research methods

• lab study

  • controlled conditions
  • IV manipulated
  • DV measured
  • low ecological validity
  • controlled extraneous variables
  • high replicability
• field study

  • natural conditions
  • IV manipulated
  • DV measured
  • high ecological validity
  • uncontrolled extraneous variables
  • low replicability
• natural study

  • natural conditions
  • IV naturally occurring
  • DV measured
  • high ecological validity
  • uncontrolled extraneous variables
  • low replicability
• quasi study

  • controlled/ natural conditions
  • IV difference in people
  • DV measured
  • low/high ecological validity
  • un/controlled extraneous variables
  • high/low replicability
  • dependent on whether experiment carried out in lab or under natural conditions
• observations

  provide psychologists with way of seeing people's behaviour without self-report methods
• covert observation

  • undisclosed - participant unaware
  • S: investigator effects + demand characteristics less likely
  • L: less ethical - lack of informed consent
• overt observation

  • participant unaware
  • S: more ethical - informed consent obtained
  • L: investigator effects + demand characteristics more likely
• participant observation

  • researcher takes part - close proximity
  • S: in-depth data gathered due to close proximity
  • L: investigator effects + demand characteristics more likely
• non-participant observation

  • researcher doesn't take part - not close proximity
  • S: investigator effects + demand characteristics less likely
  • L: researcher might miss behaviours due to lack of proximity
• naturalistic observation

  • takes part in unaltered setting
  • S: higher level of ecological validity
  • L: issued with replication + ascertaining reliability
• controlled observation

  • takes place in artificial environment
  • S: can be replicated to check reliability
  • L: lower levels of ecological validity
• structured observation

  • behaviour coded using behavioural categories
  • S: higher levels inter-observer reliability
  • L: less rich data can lack internal validity as researchers miss important behaviours
• unstructured observation

  • every instance of behaviour recorded + described
  • S: rich data leads to greater internal validity
  • L: lower levels of inter-observer reliability
• behavioural categories

  break target behaviour into set categories

  should include all potential behaviours

  • S: make structured + objective data
  • L: difficult to produce data - no possibility for further interpretation, all possible forms of behaviour, categories exclusive
• time sampling

  overview: observer records behaviour at perscribed intervals

  • S: better use of time since fewer observations made
  • L: not every behaviour of relevance to investigation counted if it occurs in-between time frames allocaed
• event sampling:

  records number of times target behaviour occurs

  • S: every behaviour of interest counted from beginning to end of observation
  • L: possibility some behaviours missed if there's too much happening at the same time, meaning some not coded
• open questions

  allows participants to answer as they wish

  qualitative data

  • S: less chance of researcher bias, detailed answers
  • L: participants may answer in socially desirable way, making answers less valid
• closed questions

  participants restricted to set responses - checklists, likert scaled, ranking scaled

  quantitative data

  • S: easy to analyse quantitative data, discover trends + replicate
  • L: predetermined list of questions limits responses ability to explore interesting answers - response bias
• questionnaires

  form of self-report in which ps provide information relating to their thoughts, feelings + behaviours, can have open, closed or both typed of questions

  • keep terminology simple
  • short as possible
  • be sensitive
  • no leading questions
  • no assumptions or sweeping statements
  • pilot + modify questions
• interviews

  form of self-report predominantly takes place face-to-face, can be structured, unstructured, semi-structured, responses usually recorded, can be group interviews in clinical setting
• structured interview

  pre-determined questions, fixed order

  • S: easier to analyse, quantitative data, discover trends, replicability
  • L: increased risk of investigator effects
• unstructured interview

  conversation like, interviewer prompts interviewee to elaborate

  • S: less chance of demand characteristics + investigator effects
  • L: difficult to analyse qualitative data, time consuming + expensive
• semi-structured interview

  many interviews combination of set of questions + others

  • S: rich + interesting qualitative data, less chance of social desirability bias
  • L: higher chance of demand characteristics in comparison to unstructured
• correlations

  illustrated strength + direction of association between 2+ co-variables - no cause + effect

  + correlation (coefficient +1): one increases, other does

  -correlation (-1): one increases other decreases

  zero correlation (0): no relationship

  correlation coefficient: measures strength + nature of relationships between co-variables

  can be presented on scattergrams
• correlations evalutations

  • S: can be used when lab experiment unethical as variables not manipulated, only correlate measure strength of relationship between variables - allowing further research
  • L: not possible to establish cause + effect
  • L: can only identify linear relationship - not curvlinear
  • L: misinterpretation/ misuse - media presents relationships as casual when they aren't
• case studies

  detailed analysis of individual, establishment or event

  doesn't refer to way research conducted

  often used to study rare behaviour

  can be longitudinal

  qualitative data: interview, observation, questionnaire

  • S: in depth detailed - insight into behaviour
  • L: generalisability - small scale
  • L: researcher's subjectivity interferes with validity of findings/ conclusions
• aim

  lays out question researcher attempting to answer

  always starts with 'to examine the effect of'
• independent variable

  variable researcher manipulated + which is assumed to have direct effect on DV
• dependent variable

  variable research measured + affected by IV
• hypothesis

  clear + precise prediction about difference/ relationship between variables

  should always contain operationalised IV + DV
• null hypothesis

  statement which predicts no effect of IV on DV
• alternate hypothesis

  effect of IV will be on DV - can be directional or non-directional
• directional hypothesis

  one tailed

  predicts nature of difference between variables based on past research

  key words: higher, lower, more, less, increase, decrease, positive, negative
• non-directional hypoethsis

  two-tailed

  predicts difference exists but not direction of difference

  previous research inconclusive + IV-DV relationship can't be predicted confidently

  key words: difference, relationship
• random sample

  every member of target population identified (equal chances) + sample randomly selected (names in hat, computer software)

  • S: likely to be representative so generalised to target population
  • L: difficult + time consuming to get full details of target population ppl selected may be unwilling to take part - becomes volunteer sample
• systematic sample

  every nth person selected from register, handbook etc

  • S: not subject to researcher bias as numerical value assigned to each person on given list + selected according to nth position
  • L: not truly unbaised as every nth selected may have similar trait in common therefore not representative
• stratified sample

  subgroups within population identified, participants obtained in proportion to occurrence within population

  • S: likely to be representative as each subsection of target population proportionally represented - generalisability
  • L: can't represent every difference
  • L: difficult + time consuming
  • L: people selected may be unwilling to take part - becomes volunteer sample
• opportunity sample

  selecting participants available + willing to take part

  • S: quicker + easier to obtain in comparison
  • L: high chance sample biased - availables not representative of target population
• volunteer sample

  participants self-select by responding to advert - offer to partake

  • S:quicker + easier to obtain in comparison
  • L: particular type os person likely to take part - biased + unrepresentative
• pilot studies

  small scale prototypes of study carried out in advance of full research to find problems in:

  • experimental design - enough time?
  • instructions for participants - instructions clear?
  • measuring instruments - categories + questions checked

  identify issues - adapt - save time + money - ensured investigation runs smoothly