Correlations, Reliability and Validity

• Positive correlation

  both variables increase together e.g increased hours watching TV and increase in aggressive behaviour
• 1. Negative correlation

  one variable increases, another decreases e.g increased stress leads to decrease in health
• Correlation

  the link between two variables - each can vary on scale

   
• Correlation coefficient

  The value of corelation co-efficient is between +1 and -1

  
  

  • +1 indicates a perfect positive correlation
  • -1 indicates a perfect negative correlation
  • 0 indicates no relationship between variables
• The correlation coefficient were looking for psychologists to agree is 0.8 !!
• Reliability

  Refers to whether the data or results is consistent i.e get the same results time and time again
• Internal reliability

  measure of extent to which something is consistent within itself
• External reliability

  this is when consistent results are produced regardless of when the investigation is used or who administers it.
• Ways of assessing reliability

  1. Test re-test
  2. Inter-observer reliability
• 1. Test re-test

  • giving same questionnaire to same participant to see if they get the same results
  • Leave enough time between the two questionnaires so they cant remember their previous answers
  • Correlate the two scores together and if they have a correlation coefficient of at least +0.8 you can assume its reliable
• 2. Inter-observer reliability

  • Both observers have agreed behavioural categories
  • Observers watching the same event
  • Record their data independently from each other
  • Data collected from the observers would be correlated to make sure they're similar
  • Correlation coefficient should exceed 0.8 for reliability
• Questionnaires (improving reliability)

  • Should be answered using test-retest
  • Comparing the two sets of data should produce correlation that exceeds 0.8
  • If it produces low reliability levels, then questions need to be taken out or reworded
• Interviews (improving reliability)

  • Have the same interviewer each time
  • All interviewers need to be properly trained
  • Structured interviews avoid ambiguous and leading questions which could impact reliability
• Observations (improving reliability)

  • Make sure all behavioural categories have been operationalised
  • Categories shouldn’t overlap
  • If reliability is low the observers may need more training in using categories or discuss their decisions
• Experiments (improving reliability)

  Having standardised procedures allows them to be consistent everytime
• What does validity refer to in research studies?

  The legitimacy / accuracy of the study's results
• What is internal validity concerned with?

  Outcomes due to manipulation of the IV
• What factors influence internal validity?

  Confounding variables and participant variables
• What does external validity relate to?

  Generalizability to other settings and populations
• What is ecological validity?

  Generalizability to other situations and settings
• What is temporal validity?

  Generalisability to different historical times
• What is population validity?

  Generalisability to different age groups and cultures
• What is face validity?

  Measure appears to assess what it should
• What is concurrent validity?

  Comparison to a similar existing measure
• What is predictive validity?

  How well a test predicts future behaviors
• How does a control group improve validity in experimental research?

  It helps assess effects of the independent variable
• What is the purpose of standardising procedures in experiments?

  To minimize participant reactivity and bias
• How do lie scales in questionnaires enhance validity?

  They assess consistency of responses
• How can anonymity improve validity in questionnaires?

  It encourages honest responses from participants
• What can negatively impact the validity of observational data?

  Broad, overlapping, or ambiguous categories
• Why are qualitative methods considered to have higher ecological validity?

  They reflect participants' realities in detail
• What is interpretative validity?

  Researcher's interpretation matches participants' views
• Explain how the reliability of the content analysis could be assessed.

  Test-retest reliability -

  • content analysis repeated on a second occasion using the same interview data
  • compare the results of the two separate analysis
  • researchers could calculate the correlation between the two ratings

  Inter-rater reliability - use a second person to work with the original researcher

  • they could tally the occurrences of each of the categories of the interviews (separately)
  • they could compare their tally charts looking for agreement and calculate the correlation between the two ratings