

Dependent variables*

• The effect

• How a specific behaviour affected by the independent variable

Confounding factors*

• One or more variables that affect both the independent and dependent variable

• Can introduce bias and make it difficult o determine id an observed effect really is caused by the independent variable

• We must

Cross-sectional studies*

• Measures many different individuals at a single point in time

• Pro: fast to run, no risk of loosing subjects due to dropout, death. etc

• Con: Cannot separate stable, trait-like patterns from a transient state

Longitudinal studies*

Carryoverr effect?*

Between-subject design*

Each animal receives only one treatment

Different animals are compared to each other

Pro: There is no risk of carry-over between conditions since no animal receives

Within-subject design (repeated measures)*

Natural experiments*

Strenght:

• Real-world scale, real world scenario

• Often large effect sizes

Randomisation

• Prevents systematic bias

• Randomly assign animals to treatment groups before the experiment

• For repeated testing: randomise the order of treatments

Blinding

• Ideally the scorer of the behaviour should not know which treatment each animal received

• Unconscious expectations biases measurements

How do you choose your sample size?*

Consider:

• Statistical power

• Ethical considerations

• Reality

Calculate your sample size before your experiment

What is power? how do you calculate it?

What is power affected by?*

• Sample size

Pseudoreplication*

True replication*

*

*

*

*

*

*