quantitative

• Reliability


  How consistent something is
• Validity


  How true it is, is it measuring what it claims to measure
• Research hypotheses

  • Directional
  • Non-directional
• Directional research hypotheses


  Based on previous research and says which direction the research will go
• Non-directional research hypotheses


  Says there will be an effect, just not which one
• Operationalisation of variables


  Putting our hypotheses/variables into measurable terms e.g. anxiety - heart rate
• Population


  All members of a group that we are interested in
• Sample


  Used to make generalisations about populations
• Sampling bias


  Over or under representation of one particular category or group in a sample
• Random samples


  Unbiased, everyone has equal chance
• Stratified samples


  Pre-define groups in population then select randomly, each group is proportionally represented. Useful when sample too small to ensure proportional representation by random sampling
• Quota samples


  Sample is stratified, but selection from each stratum is left up to researcher. May be bias
• Cluster samples


  Naturally occurring groups or clusters containing people in target population. May be bias
• Self-selecting samples


  Participants selected by own actions e.g. volunteers. May be bias
• Opportunity/convenience samples


  May be bias
• Snowball samples


  Participants contact other participants for the researcher. Bias
• Sample size


  Larger the better = more representative, less likely to be biased, easier to find significant results with statistical tests - costly & time consuming
• Purpose of experiments


  Identify cause and effect relationship
• Independent variable (IV)


  Manipulated to cause effect in dependent variable
• Dependent variable (DV)


  Measured to see effect of IV
• Extraneous variables


  Random variables that effect a participant's performance but unpredictably
• Confounding variables


  Change systematically with IV
• Independent groups design


  Split in half, random allocation to conditions (no order effects)
• Repeated measures design


  All participants do all conditions - order effects = counter balancing
• Matched pairs design


  Matched on important variables, random allocation of participants from each pair to a condition: (difficult to find perfect matches)
• Experimenter effects


  Effects caused by what the experimenter expects or wants. Double blind
• Participant effects


  "Hawthorne effect" - single blind
• Strengths of experiments

  • Tight variable control and operationalisation helps identify cause and effect relationship
  • Replication = good reliability
• Weaknesses of experiments


  • Tight variable control can lower construct validity
  • Artificial environment can lead to different behaviour (ecological validity)
• Ordinal level


  Data in relative order, or rank on a scale
• Interval level


  Equal intervals between values, e.g. IQ scores
• Ratio level


  Equal intervals between values and zero point, e.g. reaction times
• Histograms and boxplots


  • Generally used to analyse and explore data
• Bar and line charts


  • Used to present summarised results
• Mode


  Nominal level, unaffected by extreme values, represents most frequent value
• Median


  Ordinal, interval, ratio level, unaffected by extreme values, represents only middle value
• Mean


  Interval/ratio level, represents all values, distorted by extreme values
• Frequency distribution graphs


  • Histograms and box plots used for ordinal data or higher
• Range


  Distance between lowest and highest values, ordinal and interval level, easy to calculate and understand, sensitive
• Standard deviation and variance


  Measure of dispersion around the mean, average amount of deviation from the mean, interval/ratio level, represents all deviations from the mean, sensitive