Assessing Reliability

• Reliability 


  refers to CONSISTANCY of a measuring device or result. If the same results are found it means they are consistent which means results are reliable.
• Test-retest 


  -PP are tested the the SAME PP are assessed again using the same measuring device, in same location, task, instructions and questions asked.

  -results from original and retest are correlated.

  -Spearman's and Pearson's used as is a test of correlation.

  -Correlation co-efficient of 0.8+ is an agreed indicator of reliability.

  -Timing between the first and re-test. Wait too long and PP attitudes may change. Don't wait long enough then PP may remember behaviour/ questions.
• Interobserver reliability


  -Two observers watch the same interaction as each other.

  -This is done independently.

  -Use same recording sheets, codes and operationalised behaviours.

  -If both observers aren't free behaviour is recorded and analysed.

  -Results from both observers are correlated.

  -The significance id assessed using spearman's or Pearson's.

  -Correlation co-efficient 0.8+ means there is a high inter-observer reliability.