Rivermead Behavioural Memory Test - Third Edition

Why is a new version being published so shortly after the publication of RBMT-II?

The original RBMT (1985) was updated in 2003 and published as RBMT-II. This revision included an update of materials but no redevelopment of test items or re-standardisation was carried out at that stage.

The new version has been extended so that you can assess a wider range of ability. New items are included on tests and materials have been updated. A new subtest (Novel Task) has also been included to assess the ability to learn a new skill. A larger normative sample has also been collected as compared to previous versions of the RBMT.

Is there clinical data on the new form of the RBMT-3?

Yes, new clinical data has been collected with people with a range of clinical difficulties. This data is included in the new manual.

Who can administer the test?

The test can be used by clinical psychologists, occupational therapists and speech and language therapists.

When should I use Version 2 of the RBMT-3?

Version 2 should only be used for retesting and should never be used before Version 1. The normative data was collected so that Version 1 was always presented in the first session and Version 2 was always presented in the second session.

I often use the Wechsler Memory Scale (3rd Edition) (WMS-3). Why would I use the RBMT-3?

The RBMT-3 is a complementary test to the WMS-III. It provides information about everyday memory problems and is more useful for people with severe cognitive difficulties.

Can you tell me more about the new subtest - the Novel Task. What does this assess?

The Novel Task Assesses ability to learn a new task. Deficits in this skill impact on everyday life, and also influence an individual’s capacity to benefit from compensatory aids used in rehabilitation. It is therefore important to establish the level of ability in learning new tasks before attempting to teach them.

The Novel Task is based on a mathematical dissection of a 6 piece puzzle (a star for Version 1 and a square for Version 2). The puzzle is assembled in a set order by the examiner and the examinee is required to remember this. 3 learning trials and a delayed trial are given. The scoring criteria record the order and position that the examinee places the pieces of the puzzle.

Can you give me further information about the normative and clinical samples tested on the RBMT-3?

Click Technical Information.

Can you give me further information about the reliability and normative studies for the RBMT-3?

Click Technical Information.

The score obtained for the General Memory Index (GMI) seems to be lower than I would expect given the scaled scores on the individual subtests. Why is this?

When scores on components (subtests) of a composite deviate from their means in the same direction, the mean of the composite will be more extreme than the individual components. So, if subtest scores are below the subtest mean, the composite will be below the composite mean but more so than would be suggested by the simple average of the subtests.

This effect occurs with any composite (assuming that the composite is made up of correlated components) but the effect is particularly marked in the case of the RBMT because most healthy individuals will score at or near the maximum score on most subtests. Thus, if a patient scores below the maximum on a considerable number of the subtests (even if the scores are not very poor on each subtest considered individually), they are fairly unusual and this is reflected in a low composite score.

Can you explain what the GMI, Percentile Ranking and confidence intervals mean?

With regard to the General Memory Index, it is the standardised score that the test yields, and follows the basic principles of standardised assessment. It is similar to an IQ score with average being 100, and a standard deviation of 15, therefore any score between 85 – 115 is classed as an average score as these scores fall within 1 standard deviation of the mean. Classification of the scores is as follows:

For the General Memory Index / Standardised scores the following rules generally apply:

≤ 69 More than 2 standard deviations from the norm Significant Impairment

70-84

More than 1 standard deviation from the norm Borderline / Moderate Impairment

85-115 Within 1 standard deviation of the norm Average range

116-130 More than 1 standard deviation from the norm Above average range

≥ 131 More than 2 standard deviations from the norm Very high / superior

The subtest level scores are based on an average of 10 with a standard deviation of 3 with the following interpretive ranges:

≤ 3 More than 2 standard deviations from the norm Significant Impairment

4-6 More than 1 standard deviation from the norm Borderline /Moderate Impairment

7-13 Within 1 standard deviation of the norm Average range

14-16 More than 1 standard deviation from the norm Above average range

≥ 17 More than 2 standard deviations from the norm Very high / superior

For percentiles, the following rules generally apply:

≤ 1 More than 2 standard deviations from the norm Significant Impairment

2-15 More than 1 standard deviation from the norm Borderline / Morderate Impairment

16-84 Within 1 standard deviation of the norm Average range

85-98 More than 1 standard deviation from the norm Above average range

≥ 99 More than 2 standard deviations from the norm Very high / superior

A confidence interval reflects the range of scores you would expect the respondent’s scores to fall within on any given day, as it would be unrealistic to expect the individual to obtain exactly the same score each time you administered the test.

Most people tend to use a 95% confidence band as this provides you with a more accurate range of scores i.e. – you are 95% confident that their scores would fall within this range on any given day.