Experimental methods

• Hypothesis: a testable statement about the relationship between two variables: the independent variable and the dependant variable

  Alternative Hypothesis - the hypothesis a researcher tests by conducting a study and collecting data attempting to show how the null hypothesis is not supported

  Null Hypothesis- states no variables have an effect on other variables.

  This is the default hypothesis until there is evidence too reject it
• Laboratory experiment


  An experiment carried out in highly controlled environments like a laboratory. The setting is often artificial, or not where people would usually produce this behaviour.
• Field experiment


  A setting in which the behaviour would normally occur. The researcher might have confederates (people working for them) who follow a scripted behaviour.
• Natural experiment


  One in which the independent variable is not under the control of the researcher. The conditions of the experiment are already fixed, such as studies looking at gender or age differences.
• Types of experimental designs

  • Independent groups
  • Repeated measures
  • Matched pairs
• Independent groups design


  • People available to take part are divided into two groups
  • One group takes part in one condition, other group takes part in remaining condition
• Repeated measures design


  • One group of participants takes part in both conditions of the experiment
• Matched pairs design


  • People available are tested before taking part to match into pairs of similar qualities
  • One member of each pair takes part in one condition, the other takes part in the other condition
• Even though the people in the two groups are different, the researcher treats the data as if each pair of scores came from one person
• Identical twins are often considered perfect matched experimental designs to pairs in psychology
• Independent groups


  • No order effects because people only take part in one condition
  • Often, the same material can be used for the task in both conditions
  • Participants cannot work out the aim of the study because they only take part in one condition
  • Different people (participant variables) in the two conditions so that may be why the results are different
  • Need more people for the study. To get 10 in each group you need 20
• Repeated measures


  • The people in both conditions are the same, so there are no participant variables
  • Only need 10 people to get 20 results because each person produces two 'scores
  • There are order effects as people have to do two tasks
  • May need two tasks (they cannot learn the same list twice)
  • Participants may work out the aim of the study because they take part in both conditions
• Matched Pairs


  • Participant variables are reduced
  • No order effects
  • Often the same material can be used for the task in both conditions
  • Matching is difficult, time-consuming and not always successful
  • Some participant variables are still present
• Extraneous variable (EV)


  A variable that is not the IV but might affect the DV if it is not controlled
• Standardised procedures

  • A set sequence applies to all the participants when necessary
  • Makes the experiment unbiased
  • Researcher tries to control all the variables and events so that the results of the experiment can be related only to the IV
• The information that is said or written for participants might affect the way they participate in the study and therefore their scores or performances
• That could be an extraneous variable
• Usual practice for participant information

  1. Write as much of this information as possible

  2. Ensure that each participant receives the same information
• Briefing

  • What is said/written to encourage a person to agree to participate
  • Contains ethical information about consent, anonymity, the right to withdraw, and so on
• Standardised instructions


  Clear instructions about exactly what the participant will have to do in the experiment
• Debriefing


  • Explains the study in detail so that each participant is absolutely sure of the aim of the study, that is, why they were doing what they were doing in the condition(s)
  • Ethical issues are addressed again, especially the opportunity for the participant to withdraw their data if they feel unhappy with their performance
• Randomisation


  Using chance to provide an order for a procedure
• Randomisation techniques


  • Random allocation
  • Counter balancing
• Random allocation


  1. Participants are put into the conditions of the experiment using chance

  2. Researcher should not have any influence over which participants go into which condition

  3. Prevents bias in the results
• Random allocation procedure


  • 20 people available, 10 pieces of paper with 'A' and 10 with 'B' put in a bag, each participant takes a piece of paper which determines their condition
• Random allocation in matched pairs design

  1. Letters for each pair (A and a) put in a bag, one participant selects a letter which determines their condition

  2. Repeated for each pair
• With repeated measures design, there is no random allocation to conditions as each participant completes both conditions
• Counterbalancing


  • Procedure used to reduce order effects in repeated measures design
  • Half of participants complete Condition 1 then Condition 2, other half complete Condition 2 then Condition 1
• Counterbalancing does not eliminate order effects but shares them equally between the two conditions
• Strengths and weaknesses of experimental methods


  • Laboratory
  • Field
  • Natural
• Laboratory experiments


  • Lots of control of EVs means it is easier to establish cause and effect
  • The high levels of control and standardisation mean the experiment can be replicated and the results confirmed
  • There is high objectivity because opportunities for bias are reduced in the standardised procedures
  • The settings can be artificial, with unrealistic tasks, so there is a lack of ecological validity
  • Participants are aware they are in a study so their behaviour can be affected, and the results may be misleading or lack validity
• Field experiments


  • The setting is natural with realistic tasks, so there is high ecological validity
  • Replication is possible
  • Participants are unaware they are in a study so their behaviour will not be affected and the results will be valid
  • Reduced control of EVs means it is harder to establish cause and effect, so the researcher is less sure that the IV affected the DV
• Natural experiments


  • When the IV is naturally occurring, such as gender or age, this is the only type of experiment that can be conducted
  • The high levels of control (of all except the IV) and standardisation mean the experiment can be replicated and the results confirmed
  • There is high objectivity because opportunities for bias are reduced in the standardised procedures
  • There is no random allocation of participants to conditions of the experiment, as the IV is not under the control of the researcher
  • The settings can be artificial, with unrealistic tasks, so there is a lack of ecological validity
  • Participants are aware they are in a study so their behaviour can be affected, and the results may be misleading or lack validity