research methods

• self report 


  faces social desirability bias
• closed questions 


  limited responses

  +easy to analyse

  -restricted response
• open questions 


  own words (qualitative)

  +detailed

  -hard to analyse
• qualitative data 


  words

  +detail

  -hard to analyse
• quantitative data


  numbers

  +compare in graphs

  -less meaning
• primary data


  directly from source

  +tailored to study

  -time and expensive
• secondary data 


  not direct from source not direct from source

  +cheap

  -poor quality
• meta analysis


  data from numerous studies

  +greater validity

  -publication bias
• aim


  intent of investigation
• operationalised


  clearly defined and measurable
• hypothesis


  testable statement
• null hypothesis


  predicts no relationship
• directional hypothesis


  previous research, stated direction
• non directional hypothesis 


  no past research, no direction specified
• mean


  average

  +includes all scores

  -distorted by extreme values
• median


  all in ascending order, middle value

  +unaffected by extreme scores

  -extreme values may be important
• mode


  most common value

  +relevant to categorical data

  -overly simple
• naturalistic observation


  target behaviour normally occurs

  +high external validity

  -low control
• controlled observation


  control over variables

  +replication and standardisation

  -low external validity
• covert observation


  unaware being studied

  +fewer demand characteristics

  -ethical issues
• overt observations


  aware being studied

  +more ethical

  -demand characteristics
• participant observations


  researcher part of group

  +greater insight

  -loss of objectivity
• non-participant observations 


  separate form group

  +more objective

  -loss of insight
• behavioural categories (operationalisation)


  target behaviour broken up into observable categories
• event sampling


  every time it occurs

  +record infrequent behaviour

  -oversimplifies complex behaviour
• time sampling


  regular intervals

  +reduces observations

  -miss things outside time frame
• range


  difference between highest and lowest value

  +easy

  -no account of distribution
• standard deviation


  dispersion if values around the mean, larger SD means more spread out

  +more precise

  -distorted by extreme values
• tables


  raw scores in columns and rows
• scattergrams


  continuous data, correlation, data pairs
• bar charts


  categories, bar height represents frequency
• histogram


  data is continuous, no space between bars
• normal distribution 


  bell shaped, mean, median, mode all together
• negative skew


  modal scores higher, tail to left
• positive skew 


  modal scores lower, tail to right
• pilot studies 


  trial run - small scale test of procedure and techniques before doing full scale study

  aims - find what doesn't work (timings, stimulus, standardised instructions)
• lab experiments


  controlled, manipulate IV

  +replicate, EV controlled

  -low generalisability/external validity
• field experiments 


  natural setting, manipulate IV

  +ecological validity

  -EVs harder to control
• natural experiments 


  DV may be natural, IV varies

  +ecological validity

  -replicability issues
• quasi experiments


  IV pre existin, DV natural experiment

  +comparisons

  -low generalisability/external validity