Day 7 – Experimental Research Methods
There are two main sources of data. One is from observation of naturally occurring phenomena (correlational or cross-sectional), the Second is by manipulating a variable and observing the effect it has on our variable of interest (experimental).
Experimental research methods
There are 2 ways of carrying out experiments:
1. You can allocate different groups to single conditions (between-groups also called independent measures design). Here you manipulate the IV using different groups.
2. You can allocate different groups to all conditions (within-groups or repeated measures design). Here you manipulate the IV using the same groups.
Our choice of experimental method determines the statistical test used to analyse the data.
In experimental research methods, small differences caused by unknown factors (individual differences, motivation, etc) are called unsystematic variation. Small differences caused by controlled/experimental factors are called systematic variation.
The role of statistics is to discover how much variation exists in the conditions and calculate how much of is systematic and how much of it is unsystematic.
Other things being equal, experiments using repeated measures are more likely to reduce the interference of unsystematic variation on the data.
It is useful because it eliminates most sources of systematic variation thereby, leaving only the experimental variable as the only causal factor of variation.
Randomising repeated measures design:
The sources of unwanted systematic variation are mainly
a) Practice effects: the participants performance vary because they are familiar with the test.
b) Boredom effects: the participants performance vary because they are bored from completing the previous condition.
Randomisation is used to counterbalance the order in which the participants go through the conditions.
Randomising independent measures design:
The sources of unwanted variation are individual differences. The way to counterbalance these differences is by making sure that they are evenly spread across the conditions. We do this by randomly allocating participants to conditions.