Variables, designs and hypotheses
Scientific methods (3)
Experiments:
- Vary some independent variable while holding everything else constant
- Measure changes in some dependent variable.
Changes in DV must have been caused by changes in IV
- Can infer causality from this
Quasi – experiment
- Similar to experiment except IV cannot be manipulated (e.g. age of someone, or a
mental disorder, so instead have to allocate these people within their similar group)
- Potential problems with confounding variables (the people with a mental disorder
may be of different ages and so hard to distinguish which group to put them in)
Correlational design
- No manipulation
- Measure 2 (or more) variables and determine the extent to which they are related to
each other (or co-related)
- Cannot infer causality (do not know what caused what)
Variables
Independent Variable:
- Can have one or more IVs or factors
- Each IV can have 2 or more levels, e.g:
- Test ppts reading speed after a meal, before a meal and 2 hours later
- 1 IV (time of day) with 3 levels (before, after and 2 hours later)
Dependent variable:
- An experiment can have 1 or more dependent variables, i.e. what we actually
measure
- E.g. ‘does having A level maths have any effect on the ability to understand an
undergraduate psychology course?’
- ^^potential DVs: 1st year exam marks, 3rd year project marks, final degree outcome.
DVs have several types or measurement scales:
- Nominal scales (categorical):
Numbers refer to different classes
Classes not necessarily related to each other
- Ordinal scales (ranking):
Numbers indicate a relative position (rank) in a list
Rank is meaningful, items not necessarily at equal intervals
- Interval scales:
Scientific methods (3)
Experiments:
- Vary some independent variable while holding everything else constant
- Measure changes in some dependent variable.
Changes in DV must have been caused by changes in IV
- Can infer causality from this
Quasi – experiment
- Similar to experiment except IV cannot be manipulated (e.g. age of someone, or a
mental disorder, so instead have to allocate these people within their similar group)
- Potential problems with confounding variables (the people with a mental disorder
may be of different ages and so hard to distinguish which group to put them in)
Correlational design
- No manipulation
- Measure 2 (or more) variables and determine the extent to which they are related to
each other (or co-related)
- Cannot infer causality (do not know what caused what)
Variables
Independent Variable:
- Can have one or more IVs or factors
- Each IV can have 2 or more levels, e.g:
- Test ppts reading speed after a meal, before a meal and 2 hours later
- 1 IV (time of day) with 3 levels (before, after and 2 hours later)
Dependent variable:
- An experiment can have 1 or more dependent variables, i.e. what we actually
measure
- E.g. ‘does having A level maths have any effect on the ability to understand an
undergraduate psychology course?’
- ^^potential DVs: 1st year exam marks, 3rd year project marks, final degree outcome.
DVs have several types or measurement scales:
- Nominal scales (categorical):
Numbers refer to different classes
Classes not necessarily related to each other
- Ordinal scales (ranking):
Numbers indicate a relative position (rank) in a list
Rank is meaningful, items not necessarily at equal intervals
- Interval scales:
