experimental methods

• independent variable

  • the one that is changed
  • experimenter directly controls
  • to observe effects on dependent variable
  • used to establish cause and effect
• dependent variable

  • the one that is measured
  • how effect of change of IV is measured
• co-variables

  • measuring 2 separate variables and comparing them to eachother
• operationalisation of variables

  • stating a method of measuring the DV
  • eg. DV is time taken to do the task in seconds
• extraneous variables

  • things you keep the same to ensure a fair test
• cofounding variables

  • at end of experiment any extraneous variables that hasnt been controlled become cofounding variables
• key features of experiments

  • attempt to control all extraneous variables (fair testing)
  • IV is manipulated by experimenter
  • Ps are randomly assigned to groups
• lab experiments

  • all 3 key features
  • controlled setting
  • film clips of accidents - loftus and palmer, in critical Q Ps were asked to describe how fast the cars were travelling and each group of Ps were given a different verb varying in intensity. (hit, smashed, collided etc) verb smashed had a higher mean speed than contacted
  • IV - verb in critical Q
  • DV - mean estimated speed
  • EV - film clips, length of clips
  • strength - controlled EV establishes cause and effect, good internal validity
  • weakness - artificial, not same as real life (emotion, anxiety), lacks ecological validity
• field experiment

  • all 3 key features
  • in real life setting
  • power of uniform experiment - bickman, males dresses as milkman, security guard, ordinary clothes and asked public to follow instructions. guard was obeyed on most occasions, milkman too, and pedestrian the least. suggests people are more likely to obey someone wearing uniform
  • IV - uniform worn
  • DV - obedience rate
  • EV - weather, time of day
  • strength - more real life application, high ecological validity
  • weakness - not as controlled or as ecologically valid as lab experiment
• natural experiment

  • naturally occuring IV - not manipulated by experimenter
  • use if cannot ethically manipulate
  • eg. where you grew up
  • romanian orphans study - investigating to the extent good care can make up for poor early experiences. physical, cognitive and emotional development assessed at intervals throughout life. the sooner adopted, the higher the IQ. many showed signs of delayed intellectual development and undernourishment when coming to UK
  • IV - IQ
  • DV - age adopted
  • EV - who adopted, childs natural IQ
  • strength - good ecological validity, ethics
  • weakness - no control, lacks ecological validity, cant establish cause and effect
• quasi experiment

  • like natural experiment but the IV is feature of the P
  • eg. age, gender
  • digit span test - jacobs acknowledging memory gradually improves with age, conducted digit span test on 443 female students from same school, varying ages. increased number of items until Ps could no longer recall sequences
  • IV - age
  • DV - STM
  • EV - gender, test, school
  • strength - allows comparison between different types of people
  • weakness - Ps are aware theyre being studied - demand characteristics, reduced internal validity
• experimental design

  how participants are arranged in experiments
• experimental design - repeated measures

  • what is it? - using same subjects in each condition
  • strengths - no P differences between groups (fairer test), less Ps needed in total (less cost, easier)
  • weaknesses - more likely to guess what study is about (may alter their behaviour, demand characteristics, screw you effect), order effects (may do better 2nd time due to practice / worse due to boredom)
  • how to combat weaknesses - need to use counter balancing (ABBA design) to cancel out order / practice effects
• experimental design - independent measures

  • what is it? - using different subjects in each condition
  • strengths - Ps are less likely to guess aims (no demand characteristics or screw you effect), no order effects (as opposed to repeated measures)
  • weaknesses - P variables (differences between groups, test might not be as fair), more Ps needed (more cost, harder)
  • how to combat weaknesses - use random allocation for Ps and conditions to ensure a good spread
• experimental design - matched pairs

  • what is it? - using different but similar subjects in each condition
  • strengths - all the strengths
  • weaknesses - really difficult to do well (time consuming too), difficult to measure ability of Ps before hand (like driving ability, alcohol dependence)
  • best option
• pilot studies

  • small scale pilot study can be run first to help forsee any problems
  • should establish whether the design works, if Ps understand what to do, if something important has been missed out
  • researchers can practice following procedures and tackle problems before main study, potentially saving time and money
  • can be used to devise behavioural catagories for observations
• investigator effects

  • how experimenter behaves towards P (researcher bias)
  • are they friendly? - will effect Ps motivation and behaviour
  • do they let on purpose of study to Ps? - could effect results
  • how to fix - use double blind procedure (neither researcher nor P knows the aim, employ 3rd party people to run study), standardise everything (instructions, timings), pilot study
• demand characteristics

  • when Ps work out the aim of the study and change their behavior accordingly (or screw you effect - doing opposite)
  • more likely in repeated measures lab experiments
  • deception prevents this
  • how to fix - standardise everything (instructions, timings), pilot study