6.14 reliability

Reliability - the results are consistent, if the study is replicated they will get similar results

Collecting data

  • questionnaires and interviews

  • Experimental conditions (e.g EEG)

  • Observations

External reliability - the extent to which a measure is consistent when repeated

Internal reliability - the extent to which different parts are consistent with itself

Assessing reliability

Internal validity

  • split half method - assess measures that test one variable. Split the test into two parts, have participants complete both parts, test the strength of correlation. A strong correlation shows internal validity

External validity

  • test-retest - repeat the study using the same procedures/ measuring devices at different times. Test the correlation between the two versions

  • Inter-rater reliability - two or more observers record behaviours during the same observation. They then test the correlation

  • Test of correlation - use pearsons R or spearman’s Rho. A correlation of 0.8 or higher is accepted as a strong correlation

Improving reliability

  • observation - train the observers, pilot studies can identify poorly defined behavioural categories, so we can operationalise each behavioural category

  • Interviews - structured

  • Questionnaires - closed questions

  • Experiments - standardised procedures, keep environmental conditions the same