The age range is 6;0 – 89;11. There are 60 items in two equivalent forms (A and B), and it takes around 15 minutes to administer.
The test items look like Raven’s progressive matrices. It appears to measure a single dimension, and not to be criterion referenced, that is, a person’s performance on a single item cannot tell us how they would perform on ‘real life’ problems, because there do not appear to be analogies with real life problems. The only information it yields is the score, which can be compared to US norms.
The Examiner’s Manual says it is a test of abstract reasoning and figural problem solving.
The authors take a particular perspective on intelligence: that problem solving ability is ‘the essence of intelligence’ (p. 1). They equated their view of intelligence with ‘g’ and ‘fluid intelligence’ which are conceptions of a general intelligence ability that have been proposed by some experts on intelligence. They sought to test problem solving in a context that was free of language and significant cultural influence. They also wanted it to be brief and have low motor demands on the participant.
Problem solving in the test is mostly about identifying the rules that have been used to determine the symbols in the grid and applying the rule to determine what the missing one should be. This is probably a teachable skill – possibly emphasized in some cultures more than others. It seems quite mathematical. The authors identify different question types, but this is to show that the two forms (A and B) are very similar, they do not treat them as subtests, and they do not say which test items are of which types, so relative strengths cannot be identified in test takers. The types are: matching, analogy, progression, classification, and intersection. (p. 48)
The authors say that the test is free of culture, but they are still thinking within the US. They feel that being untimed and not needing pencil and paper mean that it is suitable for all cultures, but there may still be people who are unfamiliar with this type of formal, performance-based assessment and would be disadvantaged.
Stated Purposes: ‘Estimating aptitude and general intellectual functioning, identifying individuals believed to have intellectual impairments, ruling out intellectual impairment in individuals whose test performance may be confounded by concurrent language or motor impairments, verifying the validity of referrals for treatment, therapy, or special services, formulating hypotheses for intervention or further evaluation, conducting research.’
It does not appear to help a therapist to identify therapy goals or therapy methods, or to measure progress toward therapy goals.
In administration, basals and ceilings are used. The skills of the tester are defined broadly and refer to people who are skilled in giving this kind of test, but the instructions seem to assume the tester will be a psychologist. Standardized testing procedures are provided.
The test can be introduced with words, and should be unless there is a good reason not to, in which case, non-verbal instructions can be used and are provided. Verbal instructions are provided in 8 languages. There are 6 training items, which can be administered twice if necessary. If the test taker is still not clear what to do, stop the test. Data is presented showing that for people with intact language skills, it makes no difference whether the instructions are given with or without words.
The authors provide a way to compare the performance of one person on the indexed score from the TONI-4 and from another test which gives standardized scores in the same way – mean 100, SD 15 (p. 23) (such as the PLAI-2).
Correlations are given from independent studies of the TONI (previous versions) with other tests of intelligence. The PLAI-2 is not included, but the CELF-R is. This is detailed in an additional text in the pack: Critical reviews and research findings, 1982-2009.
Is there a correlation in the normal population between performance on the TONI-4 and language ability? Would a discrepancy between the two qualify as a delay or disorder?
Do test takers actually use language when solving the problems, e.g. by labeling shapes, sizes, colours? Or are the language centres in their brains inactive? Does it influence their performance? Do test takers ‘turn off’ linguistic representations after a while?
If a person with persistent language difficulties were to be tested with the TONI-4, how should the results be interpreted, especially regarding their potential for benefit from therapy?
Overall, it seems that the TONI-4 has quite a narrow purpose and applicability, and probably should not be part of a regular assessment process. It is more likely to be used where a person had been assessed and discrepancies were found that suggested that a language difficulty, e.g. ESOL, made it very difficult to measure their cognitive abilities. The TONI-4 could be used to investigate those cognitive abilities in a context free of language.
It is one-dimensional and yields a single datum – the raw score. This can be compared with norms, but there does not appear to be a criterion value or face validity relevant to SLT. The underlying construct is not well established (which is not unusual in the field of intelligence where there are many competing theories).
Written by Stephen Robinson (July 2011)