evaluations C2

• LIMITATIONS - field experiment = loss of control, more difficult to replicate, may have some ethical issues, less control of extraneous variables
• LIMITATIONS - lab experiment = Ps may be aware they are being studied, so demand characteristics are likely to occur, may have some ethical issues, lacks ecological validity.
• LIMTATIONS - quasi experiment = no direct manipulation of variables, cannot demonstrate causal relationship, lacks internal validity.
• LIMITATIONS - online experiment = too restrictive (multiple choice) so may not be accurate, may not take part as it is online so less response rate.
• STRENGTHS - lab experiment = high levels of control over variables, easy to replicate, cause + effect relationship can be established, extraneous variables are minimised
• STRENGTHS - field experiment = high ecological validity, cause + effect may be established
• STRENGTHS - quasi experiment = demand characteristics less likely, increased ecological validity , enables researchers to study real problems
• STRENGTHS - online experiment = they are replicable, easy to get more people/large sample, doesn't take long, anonymous - get more honest responses, low cost
• LIMITATIONS - privacy = there is no universal agreement on what a public place is, people may still object to being observed in public.
• LIMITATIONS - confidentiality = sometimes it may be obvious who has been involved in the study even with anonymity, sometimes possible to work out who the participants are based on certain information collected in the study e.g. location.
• LIMITATIONS - deception = decisons by ethics committee are subjective, deception means Ps cant give informed consent, even if Ps are debriefed it has alrady hppened.
• LIMITATIONS - right to withdraw:

  • from researchers point of view it may be necessary to keep Ps in the study until they hve completed the task + as a result may turn a blind eye to inclinations of withdrawal.
  • Ps may feel they cannot leave as they do not want to spoil the study.
• LIMITATIONS - informed consent:

  • having informed consent may invalidate the results of the study if too much information is given to the Ps - researchers often deceive Ps bout what it is they are actually researching - as they need to avoid Ps figuring out the aims.
  • Ps may agree to take part in study, but still not completely understand what they have let themselves in for.
• LIMITATIONS - protection from harm:

  • researchers often investigate important areas that may cause psychological harm to Ps.
  • rsearchers cannot always predict any potential harm that may occur until it happens - the damage is already done.
  • no more harm than they would 'normally' experience in everyday life = a sujective judgement of what the Ps would actually experience in everyday life.
• STRENGTHS - independant measures:

  • no order or practice effects
  • demand characteristics are less of a problem
• STRENGTHS - repeated measures:

  • extraneous participant variables are eliminated
  • ^ individual differences eliminated
• STRENGTHS - matched pairs:

  • no order effects
  • no practice effects
  • controls participant variables.
• WEAKNESSES - independant measures:

  • individual differences (participant variables) are an issue
  • potential for error due to individual differences between groups of Ps
• WEAKNESSES - repeated measures:

  • greater chnce of demand characteristics
  • there my be order effects (fatigue, boredom, practice effects)
• WEAKNESSES - matched pairs:

  • can be very time consuming + difficult to get the right pairs
  • costly
  • matches are never perfect
• CONTROL - how would you avoid order effects in REPEATED MEASURES design?

  Counterbalancing - to aviod order effects, the sample group can be splitin half so that one half does condition A first, and then the other half does condition B first = ABBA
• CONTROL - what is the best match for MATCHED PAIRS design?

  Best match would be indentical twins - the design already minimises P vriables but the more variables thhe pairs are matched on the better the match will be.
• how would you ensure FACE VALIDITY:

  • researcher can ask lay people (participants) what the study appears to be measuring
  • if they all agree on e.g. intelligence or IQ, then the study has face validity
• how would you ensure CONTENT VALIDITY:

  • a panel of experts may be asked to asses the measure for validity
• how would you ensure CONCURRENT VALIDITY:

  • test Ps with an already established test of known validity
  • if test has concurrent validity, there should be a close agreement between the scores on both measures
• how would you ensure PREDICTIVE VALIDITY:

  • by following up Ps to see if future performance is similar to performance tested
• how would you ensure CONSTRUCT VALIDITY:

  • must define what it is we are aiming to measure
  • make sure that all parts of the definition are being measured
• ways of assessing reliability (SIT) - SPLIT HALF

  • split half measures internal reliability
  • compare 2 halves of the same measure - to test to see if they have the similar score e.g. stress test, personality test
  • performance on 2 halves of test is compared
  • if test is assessing same thing in all its questions then there should be a high correlation in all the scores from both halves of the test
  • gives test split half consistency
• ways of assessing reliability (SIT) - INTER-RATER RELIABILITY

  • measures internal reliability
  • important that researchers all agree, so when results are collated they are reliable
  • if the rater's/researchers are assessing the same thing then there should be a high correlation in their ratings
• ways of assessing reliability (SIT) - TEST-RETEST

  • method of assessing external reliability
  • concerned with showing consistency over time
  • same test is carried out again, with same Ps, to see whether their results are the same over time.
  • if 2 tests achieve same scores/results on both occasions there will be a high correlation between the 2 sets of scores
• reliability issues

  1. there is only 1 researcher

  how to deal with:

  1. use more than one researcher
• reliability issues

  2. the research has not been replicated

  how to deal with:

  2. repeat study, on another occasion, using same participants
• reliability issues

  3. instructions are not given in the same way

  how to deal with:

  3. standardise instructions given to all participants so they are all the same
• reliability issues

  4. participants are not asked questions in the same way

  how to deal with:

  4. standardise questions + possible answers
• reliability issues

  5. participants do not have same experience in an experiment as procedures or not standardised

  how to deal with:

  5. standardise the procedures of experiments/research so all participants have the same experience
• reliability issues

  6. items in the questionnaire are not answered in the same way

  how to deal with:

  6. (cannot deal with it but can asses whether they have)
• STRENGTH + WEAKNESS of event sampling:

  • strength = useful if behaviour doesn't happen frequently so could be missed in time sampling
  • weakness = if too many observations happen at once, it may be difficult to record everything
• STRENGTH + WEAKNESS of time sampling:

  • strength = reduces number of observations being made
  • weakness = some behaviours may be missed as observations may not be counted when the behaviour is present. Not all behaviour is counted (something may have occurred between times)
• ADVANTAGES of observations:

  • more natural behaviour occurs if people are unaware of observation (covert), this means high ecological validity
  • findings are more valid in naturalistic + participant observations as researchers can see for themselves how participants behave - this is often different to what they say they do
• what methods produce quantitative data:

  • questionnaires (closed questions)
  • observations (tally behaviours)
  • frequency analysis (content analysis - counting)
  • measuring DV in experiments