Cave is a test of visual memory, involving spatial and temporal sequences. However, since the stimulus items for Cave can be encoded by use of verbal labels, the part played by verbal memory skills in this task is potentially as great as that played by visual memory. Although auditory-verbal memory is usually regarded as being of greatest significance where literacy skills are concerned (see next section), there is good evidence that visual memory tasks can also give good indications of dyslexia and literacy difficulties (Awaida and Beech, 1995; Beech, 1997; Singleton, Thomas and Leedale, 1996; Singleton, Thomas and Horne, 2000). Hence in cases of literacy difficulties it is important for the teacher to know whether the student’s visual memory skills are weak or strong, as these will not only affect the diagnosis but also have implications for subsequent teaching recommendations.

Although working memory is typically conceptualised as being a phonological system subserving speech, a visual equivalent known as the ‘visuo-spatial scratch pad’ has been hypothesised (Baddeley, 1996). This is believed to enable us to keep small amounts of visual information in short-term memory. Such a system is important in developing visual strategies in reading, especially those used by beginning readers (‘look and say’). Models of reading acquisition suggest that visual memory is particularly important in the early stages of learning to read (e.g. Ehri, 1995; Frith, 1985). Conclusions reported by Passenger, Stuart and Terrell (2000) from their study of 80 pre-literate students during their first year of formal schooling lend some support for this view. Stuart, Masterson and Dixon (2000) also found that visual memory influences the acquisition of sight vocabulary in students aged 5 who displayed poor graphophonic skills (i.e. those who had not yet acquired the ability to segment words on the basis of their sounds and who displayed little or no knowledge of sound-to-letter mappings). For students with good graphophonic skills, however, no association between visual memory and word learning was found. Visual memory is also essential in rapid retrieval of visual whole-word representations from the mental lexicon by older and more fluent readers when reading text (particularly of irregular words for which a phonic strategy would not be appropriate). Visual memory also comes into play when retrieving visual sequences of letters in the correct order for spelling (again, particularly where irregular words are concerned). Hence visual memory is a key component of literacy development.

There is also evidence that poor readers have a bias towards visual encoding of words. Johnston and Anderson (1998) reported that poor readers showed a preference for using pictorial rather than verbal information, which they suggest may arise from previous difficulties in learning to attach verbal labels to visual stimuli. Ellis, McDougall and Monk (1996) reported that dyslexics aged 10 years were significantly faster on some visual processing tasks (e.g. picture categorisation) than other groups, including reading age (RA) controls. On word recognition tasks in which the words are paired with either visually similar cues or phonologically similar cues, poor readers are known to perform better than RA controls on the visually similar cue items but not on the phonologically similar cue items (Holligan and Johnston, 1988; Rack, 1987). In other words, they display a less pronounced phonological similarity effect and a more pronounced visual similarity effect (Katz, 1986; Mann and Liberman, 1984).

Palmer (2000) used the Corsi Blocks test to measure visuospatial span in three groups of 14 year-old students: dyslexics, RA controls, and chronological age (CA) controls with normal reading ability. The Corsi Blocks test comprises a set of nine blocks fixed to a base in a predetermined pattern. The test administrator touches the blocks in a set sequence and the testee is required to recall that sequence by touching the same blocks in the same order. This has a direct parallel with Cave in LASS. Palmer found that the dyslexic group significantly outperformed the RA controls on this test. The results also suggested that while all participants showed evidence of using phonological coding to remember pictures, only those in the dyslexic group used visual coding. 

Another study by Palmer (2001) provides further evidence that it is useful for teachers to know about students’ visual memory skills. In this experiment, it was found that students who maintained a visual representation of words alongside a phonological representation after age 7, were significantly worse readers than those for whom the ability to switch strategies by inhibiting the visual representation had fully developed. Students with good visual memory but poor auditory-verbal memory would not only be expected to find acquisition of an effective phonological decoding strategy in reading rather difficult, but also be inclined to rely for a longer period on visual strategies. This approach is liable to run into trouble as the student’s education progresses and the number of new words with which the student is confronted steadily increases.

Some teachers and psychologists assume that problems with short-term memory are entirely verbal rather than visual. However, research suggests otherwise. Awaida and Beech (1995) found that ability to remember letter-like forms at four years of age correlated with reading skills one year later. There is a substantial literature on subtypes of dyslexia, in which visual deficits predominate (Thomson, 1993; Pumfrey and Reason, 1991). Some tests for dyslexia incorporate visual memory tests, e.g. The Aston Index (Newton and Thomson, 1982) and the Coding sub-test of WISC-IIIUK, which is one of the key elements of the A–C–I–D profile that is often used as an indicator of dyslexic difficulties (Thomson, 1993). In the CoPS project, WISC Coding given at age 8:0 had a correlation of 0.36 (p<0.05) with the Rabbits test (a forerunner of Cave) administered at age 5 (Singleton, Thomas and Horne, 2000). Fein, Davenport, Yingling and Galin (1988) found that visual memory is a factor which may be separated from verbal memory in some cases of dyslexia. Finally, there are a variety of other research themes focusing on more physiological aspects of ‘visual dyslexia’, including work on visual discomfort (e.g. Wilkins, 1991; Wilkins et al, 2001); atypical eye movements in reading (e.g. Pavlidis, 1985); ocular dominance (e.g. Stein, 1991, 1994; Stein, Talcott and Witton, 2001); and defects in the transient visual system (e.g. Lovegrove, 1991, 1993, 1994). However, perhaps with the exception of visual discomfort (sometimes referred to as the ‘Irlen syndrome’, after Irlen, 1991) the evidence on some of these physiological issues at the present time seems to be equivocal and more research is required before they can be of practical value in diagnosis and education (Stanley, 1994).

Cave also requires careful concentration and good visual attentiveness, since the stimulus items are only displayed for very brief periods of time. Therefore it is possible for a student to perform poorly on Cave not because of inherent memory difficulties, but because of attention deficit disorder. Where this appears to be a serious possibility, teachers should refer to other information about a student in order to resolve the issue, or refer the student to an educational psychologist for further investigation. Teachers should be aware that it is possible for students to have attention deficit disorder (ADD) without hyperactivity (the latter usually being referred to as AD/HD). ADD (sometimes styled ‘AD/HD without hyperactivity’) is characterised by persistently poor concentration and attention, daydreaming and passivity. Unlike AD/HD, it is more common in girls and often goes undiagnosed, but can be a significant cause of learning difficulties (Cooper and Ideus, 1995). Students with AD/HD who have hyperactive patterns of behaviour may also experience difficulties with Cave because of high impulsivity, which can disrupt the processes of memorisation and recall.

When interpreting the results from Cave, as well as determining whether scores fall below the critical thresholds (see Section 4.1.2), significant discrepancies between the scores on this module and that on the Reasoning module can also be taken into account. In such cases the procedure described in the section ‘Calculating discrepancy’ (Section 4.3.3) should be employed. Teachers should be aware that students with very good scores on Cave (or who show marked discrepancies between scores on this test and Mobile) may develop over-reliance on visual strategies in reading, with a consequent neglect of phonic strategies.