Experimental designs and controlling extraneous variables Flashcards
What is an experimental design? And the three types of experimental designs ?
The experimental design of a study is how the participant are assigned to different conditions. There are 3 main types of experimental deigns.
Independent groups
Repeated measures
Matched pairs
Independent groups
In the independent groups design different participants are used in each of the conditions, so each group of participants are independent from each other, people are randomly assigned to groups to prevent participant variables
Independent groups evaluation
+) order effects will not occur. This is when the sequence in which the participants take part in conditions effects behaviour, such as people getting tired or better with practice.
+) the chance of demand characteristics is lowered as they each only do one task so have less information with which to guess the purpose of the study.
+) birth groups can be tested at the same time with enough researchers which saves time
- ) more participants needed
- ) different results between the groups could be due to participant variables (PV) which makes it harder to establish cause and effect.
Repeated measures
In repeated measures each participant is tested in all conditions of the experiment
Repeated measures evaluation
+) same people are measured in each condition so no participant variables between the conditions
+) half as many participants needed then the independent groups design
- ) order effects can effect the results, this can be mitigated by counterbalancing. This is when half the participants do condition A first then B while the other hand dies the opposite, it doesn’t get rid of order effects but controls the impacts of effects such as fatigue, practice, boredom etc and makes it equal in both conditions.
- ) demand characteristics more likely as participants have more information
- ) this experiment design takes more time especially if there is time gap between conditions.
Matched pairs
In the matched pairs design the different participants are used in all of the conditions like in the independent groups design but the participants in the twin groups are matched in characteristics important for the study such as age, gender, education etc. Identical twins are often used.
Matched pairs evaluation
+) there is less risk of rider effects
+) less risk of demand characteristics
+) participants variables are unlikely as the groups are closely matched
- ) twice as many participants are needed then then repeated measures design
- ) the matching processes is difficult and expensive ,also even two closely related participants will still have many differences.
- ) the matching processes is very time consuming
Controlling extraneous variables
In order for a study to have validity extraneous variables (EV’s) need to be controlled for before the coke confounding variables (variables the effect the results of an experiment)
What are the 4 types of extraneous variables?
Participant variables- there are characteristics of the participants why’d can effect the DV such as age or intelligence.
Environmental variables- these are factors in the environment that could effect the DV such as time of day, noise, temperature etc.
Investigator aims- this is when the person collecting the information may inadvertently effect the results of the research.
Demand characteristics- these are features of a research study that allows participants to guess a study’s purpose which can lead to them changing their behaviour.
How to control for participant variables?
Choosing a certain type of experimental design may help, such as matched pairs or related measures. However repeated measures can lead to order effects so counterbalancing may be needed.
Random allocation of participants can also ensure groups are nit biased such as in independent groups however you can’t do this in a quasi experiment.
How to control for environmental variables
Environmental variables can be controlled by standardisation, which is when all the conditions, instructions and materials are the same for all participants.
How to control for investigator effects?
Observer bias is a type of investigator effect which occurs when the person collecting the data knows the aim of the research. This can be solved by the double blind technique so neither the participants or investigator know the aims of the study.
Also certain physical characteristics of the investigator such as age, gender and ethnicity can effect results so standardised scripts should be written so the investigator acts the same way to everyone. And some study’s should have the investigator be the same gender, age, ethnicity etc as the participants.
How to control for demand characteristics
Acting unnaturally out of nervousness or due social desirability bias are types of demand characteristics so observer effects and interviewer effects are types of demand characteristics and need to be controlled.
Demand characteristics can be controlled by the single blind technique in which participants don’t know the hypothesis or what condition they are in this is more difficult using a repeated means red design.
