RESEARCH METHODS

• Ways to test for reliability

  • Test-retest
  • Inter-rater reliability test
• Things to consider when choosing statistical test

  • Difference or correlation, design, data
• Features in designing questionnaires

  • Likert scales
  • Rating scales
  • Fixed choice option
  • Closed Qs
  • Open Qs
• Aim is a general statement of what the researcher intends to investigate and the purpose of the study
• Attachment - role of the father

  • Bowlby's 'Monotropy Theory' - only child and mother form secure and lasting monotropic bond
  • Schaffer & Emerson's 4th Stage of Attachment 'Multiple Attachments' - important to have more than one attachment including father
• Case studies limitations

  • Researcher bias - the information on the final report is subjectively selected by researcher therefore could miss out vital information that was deemed insignificant
  • Inaccuracy and memory decay - personal accounts from family and friends therefore information isn't 100% reliable
  • Small sample sizes therefore low generalisability
• Case studies strengths

  • Rich detailed insight (vs. experiments/questionnaires = superficial single-focused investigation) therefore increase external validity
  • Studies unusual infrequent behaviours therefore paves way for further research
• Case study

  Detailed study over time of single individual, institution or event providing rich record of experience
• Case study methods

  Diary records, interviews, psychological tests
• Confounding variable

  Type of EV, does systematically vary with IV (related to IV) but affects DV
• Conventions of positively skewed distributions

  • For very difficult tests
  • Mode = the highest scores/values
  • Mean = the extremely low scores/values
  • Dragged to the RIGHT - furthest from y-axis
• Correlation

  The measure of the association between co-variables
• Correlation limitations

  • Can't infer causality - relationship could be due to third/intervening variable(s)
  • Lack of experimental manipulation and control of EVs - only shows relation therefore not know which caused change in DV
• Correlation strengths

  • Precise and quantifiable measure therefore can suggest future ideas if strongly related
  • Quick and economical - controlled environment and variable manipulation unnecessary therefore can be used in natural conditions
  • Can use secondary data therefore less time-consuming
• Critical value (CV)

  Value that's worked out from the table
• Directional hypothesis

  States the direction of the difference or relationship
• Median is easy to calculate and not affected by extreme scores due to low sensitivity BUT is less representative as not all data is used
• Range is easy to calculate BUT only uses two most extreme values therefore not useful and less representative
• Thematic analysis process

  Data is transcribed (where necessary) and is reviewed repeatedly for the identification of trends in the meaning conveyed by language. Re-analyse identified themes so it becomes more refined and relevant. Give short hand codes then annotate transcript with them. The identified themes are used to support or challenge existing theories, with specific examples of data as supporting evidence.
• Field experiment

  An experiment conducted in the participants' natural environment with still controlled variables
• Identified themes

  Used to support or challenge existing theories, with specific examples of data as supporting evidence
• Qualitative methods


  • More depth and detail that reflects P's reality therefore higher mundane realism (eg in case studies and interviews)
  • Triangulation = using a number of different sources - therefore concrete evidence and not researcher's assumptions
• Inter-rater reliability

  Number of agreements/total number of observations
• Hypothesis

  A precise testable statement stating relationship between investigated variables
• Attachment

  Role of the father
• Treatments

  Mental illnesses
• Interval data

  • More informative than nominal and ordinal
  • Equal values between intervals - objective therefore very reliable
• Interval data limitation

  Can be arbitrary (random)
• Investigator effects

  Investigator's behaviours affects the research's outcome (DV)
• Lab experiment

  An experiment that takes place in a controlled environment where the researcher manipulates the IV and records the effect on the DV while maintaining strict control of extraneous variables
• Level of significance

  p <_0.05 (5%) = the likelihood results are due to chance is less than/equal to 5%
• Likert scales


  Questionnaires that require individuals to indicate their degree of agreement or disagreement with a set of statements
• Mean

  • Includes all data therefore most representative and sensitive
• Median

  Ordinal data - wilcoxon, mann-whitney, spearman's rho
• Meta-analysis

  • More generalisable because different samples are combined therefore more representative results
  • Useful for studies on unusual events where samples are quite small
• Meta-analysis weaknesses

  • Lack of comparability - likely to use slightly different methods therefore not valid to combine results
  • Selection bias - researcher selects the studies therefore possibility to select studies that complement their hypothesis and exclude non-significant results therefore low internal validity
• Mode

  • Nominal data - sign test or chi-squared
  • Easy to calculate
• Mundane realism

  The extent to which events in an experiment are likely to occur in the real world
• Natural experiment

  Where the IV naturally exists and isn't directly manipulated
• Nominal data

  • Easy to obtain from closed Qs - interviews or questionnaires therefore more accuracies
  • More data collected quickly therefore efficient therefore high reliability