7. Research methods
The experimental method
Information
The scientific method:
1. Observe
2. State expectations
3. Design a study
4. See if your expectations are correct
Experiment Key Words:
Aim: A statement of what the researchers are intending to do/find (research question
or intent- e.g. To investigate the effects of TV on homework performance)
Independent Variable: The two conditions (or the two levels of the independent
variable.) This allows for comparison.
Dependant Variable: What’s being measured (e.g. Number of words recalled after
10 minutes). These should be operationalised (specific) to make sure every participant
is doing the same thing and the experimented can be repeated.
Casual relationship/conclusions: A change in one variable causes a change in the
other.
Hypothesis: What you’re expecting to find. The expected relationship between the
independent and dependant variable.
Standardised Procedure: What the participants are doing. Making sure each
participant does exactly the same thing in each condition removed extraneous
variables as otherwise the results may change due to the procedure and not the
independent variable.
Ethical Key Words:
Informed Consent: Telling the participants what they’ll be expected to do and allow
them to pull out.
Age Restrictions: Never using anyone under the age of 16, or anyone who may be
described as ‘vulnerable.’
Debrief: Tell them of any deception that may have taken place and allow them to
withdraw their data. Always consult others on the script you’re going to be using
when it comes to the debrief and informed consent.
,Control of variables
Information
Control: The extent which a variable is regulated by a researcher.
Confounding Variables: The experimenter hasn’t tested what they intended to test, but
instead the influence of a different variable. Varies systematically with the independent
variable
Extraneous Variable: Doesn’t vary systematically with the independent variable but still
may change the results of the dependent variable. Nuisance variables.
Realism: Making sure the set-up of the study isn’t too artificial so that people will act as they
do usually.
Mundane Realism: How much the experiment mirrors the real world.
Internal Validity: What’s going on within the study. Confounding variables, extraneous
variables, mundane realism. Are they testing what they intended to test?
External Validity: How the findings can be generalised to the rest of the world. Ecological
validity, population validity, historical validity. The study materials shouldn’t really be
contrived, as mundane realism should be high.
Hypotheses, Pilot studies, confederates
Information
Directional Hypothesis: You state the direction of your results. (e.g., People with plentiful
sleep- over 8 hours per night for a period of one month- have higher test markers than people
with a lower sleep average.) Used when past research suggests that findings will yield a
specific result.
Non-Directional Hypothesis: States there’s a difference between the conditions but not what
the direction of the difference. (e.g., People with plentiful sleep- over 8 hours per night for a
period of one month- will have different test scores than people with a lower sleep average.)
Used when there’s no past research (theories or studies) or the research is contradictory.
Pilot Studies: A small scale trail run of a research design before doing the real thing. Allows
us to see what does and doesn’t work- participants may get boarded at question 50 or guess
the aim of the study by question 12. This allows a lot of money to be saved. The results don’t
matter.
Confederates: An induvial taking part in the experiment who is not a real participant but has
been instructed by the investigator
,Experimental Design
Information
Repeated Measures Design: All participants receive all levels of the independent variable
(they do both conditions, then we compare their results/the dependent variable).
Strength: Individual differences reduced, less time and money
Disadvantages: Order effect (boredom effect, practice effect), demand
characteristics.
Dealing with disadvantages: Use two different tests that are equivalent to
reduce practice effect. Counterbalancing (will discuss later) can also reduce order
effects. Deception can be used to stop demand characteristics.
Independent Group Design: Separate groups of participants are placed in separate
conditions, and we compare the results/dependent variable of the two groups.
Strengths: No order effects, higher external validity as more participants are used
Disadvantages: Individual differences (participant variables) are extraneous
variables, more people (time and money) for the same amount of data
Dealing with disadvantages: Randomly allocate so theoretically the participant
variables are distributed evenly.
Matched Pairs Design: Separate groups in separate conditions but they are matched by key
characteristic believed to effect performance on the dependent variable.
Strengths: No order effects, same materials can be used in all conditions, fewer
demand characteristics, fewer participant variables
Disadvantages: Time consuming and expensive to match the pairs up, impossible
to match every variable that’s relevant
Dealing with disadvantages: Restrict the number of variables to match on to
make it take less time, have a pilot study to find key variables that might be important
Counterbalancing: In the repeated measures design it is used to balance out the order
effects. There are two ways.
1. AB BA- In this one you split the group into two, and group 1 does condition A,
then B, whilst group 2 does B, then A.
2. ABBA- In this one you keep the group together, but they do the conditions twice.
Basically, if it were a memory test and the order effects was tiredness they would
do 4 trails:
Trial 1. Condition A in the morning.
Trail 2. Condition B in the afternoon.
Trail 3. Condition B in the morning.
, Trail 4. Condition A in the afternoon.
The order effects are balanced throughout the method so they’re not just effecting one
condition.