Good Practice Made Easy

Home

Reliable Difference Scorer (Adult Tests)

IMPORTANT: The 'reliable difference' of two scores on psychometric tests refers to the statistical likelihood that those scores are (or are not) the result of measurement error alone.

Reliable difference should not be confused with measures of how 'abnormal' or 'unusual' the difference is. It is entirely possible for two scores to reliably differ but for that difference to occur naturally (i.e. natural strengths or weaknesses)

To calculate how 'abnornal' or 'unusual' a difference is requires either knowledge of the distribution of the differences (e.g. population base rates from the test manual) or by statistical methods such as those described by Crawford (2003).

This calculator is intended as an aid to manual scoring, not a replacement. Like all statistical analyses, reliable difference scores are an interpretation tool, which have their own strengths and limitations and should be understood in terms of a wider clinical formulation.

Client Age: Test you wish to compare scores against:

Results

Comparison	Difference	SEdiff	Z	Base Rate (%)	Label

*These calculations utilise the best available reliability data for each test- in most cases taken from rates published by the test manufacturer. Age-stratified, internal consistency/Cronbach Alpha coefficients are generally utilised unless otherwise contraindicated, as in the case of CVLT3. In the case of multiple possible reliability coefficients, we have utilised the most conservative. Methods based on classical statistical theory and those suggested by Chelune et al (1993) and Lezak et al (2012).