Tactile and kinesthetic perceptual processes within the taxonomy of human cognitive abilities.
The present investigation examines tactile and kinesthetic processes in relation to the major, broad organizations of human cognitive abilities. In contrast to the two perceptual domains, most extensively studied previously by differential psychologists -- vision and audition -- "tactile perception" (or "touch") includes modes of perception that do not all originate from excitation of the same receptor site. Consequently, unless the exact nature of the task at hand is described, it is not entirely clear what is meant by touch. Similarly, the kinesthetic sense needs to be precisely defined since it has its own particular variants. This paper begins by examining the various definitions of touch and kinesthesia that guided the current empirical investigation.
Tactile sensations arise from the stimulation of receptors in the skin, but other receptors (from muscles and joints) may contribute as well. Traditionally, a distinction has been drawn between several different skin sensitivities -- to pain, to temperature, and to pressure (see Boring, 1942). The present investigation is not concerned with sensitivity to temperature or pain and thus these concepts will not be discussed any further. The modes of tactile perception with which the present study is concerned have variously been labeled passive and active touch (Gibson, 1962). There is no consensus in usage of these terms such that the basic distinction presently offered should be regarded as a useful guideline in understanding the present research. A more comprehensive classification of different modes of touch can be found in, for example, Loomis and Lederman (1986).
Passive touch is perception arising from the contact of objects with the skin of a resting body part. Its basis is activation of receptors in the skin only. The objects perceived may be stationary or moving. A pat on the shoulder and the shape of a coin pressed into one's palm are examples of stimuli sensed by passive touch.
Active touch (also called haptic perception; see Gibson, 1966) involves movement and active exploration of object's properties. Finding keys in a bag and perceiving properties of a fork one is handling are examples of active touch. Apart from the receptors in the skin, receptors from muscles, tendons, and joints also supply information in this mode of perception -- in other words, kinesthesia (see below) is involved as well.
Two poles of tactile experience can be distinguished -- objective and subjective (Gibson, 1966, p. 99). On the objective side, the properties of objects that an individual can discern by active and/or passive touch include texture, consistency of a substance, size, and shape. On the subjective pole, different amounts of pressure can be discerned, as well as that part of the skin that is stimulated, and so forth.
Kinesthesia has been defined as a sense of movement (active -- initiated by the individual, or passive -- imposed on one's body parts by some external force), of position of body parts, and of muscle tension. It may be used interchangeably with the term proprioception (Clark & Horch, 1986; Sage, 1977). In providing feedback signals for limb position, contributing to motor programming, and so on, its importance in normal human perception and functioning is significant but not commonly recognized. In contrast, people who can see may appreciate the importance of vision by closing their eyes and depriving themselves of this sense for a while. It is harder to imagine what functional purpose the kinesthetic sense serves because we cannot deprive ourselves of input from it voluntarily (Sacks, 1985). People deprived of kinesthesia (due to illness) cannot identify the position of their body parts. They therefore have to learn to compensate for this lack of feeling by controlling their posture and movements through visually based strategies (see Fleury et al., 1995; Sacks, 1985). Sacks (1985, p. 65) has described another type of impairment -- kinesthetic hallucinations -- by recourse to the following case study:(1)
[W]hat he experienced was ... a flutter of ever-changing illusions -- suddenly the floor seemed further, then suddenly nearer, it pitched, it jerked, it tilted -- in his own words `like a ship in heavy seas.' In consequence, he found himself lurching and pitching, unless he looked down at his feet.
In addition to informing the individual of their body's state, kinesthesia provides information about the attributes of external objects (for example, about lengths -- via extent of movement; about weights -- via the amount of tension felt in muscles; and so forth). Kinesthetic perception is based on input from receptors in muscles, tendons, joints, and skin, as well as on sensory experience that is a direct consequence of the commands for voluntary movements issued to the muscles (Clark & Horch, 1986). The latter source of information is absent from the sensation of passive movements.
1.3. Empirical studies of tactile and kinesthetic abilities: a review
In undertaking a review of the literature, it was found that empirical studies of tactile and kinesthetic abilities derive from four (largely disparate) sources.(2) First, there have been experimental studies of tactile and kinesthetic processes, as well as studies examining the relationship between percepts obtained by different sensory modalities and outcomes derived from so-called intermodality conflict (for comprehensive reviews of this literature, see, in particular, Clark & Horch, 1986; Loomis & Lederman, 1986; Welch & Warren, 1986). Flexibility of perceptual-motor coordination, as an ability to adapt to discrepancy within or between sensory modalities, has also been studied (Welch, 1986). Such extant experimental studies provide information that may be useful for the development of new psychometric instruments. The systematic coverage of all the experimental paradigms found in the cited literature would obviously be impractical. Nevertheless, the present research included tasks representative of each of the aforementioned fields of study. Parenthetically, there has been a push within this literature to acknowledge the limited attention given to tactile-kinesthetic processes, with one major proponent even going so far as to "promote an image of muscles as smart instruments" (Turvey, 1996, p. 1151).
Second, tactile-kinesthetic abilities are critical in understanding the psychological processes of the visually impaired whose navigation skills, for instance, depend on nonvisual cues (e.g., Klatzky, Golledge, Loomis, Cicinelli, & Pellegrino, 1995). These studies are most often applied (rather than theoretical) in their focus. Importantly, within this field of inquiry, several studies report on the development of new tactile tests for the blind. Very little is known of how well many of these tasks would be performed by sighted individuals, although this information is clearly pertinent both to psychological theories (e.g., see Thinus-Blanc & Gaunet, 1997) and to scale and test development. Accordingly, tasks deemed representative of this literature were employed in the current design including, in particular, a tactile test of working memory (i.e., Tactile Shapes; see Heller, 1989; Kainthola & Singh, 1992)).
Third, there are a number of neuropsychological studies providing the largest source of information relevant to the present investigation. This is perhaps not surprising because many theoretical and methodological developments of tests within this domain rely on the psychometric approach. Two of the most widely employed neuropsychological batteries -- the Halstead-Reitan (Jarvis & Barth, 1984) and Luria-Nebraska (Golden, Hammeke, & Purisch, 1981) tests -- contain several measures involving stimulation of the tactile and kinesthetic senses. Although both batteries have been subjected to empirical investigation, findings from studies of the Halstead-Reitan Neuropsychological Test Battery (HRB) are of particular relevance to present concerns. This comprehensive battery consists of tests of general abilities and measures of visual, auditory, and tactile processing. In addition, other measures of neuropsychological function (e.g., a test screening for aphasia) are administered during the complete protocol. Although neurological impairment may not be directly relevant to the discussion of the performance of "normal" individuals on all tasks, several factor analytic investigations of this battery have been published within the past 20 years. These studies indicate that the HRB may indeed tap meaningful individual differences (e.g., Moehle, Rasmussen, & Fitzhugh-Bell, 1990; Newby, Hallenbeck, & Embertson, 1983; Royce, Yeudall, & Bock, 1976; Swiercinski, 1979). Two findings from that research corpus are of direct relevance. According to Moehle et al. (1990), there is considerable agreement in showing that one test from the HRB -- Tactual Performance (both time and location scores) -- loads on a factor defined by four Wechsler Adult Intelligence Scale (WAIS) tests. Notably, these WAIS tests would otherwise form the Performance subscale. In terms of fluid intelligence (Gf)/crystallized intelligence (Gc) theory, tactual performance (TP) appears, therefore, to be a good measure of Gf (see Pallier, Roberts, & Stankov, 2000). Furthermore, in two studies, a well-defined tactile perception factor was found (see Moehle et al., 1990; Swiercinski, 1979). This factor was defined by tests of Finger Tapping (identify which finger, out of the participant's field of vision, was tapped), Stereognosis (recognition of an object derived through touch), and Graphesthesia (the ability to recognize characters drawn on the skin).
Fourth, individual differences research aimed at discovering the structure of tactile and kinesthetic abilities has been conducted but remains particularly sparse. Thus, Carroll (1993, pp. 546-547) cites three studies in which tactile (TS) and kinesthetic sensitivity (KS) factors appeared. However, since one of these studies had a single tactile test (Moursy, 1952), little can be gained from considering it in the current context. Of the three tactile factors identified in the other two studies (Adevai, Silverman, & McGough, 1968; Royce et al., 1976), one is `inconclusive.' The two other factors are `TP' (defined by the Halstead-Reitan test of the same name) and a factor tentatively labeled as "General Orientation of Bodily Locations." Both of these latter factors appear replicable and relevant to issues of interest in understanding individual differences in tactile and kinesthetic perception.
These findings notwithstanding a study by Roberts, Stankov, Pallier, and Dolph (1997) were an attempt to redress the meager evidence for tactile-kinesthetic sensitivity factors suggested in Carroll's (1993) review of the psychometric literature. This study examined eight tactile tasks, including the Fingertip Writing and Tactual Performance subtests of the HRB. One of the eight tests -- Tactile Texture (a task requiring the participant to match one of five textures of sandpaper felt with the dominant hand to a single patch felt with the nondominant hand) -- resulted in low communality. The remaining tests all loaded on the same factor. However, the derived factor had equally high (and salient) loadings from established markers of broad visualization (Gv). The conclusion was relatively unequivocal: tactile/kinesthetic (TK) abilities appear inextricably linked to the psychological processes defining Gv. Furthermore, since the intercorrelation between TK/Gv and a factor that was interpreted as Gf was very high (.80), it became clear that the processes underlying TK, Gv, and Gf are difficult to disentangle. Indeed, in an exploratory factor analysis (EFA) conducted by Roberts et al. to reduce experimental dependence, TK measures remained linked to paradigms assessing Gf and Gv. Under the circumstances, there appeared little option but to conclude that primary tactile factors (e.g., the TP factor identified during this investigation) were perhaps an aspect of Gv, Gf, or both. In turn, this outcome made it doubtful that a broad TK factor [similar to broad auditory reception (Ga) or Gv] could be isolated from other second-stratum ability constructs.(3)
1.4. Present study: overview
1.4.1. Guiding principles for task selection
Roberts et al. (1997) suggested that one reason TK could not be distinguished from Gf and Gv might be because most tests used in their study (particularly those employed in neuropsychological assessment) were cognitively demanding (i.e., complex). In short, paradigms containing an element of memory or reasoning place greater load on the cognitive architecture than tasks requiring low level, sensory processing. Roberts et al. hypothesized that by including a number of simpler tests (like the low communality Tactile Texture test) within the context of a multivariate design, a TK factor (or factors) should emerge.(4) To this end, while retaining the majority of previously used tests, a battery of new, simpler tests was developed for the present investigation. Indeed, of the 14 tests of tactile and kinesthetic abilities employed herein, only five place demands on memory and reasoning to any significant degree. The choice of tests to be included in the multivariate design was guided by two additional considerations.
First, tests used in neuropsychological assessment call upon processes engaging neither the tactile nor kinesthetic modalities in isolation but rather some combination of these two media. Notably, virtually all tests originally used by Roberts et al. (1997) contained mixtures of tactile and kinesthetic processes to varying degree. The choice of simpler tests for the present study opened the possibility of including more nearly "pure" tactile and kinesthetic measures. Thus, the Localization Error test (see Variable 21 below) would appear a prototypical measure of tactile processing, while several other tests (Variables 18, 19, and 21) are more pure measures of proprioception. With this (and the preceding modification), it was hypothesized that separate TS and KS factors would be identified more uniquely in the course of the current investigation.
Second, in the search for new paradigms to be used in individual differences research and in particular perceptual paradigms with low cognitive demand, a cross-modal matching task was included in the present design. Experimental studies of cross-modal matching of length between kinesthesia and vision have shown that there exists a systematic bias in length estimation based on kinesthetic input relative to visual output (and vice versa; e.g., Jastrow, 1886; Seizova-Cajic 1998). Noteworthy individual differences have been observed in one of these studies (Seizova-Cajic, 1998), although this finding requires replication and extension to an abilities framework.
1.4.2. A system for classifying tactile and kinesthetic tasks
In the present investigation, both the properties of objects perceived and the number and type of modalities employed vary across tasks. With respect to the object of perception, the tasks included in the current study require the perception of texture, two- and three-dimensional forms, extent and spatial orientation of objects, location of the stimulus on the skin, and body position. In attempts to classify the tasks with respect to the number and type of modalities involved, care should be taken in order to avoid simplistic and artificial definitions of modality. Thus, in certain measures of active touch, the role played by kinesthesia would appear negligible in comparison to the role of the skin sense. (For example, Test 10 where the participants are required to rub the surface of sandpaper to determine texture.) In other paradigms, the reverse applies (e.g., Test 17 where participants are required to judge an extent by running their hand over a surface). Similarly, kinesthesia is not a simple modality including, in particular, sensitivity for active movements, passive movements, joint position, and muscle tension. Obviously, these sensory elements are not all based on the same type of processing or are they necessarily related in terms of performance outcome. Indeed, factorial studies of kinesthesia have shown that it is composed of a number of specific components (see Sage, 1977 for a review).
Taking the limitations suggested by the preceding discussion into account, it should nonetheless be possible to provide a rudimentary classification of tactile-kinesthetic tasks according to both modality and certain fundamental properties of the objects that are perceived. Although it is possible to generate other criteria for classification (e.g., task requirements -- detection, discrimination, recognition, and matching), the decision to employ the present system was based on the likely high salience of modality (and perceptual object) across all tests. In Section 2, relevant `taxonomic' information is provided for each task composing the `experimental' portion of the test battery. Note that tasks are first grouped under specific modalities, with objects of perception underlined. In providing this system of classification, an attempt is made to highlight the `breadth' of tactile and kinesthetic processes sampled across the present multivariate design.
1. Kinesthesia: Spatial Orientation (Test 18); Two-Dimensional Form (Test 19); Body Position and Movement (Test 22).
2. Passive Touch: Location on the Skin (Tests 9 and 21); Two-Dimensional Form (Test 15).
3. Active Touch: Texture (Test 10); Three-Dimensional Form (Test 13).
4. Intermodal comparison between active touch and vision: Two-Dimensional Form (Tests 11, 12, and 14); Three-Dimensional Form (Test 16); Extent (Test 17).
5. Intermodal comparison between passive touch and vision: Two-Dimensional Form (Test 12).
Two of the aforementioned tasks served an additional purpose -- assessment of motor performance without the aid of vision [Halstead-Reitan Tactual Performance-Time (Test 13) and Bead Construction (Test 16)]. Given the broad aims of the investigation, it was also deemed expedient to measure the flexibility of visual-motor coordination (Mirror Tracing, Test 20). A specific rationale for including this task is also provided by recent neurological research with deafferented patients using analogous procedures (e.g., Fleury et al., 1995; Lajoie et al., 1992; Teasdale et al., 1993). This research has been influential in broadening an understanding of tactile and kinesthetic perceptual processes (Azar, 1998).
It is worth noting that visual imagery could be used by many individuals to help solve almost all of the non-visual tasks (except, perhaps, Tactile Texture and Localization Error). However, attempting to limit the influence of this factor would appear contrary to the main aims of the current investigation. In support of this claim, consider the fact that the use of visual imagery is likely to be a natural tendency for sighted people when deprived of vision. Note also that imagery is thought to be an important component of cognition whether the individual is sighted or not (Thinus-Blanc & Gaunet, 1997).
The participants in this study were 116 (77 female) first-year Psychology students from the University of Sydney, Australia. The average age of the sample was 20.77 (S.D. = 4.13) years.(5) Because of the potential importance of handedness in tactile perception, consideration was also given to this variable. The majority of participants (i.e., N = 104) were right-handed, with an equal number (N = 6) reporting being either ambidextrous or left-handed. Representation within each of these respective groups was as might be expected across the entire population. In addition, for 64 participants, height measurements were recorded. These ranged from 152 to 201 cm with a mean of 171.13 (S.D. = 10.38) cm. This information was obtained in order to partial-out possible confounding effects caused by individual differences in the related physical variable of reach, which might have been relevant in successfully performing some kinesthetic tasks. However, this variable is not considered further because no significant and/or meaningful relation between height and any of the psychological measures was observed.
2.2. Psychometric measures
The psychometric tests employed in this investigation have routinely been used in studies of intelligence carried out at the University of Sydney and elsewhere (e.g., Horn & Noll, 1994; Roberts et al., 1997; Stankov & Crawford, 1997). Within the framework provided by the theory of crystallized (Gc) and fluid intelligence (Gf), these tests are often used, as they provide added clarity to examination of the structural model underlying cognitive ability concepts (e.g., see Roberts, Goff, & Kyllonen, 1998). However, because a review of the literature had suggested little relationship between Gc and tactile measures (a finding replicated by Roberts et al., 1997), it was decided a priori to focus on the constructs of Gf and Gv. A psychometric measure of processing in the auditory medium (Test 8) was also introduced into the design in order to explore the relations between lower level, sensory functions.
The tests employed in this psychometric battery serve to demarcate second-stratum cognitive ability constructs that are consistent with the theory of Gf and Gc. Indeed, references to these tests may be found in the sources cited above. For the sake of brevity, readers are referred to these citations for more complete descriptions of each test.
2.2.1. Fluid intelligence (Gf)
1. Raven's Progressive Matrices Test.
2. Letter Series test.
3. Letter Counting test.
2.2.2. Broad visualization (Gv)(6)
4. Paper Folding test.
5. Hidden Words test.
6. Card Rotations test.
7. Line Length test. In this task, participants were asked to select a line that is longer than four other vertical lines presented on a computer screen. This test has been used extensively in recent studies of self-confidence (e.g., see Stankov & Crawford, 1997).
2.2.3. Broad auditory reception (Ga)
8. Pitch Discrimination test. This test was similar in form to the Line Length test (Test 7). Participants had to choose from among five binaurally presented tones the one that was higher in pitch than the remaining four. In all items, tones were randomly chosen from the range between 400 and 900 Hz, and one tone differed from the remaining four by between 2 and 5 Hz.
In order to make comparisons with mean presented in Roberts et al. (1997), all eight tests listed above were scored in terms of `percentage correct' (i.e., number correct divided by the number of items in the test multiplied by 100). This procedure was adopted because a fewer number of items constituted the tasks of this study relative to the Roberts et al.'s investigation.
2.2.4. Tasks used in the Roberts et al.'s (1997) study
Unless otherwise stated, scoring of tactile-kinesthetic tasks was also in terms of a `percentage correct' score. A more extensive description of these tasks, along with their original source, may be found in the Roberts et al.'s (1997) paper.
9. Finger Counting test. In terms of its cognitive requirements, this test is similar to Test 3 (Letter Counting). Thus, it measures memory for the number and serial position of sequentially administered stimuli presented, however, in the tactile modality. To this end, the four fingers of the participants' right hand were tapped one or more times in a prearranged sequence. Upon the cue word `Recall,' participants recalled the number of times each finger had been touched (12 items).
10. Tactile Texture test. In this test, participants were required to match the texture of a large square of sandpaper, felt with the nondominant hand, to one of five smaller squares that were touched using the dominant hand (12 items).
11. Tactile Shapes test. In this cross-modal matching test, shapes (i.e., 5-mm wide unclosed and continuous lines, inscribed into cardboard) requiring active touch were subsequently matched to one of five visually presented analogues printed on paper. The visually presented lines were 2-mm thick and proportionately smaller (but otherwise identical in shape) to the tactile stimuli (12 items).
12. Gibson's Touch test. There were actually two tasks composing this paradigm.
(a) Stationary Stimulus Condition. This task involved the perception, by passive touch, of stationary shapes (cookie cutters) pressed into the palm of an individual's dominant hand for 3 s. The participant's task was to match the tactile stimuli to one of five visually presented two-dimensional shapes (15 items).
(b) Rotating Stimulus Condition. This task was identical to the Stationary Stimulus subtest, except for one important modification. Instead of the stimulus remaining stationary when pressed into the participants palm, the shape was pressed on the same region and then rotated four times in the following sequence: clockwise, anticlockwise, clockwise again, and thence anticlockwise. Note that this task differs from the one reported by Roberts et al. (1997) who carried out only one (180 [degrees]) rotation of the stimulus.
The sum of the two conditions constituting Gibson's Touch Test was employed in all subsequent analyses. A rationale for this decision is provided in Section 3.1.
13. Halstead-Reitan Tactual Performance (Time) Test. This test required the participant to discriminate and identify the shape of blocks manipulated with the hand (or hands), as well as the forms of shapes incised into a formboard. The requisite response involved matching these two stimuli and, more especially, positioning each block within its appropriate location. The task consisted of three trials -- a condition where the dominant hand was used, another where the nondominant hand was employed, and a final trial, where both hands could be used, to solve the `puzzle.' The score used in the ensuing analyses represents total time (in s) across the three experimental conditions.(7)
14. Halstead-Reitan Tactual Performance (Level) Test. Upon completion of the third trial of Test 13, the formboard was removed from the participant's field of vision. At this stage, the participants were asked to remove their blindfold and draw a formboard with each shape in its correct location. Following conventional practice, two performance measures were obtained from this test: the number of shapes correctly recalled (i.e., `Memory' score) and the number of shapes correctly recalled in their exact position (i.e., `Location' score). A composite `Memory' and `Location' score was derived for use in the current context (see Roberts et al., 1997 for a rationale underlying derivation of this composite).
15. Fingertip Writing test. This test (also adopted from the HRB) requires perception of letters drawn sequentially on each of the participant's fingers. The stimuli were the capital letters `C,' `D,' `G,' `M,' `N,' `O,' `P,' `R,' `S,' `U,' `V,' `W,' and `Z.' These characters were written at the rate of 3-5 s each. The criterion for selecting letters was that they could be written in a single movement (20 items).
2.2.5. `New' measures of tactile and kinesthetic abilities
16. Bead Construction test. In this task, the participant was required to place beads on to a pole, in a prescribed sequence, as fast as possible.(8) In so doing, the test measured perception of three-dimensional forms by active touch, as well as motor performance (since it required participants to put beads on a stand). This test actually represents a major modification of the Bead Memory test of the Stanford-Binet Intelligence Scale: Fourth Edition (Thorndike, Hagen, & Sattler, 1985). All of the equipment used in that test was employed here. The sequence was prescribed by presenting a drawing of the beads on the stand. This visual standard was present all the while participants performed the task. In order to see the standard, but neither the experimental materials nor their own hands, participants wore special opaque glasses lowered towards the tip of their nose, with their head slightly inclined. In total, the test consisted of nine trials: three with the dominant hand, three with the nondominant hand, and three in which both hands were used. The final score, presented in Section 3, depicts total time to complete each picture pattern across the nine trials (measured in s).
17. Cross-Modal Matching of Length test. There were two subtasks (i.e., experimental manipulations) composing this test.
(a) No-Feedback Condition. In this cross-modal matching task, participants were required to match the length of a rod perceived by vision to a length of a groove using the kinesthetic sense. Five visual standards (wooden rods) of length 5, 10, 20, 40, and 50 cm were employed. One rod at a time was placed on a stand in a horizontal position, approximately at eye level and 1 m away. A response bar (80-cm long, with a groove cut in its upper surface) was fixed on the table in front of the participants in a frontoparallel plane. The groove had a movable stopper positioned on top of it, which allowed the participant to alter the length of the groove to any value between 0 and 60 cm. To occlude movement from vision, while still allowing the standards to be viewed, participants wore opaque glasses lowered towards the tip of their nose. In total, there were 15 trials given in a completely randomized order to each participant (i.e., three trials at each of the five lengths of the visual standard).
(b) Feedback Condition. Except for the modifications listed below, the design of this task was identical to the previous one. (1) Feedback was provided after each trial -- the rod was taken off the stand and placed next to the groove so that participants could see the discrepancy between their response and the actual length of the rod. Participants were also instructed to use this feedback to correct their next response. (2) Each length was presented on three successive trials so that feedback for that particular length could be used immediately.
The measures chosen to describe performance on both subtests were absolute proportional errors. Responses to the given rod were averaged (across three trials in the first subtest but only across the two trials following feedback in the second subtest). These error scores (i.e., a difference between response and the real length of the rod) were thereafter expressed as a proportion of the actual rod length. The score for each subtest used in all further analyses was an average of these proportions (errors) for different lengths of the standard.
18. Kinesthetic Acuity test. This test measured KS to the orientation of passive arm movement in the horizontal plane and the ability to match this information to the position of visually presented targets. It involved a paradigm first used by Livesey and Intilli (1996) (see also Livesey & Kangas, 1997). The apparatus consisted of two boards (610 x 360 mm) suspended 235 mm vertically apart. In the middle of the top board was a circle (450 mm in diameter) with the numbers 1-28 spaced equally along its perimeter. On the lower board, the same size circle was cut out, with a turntable disk installed in the hole. The disk had a slot (225 mm long x 23 mm wide) cut into it from the center to the edge and a stylus that could be moved bidirectionally along the slot. This stylus was yoked to a handle. At the commencement of the test, the participants were asked to place their dominant hand underneath a curtain in between the two boards and lightly hold the stylus whilst the experimenter took hold of the yoked handle. Each test trial began from the center of the circle. The experimenter moved the stylus along the slot to a target position right below a number on the top board. With the hand held in that position, participants were asked to name the number they thought their hand was under. The score employed in subsequent analyses reported in this paper represented the total number of positions away from the actual target that the participant felt the handle had stopped (16 trials).
19. Kinesthetic Perception (and Memory) test. This task involved KS to passive movements, memory, and cross-modal matching of shape from kinesthesia to vision (Laszlo & Bairstow, 1985). Blindfolded participants were instructed to hold a stylus in their dominant hand in a relaxed manner. Thereafter, their hand was guided around a closed two-dimensional pattern cut into a Perspex disc (which was placed on a turntable in a horizontal plane). Participants were further instructed to feel the shape and to remember its orientation, as they were to reproduce this information at the completion of each trial. Their hand was guided around the pattern twice at relatively slow speed. The hand was then removed from the pattern, which the experimenter then rotated a certain number of degrees. The participants removed their blindfold and tried to reorient the pattern back into its original position. Error (in degrees) was read from the turntable and recorded. Twelve different patterns were used. The final score represented a measure of average error.
20. Mirror Tracing test. This task measured the ability to modify motor control of the hand in order to accommodate changes in visual information that were due to mirror reversals of natural scenes. A specially constructed apparatus allowed participants to see only the mirror image of a drawing and their hand. The participant's task was to trace a path between 4-mm wide double lines depicting two images (the Star of David and the number 69 rotated 90 [degrees]). The score was average time (in s) to complete these two conditions.
21. Localization Error test. This task is slightly modified version of a paradigm used by Weinstein (1968). It measures the discrimination of positions of tactual stimuli successively applied to the palm of the resting hand. Participants were blindfolded and as a precursor to this test had an Y-shaped grid stamped on the palm of their dominant hand. Gradations (3 mm apart) were marked on the grid. Using these markings, the experimenter located stimuli in 2-mm steps. Inside this grid, participants had their palm touched twice with a pencil tip (approximately 1.5 mm wide). The participant's task was to state whether or not they had been touched on the same location of the palm in each instance. The distance between the two locations varied in an ascending or descending series of steps. The largest distance in a series, for which participants claimed that the spot touched was the same, became the localization error for that series. The score for this task was an average of localization error (in mm) over six series of trials.
22. Nail Pointing test. This task measured KS (for position and active movement) and motor performance. Blindfolded participants placed their hands on the table, shoulder-width apart. One hand was placed palm down and fingers spread, and the other upside down, with the index finger pointing skyward. The participant's task was to touch the specified fingernail (e.g., the nail on the middle finger) with the index finger of the other hand. Accuracy was measured across 16 trials (two for each finger except the thumb).
Only one psychometric test (Card Rotations) was given in the traditional paper-and-pencil format. All other psychometric tests were administered via Macintosh computers. Testing was carried out in two sessions, each of approximately 2-h duration, spaced 1 week apart. Each tactile-kinesthetic task was undertaken on a one-on-one basis and involved comprehensive demonstrations by the experimenter and several practice trials.
3.1. Preliminary analyses
The majority of tests forming the present test battery offer a single score reflecting a particular type of processing. In subsequent factor analyses, these scores are employed. The exceptions are two tests (represented by composite scores) that have somewhat mixed processes underlying them: Gibson's Touch (Test 12) and Cross-Modal Matching (Test 17). In each instance, these tasks consist of two subtests. The subtests share much in common but are not identical. The subtests comprising Gibson's Touch differ in terms of the manner of stimulation, while the subtests of Cross-Modal Matching have distinct task requirements. In both cases, there were two main reasons for employing composite scores. First, `component' scores, at least partially, reflected the same, underlying construct. Second, in undertaking various preliminary analyses, the two `component' scores tended to define a single factor. In order to clarify the meaning of the scores used in factor analyses, a brief description of performance in these two tasks follows.
In Gibson's Touch Test, performance in the Rotated Stimulus Condition (M=70.3%, S.D.=16.9%) was superior to the performance in the Stationary Stimulus Condition (M=51.7%, S.D.=17.6%). This is in accordance with the findings of Gibson (1962) and shows that more information is available to the perceptual system if the stimulus is moving over the receptor surface (skin) than if it is stationary. The correlation between the two scores is .50 -- most probably due to the fact that materials and response were analogous in both instances.
In the Cross-Modal Matching test, the score used in both conditions (average absolute error expressed as proportion) represents the amount of mismatch between visually perceived, and kinesthetically reproduced, lengths. In the first subtest, the mismatch reflects natural calibration of the intermodal relationship (for example, a participant might consistently respond to a 50-cm visual standard with a 43-cm movement). However, in the second subtest, this score also reflects learning (i.e., the participant's ability to correct kinesthetically based movement responses using visual feedback). `Natural calibration,' or original mismatch (in the no-feedback condition), corresponds essentially to an intermodal illusion.(9) Therefore, the first score might be said to reflect the size of the illusion. The second score, on the other hand, represents the size of the illusion reduced by the participant's ability to counter the illusion in the presence of feedback. Because both scores reflect individual differences in size of illusion, the correlation between them is significantly different from zero (i.e., r=.38).
3.2. Descriptive statistics
Table 1 presents mean and S.D. for the 22 variables employed in this study. Because most scores are expressed in terms of `percentage correct,' difficulty levels for the majority of variables may be compared quite easily. For example, based on the mean obtained for percentage correct score, relatively easy tests in this battery would appear to be Letter Series (Test 2) and Paper Folding (Test 4). In contrast, the two most difficult tests are Nail Pointing (Test 22) and Letter Counting (Test 3).
Table 1 Descriptive statistics of psychometric and tactile-kinesthetic tests employed in the investigation Mean (Roberts Measure Mean S.D. et al., 1997) Psychometric tests 1. Raven's Matrices 67.86 18.52 85.17(**) 2. Letter Series 78.07 12.70 75.33(*) 3. Letter Counting 47.69 25.85 47.26(**) 4. Paper Folding 77.96 17.70 -- 5. Hidden Words 62.13 14.11 62.96 6. Card Rotations 50.87 14.90 64.95(**) 7. Line Length 60.49 16.45 -- 8. Pitch Discrimination 70.50 22.11 -- Tactile and kinesthetic tests 9. Finger Counting 58.41 19.18 51.78(**) 10. Tactile Texture 69.68 12.64 66.77(*) 11. Tactile Shapes 64.69 18.50 51.91(**) 12. Gibson's Touch 61.02 14.72 75.66(**) 13. Halstead-Reitan (Time)(a) 154.79 54.29 166.52(*) 14. Halstead-Reitan (Memory) 64.10 18.10 65.50 15. Fingertip Writing 79.74 14.47 56.62(**) 16. Bead Construction(a) 459.87 67.21 -- 17. Cross-Modal Matching(b) 0.34 0.12 -- 18. Kinesthetic Acuity(b) 9.21 3.82 -- 19. Kinesthetic Perception(b) 124.54 71.02 -- 20. Mirror Tracing(a) 290.77 154.53 -- 21. Localization Error(b) 2.83 1.44 -- 22. Nail Pointing 36.00 15.11 -- The final column presents data from Roberts et al. (1997) on common measures. All other indices were scored percentage correct. Asterisks indicate significant t tests. (a) Indicates a test in which speed of performance (measured in s) was the dependent variable. (b) Indicates a test in which error score was the dependent variable. (*) p<0.05. (**) p<0.01.
For comparison purposes, the last column presents suitably rescaled mean for paradigms employed in the Roberts et al.'s (1997) study that are common to the current test battery. Six out of the 12 tests are of about the same difficulty level in both instances (Tests 2, 3, 5, 10, 13, and 14). The asterisks indicate the conventional levels of significance for the t tests between the mean that is sensitive to sample size. Using the difference of one S.D. as the threshold for pronounced difference between the two groups, it is apparent that one test [i.e., Fingertip Writing (Test 15)] is easier for this group than it was in the earlier study. Notably, Roberts et al. did not present letters quite so slowly. That fact could partly account for this discrepancy. In five tests common to both studies (Tests 1, 6, 9, 11, and 12), mean differences are pronounced yet smaller than one S.D. In three of these tests, the present average is lower than found previously, while in two tests the outcome is slightly higher. Is it possible to account for these inconsistencies? It appears that several inconsistencies between the two investigations may be explained in a particularly cogent fashion. For example, a 60-item Standard Progressive Matrices test was used in the earlier study, while a selection of 20 items from Standard and Advanced Progressive Matrices was employed in the present design. Accordingly, the present version of the Raven's Matrices Test is, in reality, considerably more difficult. Note also that certain measures are likely to be somewhat less reliable because there were, for administrative purposes, far fewer items presently employed in many of the tactile-kinesthetic paradigms than had been the case in the past. Within this context, one should not rule out practice effects influencing mean performance as well.
Table 2 presents the results of EFA for the 12 common measures.(10) It is apparent that the two factors extracted in these two studies are highly similar in the sense of configural invariance (see Horn & Noll, 1994). In accordance with interpretation of factors presented later in the current study, the first factor of Table 2 represents Gf, while the second factor is undoubtedly `TP.' Overall, the difference in performance between studies is within the broad range of expectations, with Gibson's Touch Test being the only measure showing loading on different factors across the two investigations. In light of close correspondence between the two solutions, it may be concluded that all present measures are behaving within expectations.
Table 2 A comparison of the factor structure of (overlapping) tests used in the present investigation with those employed in Roberts et al. (1997) Roberts et al., Present study 1997 Measure Gf TP Gf TP 1. Raven's Matrices .63(*) .08 .48(*) .38(*) 2. Letter Series .83(*) -.25 .56(*) .18 3. Letter Counting .48(*) .19 .60(*) .08 5. Hidden Words .46(*) .05 .59(*) -.19 6. Card Rotations .29 .39(*) .52(*) .01 9. Finger Counting .27 .47(*) .40(*) .33(*) 10. Tactile Texture -.12 .29 .11 .15 11. Tactile Shapes .25 .23 .43(*) .27 12. Gibson's Touch .35(*) .13 -.04 .63(*) 13. Halstead-Reitan (Time) -.22 .66(*) .26 .46(*) 14. Halstead-Reitan (Memory) .14 .63(*) .16 .41(*) 15. Fingertip Writing .07 .32(*) -.08 .45(*) The factor intercorrelation between Gf and TP in the present study was .40, while the factor intercorrelation between Gf and TP from reanalysis of the Roberts et al.'s (1997) data was .45. (*) All loadings above .30 have been underlined.
Table 3 presents correlations between all tests of this battery. It is clear that some variables have systematic negative correlations. The reasons for negative correlations are twofold. First, for three tests [Halstead-Reitan Tactual Performance-Time (Test 13), Bead Construction (Test 16), and Mirror Tracing (Test 20)], the dependent variable was time. Second, for four tests [Cross-Modal Matching (Test 17), Kinesthetic Acuity (Test 18), Kinesthetic Perception (and Memory) (Test 19), and Localization Error (Test 21)], the dependent variable was some kind of departure from optimal performance, or error score. In both cases (i.e., with time and error measures), low scores indicate superior performance. If these seven variables were to be reflected, positive manifold would become readily apparent. Indeed, in order to aid interpretation, these variables are reflected in all subsequent factor analyses reported in this paper.
Table 3 Correlations between psychometric, tactile, and kinesthetic measures sampled in the present investigation (n=116) Variables 1 2 3 4 5 1. Raven's -- Matrices 2. Letter Series .470 -- 3. Letter Counting .452 .317 -- 4. Paper Folding .542 .483 .319 -- 5. Hidden Words .330 .366 .327 .371 -- 6. Card Rotations .375 .242 .285 .283 .218 7. Line Length .333 .192 .121 .407 .142 8. Pitch .186 .060 .102 .232 .229 Discrimination 9. Finger Counting .254 .281 .329 .206 .182 10. Tactile Texture .081 -.100 .062 -.052 .009 11. Tactile Shapes .285 .166 .294 .304 .027 12. Gibson's .239 .301 .292 .252 .076 Touch 13. Halstead -.338 -.230 -.348 -.280 -.310 (Time) 14. Halstead .328 .116 .333 .357 .253 (Memory) 15. Finger Writing .192 .109 .083 .190 .087 16. Bead -.033 -.084 -.117 -.159 -.139 Construction 17. Cross-Modal -.017 .007 -.198 -.153 -.096 Match 18. Kinesthetic -.366 -.332 -.165 -.338 -.204 Acuity 19. Kinesthetic -.139 -.171 -.128 -.205 -.161 Perception 20. Mirror -.173 -.200 -.083 -.350 -.227 Tracing 21. Localization -.092 -.162 .121 -.064 .024 Error 22. Nail Pointing .168 .053 .015 -.011 .060 Variables 6 7 8 9 10 1. Raven's Matrices 2. Letter Series 3. Letter Counting 4. Paper Folding 5. Hidden Words 6. Card Rotations -- 7. Line Length .191 -- 8. Pitch .222 .165 -- Discrimination 9. Finger Counting .354 .026 .239 -- 10. Tactile Texture .169 -.082 .072 -.016 -- 11. Tactile Shapes .258 .158 .162 .238 -.043 12. Gibson's .211 .128 .079 .249 .104 Touch 13. Halstead -.386 -.064 -.252 -.517 -.236 (Time) 14. Halstead .443 .106 .219 .441 .031 (Memory) 15. Finger Writing .137 -.068 .136 .251 .238 16. Bead -.130 -.063 -.172 -.241 -.013 Construction 17. Cross-Modal -.177 -.089 -.027 -.148 -.143 Match 18. Kinesthetic -.271 -.129 -.124 -.128 .000 Acuity 19. Kinesthetic -.305 -.006 -.205 -.107 -.058 Perception 20. Mirror -.319 -.163 -.177 -.307 -.134 Tracing 21. Localization -.047 .101 -.084 -.122 -.100 Error 22. Nail Pointing .099 -.104 .120 .170 .145 Variables 11 12 13 14 15 16 1. Raven's Matrices 2. Letter Series 3. Letter Counting 4. Paper Folding 5. Hidden Words 6. Card Rotations 7. Line Length 8. Pitch Discrimination 9. Finger Counting 10. Tactile Texture 11. Tactile Shapes -- 12. Gibson's .258 -- Touch 13. Halstead -.256 -.286 -- (Time) 14. Halstead .280 .169 -.522 -- (Memory) 15. Finger Writing .082 .103 -.317 .206 -- 16. Bead .000 -.195 .378 -.070 -.127 -- Construction 17. Cross-Modal -.104 -.275 .109 -.043 .023 .007 Match 18. Kinesthetic -.238 -.144 .067 -.075 -.141 .068 Acuity 19. Kinesthetic -.126 -.122 .213 -.216 -.227 .145 Perception 20. Mirror -.102 -.087 .303 -.304 -.151 .125 Tracing 21. Localization .010 -.086 .157 .032 -.328 .168 Error 22. Nail Pointing .091 .115 -.202 .007 .243 -.151 Variables 17 18 19 20 21 22 1. Raven's Matrices 2. Letter Series 3. Letter Counting 4. Paper Folding 5. Hidden Words 6. Card Rotations 7. Line Length 8. Pitch Discrimination 9. Finger Counting 10. Tactile Texture 11. Tactile Shapes 12. Gibson's Touch 13. Halstead (Time) 14. Halstead (Memory) 15. Finger Writing 16. Bead Construction 17. Cross-Modal -- Match 18. Kinesthetic .080 -- Acuity 19. Kinesthetic .065 .277 -- Perception 20. Mirror -.057 .045 .114 -- Tracing 21. Localization -.072 .002 .014 .005 -- Error 22. Nail Pointing .129 -.242 -.120 -.132 -.170 --
3.3. Exploratory factor analysis
The correlational matrix of Table 3 was analyzed using the maximum likelihood extraction procedure and oblimin rotation. Root-one criterion suggested eight factors but two variables showed a tendency towards a Heywood case (and singlet factors in the rotated EFA solution) when more than five factors were extracted. These problems were not present in the five-factor EFA solution, which in fact had an acceptable [chi square] goodness-of-fit index (i.e., 104.32; df = 131; P = .96). It is worth noting that the first principal component accounted for 22.8% of total variance, indicating that general factor is not strong in these data (see Roberts et al., 1997 who note a similar finding). The factor pattern matrix from his solution is presented in Table 4. All factor loadings greater than .30 are underlined.
Table 4 First- and second-order EFA (maximum likelihood followed by oblimin rotation) of psychometric and tactile-kinesthetic measures Measure Gf Gv(?) KS (a) Factor pattern matrix (first-order EFA) 1. Raven's Matrices .36(*) .33(*) .26 2. Letter Series .34(*) .39(*) .17 3. Letter Counting .74(*) .03 .01 4. Paper Folding .08 .72(*) .12 5. Hidden Words .25 .23 .06 6. Card Rotations .20 -.03 .28 7. Line Length .03 .46(*) .01 8. Pitch Discrimination -.04 .07 .12 9. Finger Counting .37(*) -.14 .00 10. Tactile Texture .04 -.22 .06 11. Tactile Shapes .25 .07 .19 12. Gibson's Touch .41(*) .04 .04 13. Halstead Time .39(*) -.15 -.09 14. Halstead Memory .21 .01 -.04 15. Fingertip Writing -.05 -.05 .16 16. Bead Construction .11 -.02 -.05 17. Cross-Modal Matching .27 .01 .00 18. Kinesthetic Acuity .04 .13 .70(*) 19. Kinesthetic Perception -.02 -.02 .36(*) 20. Mirror Tracing -.09 .20 -.02 21. Localization Error -.11 .02 .00 22. Nails Pointing -.02 -.23 .37(*) Measure TS TP [h.sup.2] (a) Factor pattern matrix (first-order EFA) 1. Raven's Matrices .11 .05 .51 2. Letter Series .30(*) -.17 .47 3. Letter Counting -.07 -.00 .55 4. Paper Folding .16 .25 .77 5. Hidden Words .10 .13 .24 6. Card Rotations -.10 .41(*) .41 7. Line Length -.08 .13 .16 8. Pitch Discrimination .07 .32(*) .26 9. Finger Counting .18 .34(*) .42 10. Tactile Texture .13 .12 .10 11. Tactile Shapes -.10 .16 .21 12. Gibson's Touch .18 -.04 .23 13. Halstead Time .32(*) .49(*) .68 14. Halstead Memory -.13 .69(*) .58 15. Fingertip Writing .39(*) .19 .30 16. Bead Construction .35(*) .13 .18 17. Cross-Modal Matching -.11 -.01 .08 18. Kinesthetic Acuity .08 -.14 .59 19. Kinesthetic Perception .02 .21 .20 20. Mirror Tracing .12 .47(*) .28 21. Localization Error .59(*) -.06 .33 22. Nails Pointing .27 -.03 .29 (b) Factor intercorrelations (first-order EFA) Factor Gf Gv KS TS TP F1. Gf 1.00 F2. Gv(?) .25 1.00 F3. KS .30 .16 1.00 F4. TS .11 -.09 .28 1.00 F5. TP .38 .09 .30 .27 1.00 (c) Factor pattern matrix (second-order EFA) Factor Factor 1 Factor 2 [h.sup.2] 1. Gf .69(*) .36(*) .50 2. Gv(?) .42(*) .00 .18 3. KS .36 .48(*) .28 4. TS .05 .68(*) .49 5. TP .40(*) .51(*) .33 (d) Factor intercorrelation (second-order EFA): r=.21 (*) All loadings above .30 are underlined.
Two tests in the present battery -- Cross-Modal Matching (Test 17) and Tactual Texture (Test 10) -- have low communalities. This outcome makes it extremely unlikely that either test will exhibit salient loading on the common factors. Note also that Hidden Words (Test 5) and Tactile Shapes (Test 11) have no loadings above the .30 level, due largely to the fact that their common variance is split equally across two or more factors. Interpretation of the five-factor solution is as follows.
3.3.1. Factor 1: Gf
The first factor would appear, unequivocally, to represent Gf. Thus, it has salient loadings from all three markers of Gf (i.e., Raven's Matrices, Letter Series, and Letter Counting). Finger Counting (Test 9) is analogous to Letter Counting and therefore represents a natural addition to the above list. Halstead-Reitan Tactual Performance (Time) (Test 13) is a relatively complex tactile test of the present battery that appears also to be measuring Gf. Salient loading from another tactile task [Gibson's Touch (Test 12)] derives from a measure that appears less cognitively demanding. Nonetheless, for all tactile tasks loading on Factor 1, it would appear important to find a solution using visual imagery. Interestingly, the highest loading from known Gf markers (Letter Counting) similarly requires efficient use of visual imagery techniques for optimal performance (see Monty, 1973). Of course, visual imagery is also seen as a critical component of performance in solving items from the Raven's Progressive Matrices Tests (e.g., Jacobs & Vandeventer, 1968).
3.3.2. Factor 2: Gv?
This factor has the highest salient loadings from two perceptual tests of visual processing (Paper Folding and Line Length) and is therefore a good candidate for representing Gv. However, there are two reasons for using this label in a tentative manner. First, the factor has salient loadings from two markers of Gf (Raven's Matrices and Letter Series), only one of which is based on actual pictorial material. The overlap between Gf and Gv was also observed in the Roberts et al.'s (1997) study. It appears difficult to separate Gf and Gv in a battery of tests that includes tactile-kinesthetic measures (see Roberts et al., 1999 for a possible explanation). Second, two visual tests (Hidden Words and Card Rotations) do not exhibit high loadings on this factor.
3.3.3. Factor 3: KS
Three tests of kinesthesia have salient loadings on this factor. The highest loading derives from Test 18 (Kinesthetic Acuity), which measures sensitivity to orientation of a passive arm movement and hand position. The second highest loading [Kinesthetic Perception-Memory (Test 19)] involves perception and memory of passive complex movement. Finally, Test 22 (Nail Pointing) requires the perception of active, pointing movement, and the position of fingers of a resting hand.
It is possible that, in addition to kinesthesia, all three tests had in common the use of visual imagery based on proprioreceptive input. In order to perform Test 18, the participant likely imagined how their hand was moving underneath the upper board of the apparatus. Similarly, in Test 19, the participant plausibly imagined the complex shape and reproduced its position, while in Test 22 they might have formed an image of where each fingernail lay. This may have helped participants with good visual imagery to succeed in Tasks 18 and 19, as they involved matching the kinesthetic percept to a visual counterpart. It is less easy to envisage how imagery might have aided performance in Test 22 because no visual match was required (see, however, Livesey & Kangas, 1997 for an alternative perspective). In any case, Gv is undoubtedly related to visual imagery (Burton, 1998). The near-zero loadings on that factor exhibited by the three aforementioned tests, and low correlation observed at the second order, suggest that this factor is minimally influenced by individual differences in visual imagery.
Finally, given an aim of the present investigation was to isolate a tactile-kinesthetic factor (or factors) with simpler measures than those previously employed, some comment on the cognitive demands of the tasks defining this trait is in order. Within this context, the most complex task seems to be Test 19. Because the response was delayed, a memory component is introduced to performance in this task. Importantly, the two other tasks defining this factor (i.e., Tests 18 and 22) appear to place very little demand on an individual's cognitive capacity.
3.3.4. Factor 4: TS
Loadings on this factor derive from tactile tests involving both passive and active touch. The two tests with the highest loadings [Localization Error (Test 21) and Fingertip Writing (Test 15)] require the use of passive touch and in particular fine discrimination between point stimuli successively applied to the skin. Two other tactile tests with loadings on this factor require the use of active touch in order to detect the shape of three-dimensional objects (beads in Test 16 and geometrical forms in Test 13). Vision was explicitly involved only in Test 16 where the participants were using a visually presented model of a pattern that they reproduced without looking at the objects (beads). Arguably, visual imagery might also have aided performance in Fingertip Writing. However, it is not obvious that this `strategy' would help in performing the other two tasks, especially since there was no visual matching involved (cf. Livesey & Kangas, 1997). Memory and reasoning demands of the tasks with high loadings on this factor are not pronounced. However, two tests with low loadings [Letter Series (Test 2) and Gibson's Touch (Test 13)] do have a memory and/or reasoning component. Notice, however, that these latter tests share considerable overlap with another factor having a strong reasoning component (i.e., Gf). It is therefore appropriate to assume that the present factor is largely sensory in nature.
3.3.5. Factor 5: TP
Two measures derived from the Halstead-Reitan Tactual Performance Test (Variables 13 and 14) have the highest loading on this factor. Additional salient loadings derive from a psychomotor measure [Mirror Tracing (Test 20)], a measure of spatial orientation [Card Rotation (Test 6)], and a measure of auditory reception [Pitch Discrimination (Test 8)]. All three tests are less cognitively demanding than the Halstead-Reitan Tactual Performance Test. However, the psychological processes involved in two of these paradigms (Mirror Tracing and Card Rotations) are easily seen to be linked to the former. Indeed, this fact and the loading of Finger Counting (Test 9) on this factor point to the importance of working memory in performing certain tactual tests (Roberts et al., 1997). Arguably, the factor is broader in scope than the label TP denotes. Its appearance in the present context provides some support to the notion of what Halstead (1947) called `biological intelligence,' a factor linked to complex, Gf-like problem solving tasks that require some type of physical manipulation (see Pallier et al., 2000).
3.4. Second-order EFA
The factor intercorrelation matrix of Table 4 was subjected to factor analysis (maximum likelihood with oblimin rotation). The resulting factor pattern is presented at the bottom section of Table 4. Two factors have latent roots greater than one. Although there is a degree of overlap between the factors (i.e., three out of five first-order factors have salient loadings on both second-order factors), their interpretation is relatively unambiguous. The first factor is mostly Gf and Gv(?), with lower loadings from TP and KS. The second factor has salient loadings from TS, KS, TP, and somewhat lower loading from Gf. In sum, this pattern appears to point to a distinction between traditional ability factors (such as Gf and Gv), on the one hand, and low-level, sensory (i.e., TS and KS) factors on the other. In turn, this finding indicates that, as speculated, the level of cognitive complexity of tactile and kinesthetic measures influences the derivation (and separation) of these sensory constructs.
3.5. Confirmatory factor analysis (CFA)
A series of CFA was also conducted in order to address two issues of theoretical importance. First, it needed to be ascertained whether or not separate Gf and Gv constructs could be identified in the present data set. Second, it needed to be determined whether each of the factors tied specifically to tactile and kinesthetic processes (KS, TS, and possibly TP) would collapse to define a single, broad tactile-kinesthetic [or `general haptic' (Gh)] factor.(11) To resolve these issues, the correlational matrix of Table 3 was analyzed using RAMONA (Version 4.0; see Browne & Mels, 1994).
In order to address the first question, a five-factor solution was attempted (with Tests 1-3 defining Gf and Tests 4-7 defining Gv). For this solution, the three other factors followed the KS, TS, and TP pattern of loadings in Table 4. Although the resulting solution produced acceptable goodness-of-fit statistics, the correlation between the Gf and Gv factors was .990. In other words, the overlap between Factors 1 and 2 (i.e., Gf and Gv?) in Table 4 translates into an extremely high correlation between the underlying constructs. Furthermore, the correlation between Gv and TP was also high (i.e., r=.755). Plausibly this was because Card Rotations defines a TP factor in Table 4 and yet it was "forced" to define the Gv factor in this solution. Overall, this initial analysis suggested that the Gf, Gv, and TP factors derived from CFA share much in common with constructs given in the EFA solution of Table 4. Based on the correspondence between these two solutions, and the results obtained with the second-order EFA solution, it was decided to attempt to fit a model postulating three factors. In this configuration, the first factor spans Gf, Gv, and TP processes, while the two other factors correspond to KS and TS.
The resulting solution is presented in Table 5. This solution produced a [chi square] = 276.51 (df = 204; P = .001). The root mean square error of approximation (RMSEA) equals 0.056. These statistics indicate an acceptable degree of model fit. All but one of the loadings presented in Table 5 [i.e., Cross-Modal Matching's (Variable 17) loading on Gf] are significant. Interestingly, the correlation coefficient between Factors 2 and 3 (KS and TS) is not significant. Although fixing these two values at zero improves the goodness-of-fit statistic, we do not presently report this solution largely because of a conviction that the actual size of these two parameters may be of substantive interest. In particular, the potential zero correlation between KS and TS factors provides a direct answer to the second issue raised in the introduction to this section. The model with "collapsed" KS and TS factors has to produce unacceptable fit statistics. Furthermore, the results of this study indicate that Cross-Modal Matching engages processes that are not required by complex tasks that comprise intelligence test batteries. Interpretation of the three factors presented in Table 5 is as follows.
Table 5 CFA of psychometric, tactile, and kinesthetic measures Measure Gf/Gv KS TS E's 1. Ravens Matrices .68 .54 2. Letter Series .54 .71 3. Letter Counting .56 .69 4. Paper Folding .66 .56 5. Hidden Words .48 .77 6. Card Rotations .58 .66 7. Line Length .33 .89 8. Pitch Discrimination .34 .89 9. Finger Counting .54 .71 10. Tactile Texture .32 .90 11. Tactile Shapes .43 .82 12. Gibson's Touch .40 .84 13. Halstead (Time) .43 .56 .35 14. Halstead (Memory) .59 .66 15. Fingertip Writing .27 .40 .74 16. Bead Construction .44 .80 17. Cross-Modal Matching .18 .97 18. Kinesthetic Acuity .65 .58 19. Kinesthetic Perception .46 .78 20. Mirror Tracing .41 .84 21. Localization Error .33 .89 22. Nail Pointing .32 .90 Factor intercorrelation matrix Factor Factor 1 Factor 2 Factor 3 F1. Gf/Gv 1.00 F2. KS .58 1.00 F3. TS .32 .11 1.00
3.5.1. Factor 1: Gf
This is a broad factor that combines processes underlying traditional measures of Gf (Tests 1-3) and broad sensory (i.e., Gv and Ga) constructs (Tests 4-6 and 8). It also includes particularly complex and/or `impure' measures of tactile-kinesthetic phenomena (Tests 9, 11-14, 17, and 20). The main reason for interpreting this factor as Gf (rather than some combination of these various constituent processes) follows from two lines of evidence.
(a) Theoretically, Halstead's (1947) view of `biological intelligence' is highly similar to what Cattell (1987) and others have subsequently meant by the concept of `Gf.' This proposition has received empirical support in the Roberts et al.'s (1997) investigation (see also Pallier et al., 2000) and in EFA of the present study. It appears that the tactile-kinesthetic tests employed in neuropsychological assessment measure a primary ability (or abilities) of Gf rather than a distinct tactile-kinesthetic factor akin to the other broad, perceptual organizations (i.e., Gv and Ga).
(b) Although there is ample evidence that Gf and Gv are independent constructs (Carroll, 1993), the separation between these two domains is difficult to achieve in the presence of tactile measures (see Roberts et al., 1999). This may derive from the fact that complex (or otherwise `impure') measures of tactile and kinesthetic performance depend to a significant degree on visual imagery. Overall, however, the largest amount of common variance captured by this factor is due to processes underlying Gf, rather than Gv. Note that even in those instances where no tactile measure is included in a test battery, difficulties have been encountered in separating Gf from the processes underlying Gv (e.g., Horn, 1988; Humphreys, 1962).
3.5.2. Factor 2: KS
Because this factor is essentially the same as Factor 3 of the EFA presented in Table 4, its interpretation as KS would appear unequivocal.
3.5.3. Factor 3: TS
In terms of its overall pattern of salient loadings, this factor would appear similar to Factor 4 of Table 4. However, it is necessary to note that the present factor may be more complex than the one obtained through EFA. Thus, two tasks (Tests 13 and 16), requiring the perception of three-dimensional shapes and motor performance, have higher loadings on the present factor than simple measures of TS.
The pattern of factor intercorrelations suggests that both KS and TS share some common variance with the Gf construct. In fact, the correlation between Gf and KS is relatively high (r=.585). Importantly, the correlation between TS and KS is not significantly different from zero, suggesting that a broad tactile-kinesthetic factor may be difficult to identify in future research. Rather, it is likely that two separate perceptual factors, corresponding to independent tactile and kinesthetic modalities, would emerge.
4.1. Tactile and kinesthetic processes: the big picture
As mentioned in Section 1, the results of Roberts et al. (1997) were inconclusive with respect to the existence of an independent broad TK (i.e., `Gh') ability. The present study has helped clarify important issues surrounding the overall structure of human cognitive abilities. It appears that the most crucial feature for the changed outcome is the presence of low-complexity (and relatively `pure') tactile and kinesthetic tests in the present battery. In this section, a synthesis of the findings from these two studies is attempted.
In the Roberts et al.'s (1997) study, tactile-kinesthetic tests (employed in neuropsychological assessment) and measures of visualization defined a single factor (TK/Gv). Furthermore, the correlation between this construct and Gf was extremely high, indicating substantial communality between all three traits. Essentially, the CFA employed in this study resulted in a similar outcome -- TP, Gf, and Gv measures defined a single trait. It would appear that existing neuropsychological tests use tactile and kinesthetic tasks to measure relatively complex cognitive abilities and thus essentially assess Gf and/or Gv. Exploratory analysis of the present study provides a somewhat different picture. In particular, the analysis suggests that it is possible to isolate a separate first-order factor that is based on relatively complex tactile-kinesthetic tasks. This factor was labeled `TP.' However, in agreement with the results of CFA, TP shares some common variance with the Gf and Gv factors at the second order.
Within the neuropsychological literature, there are reports of factor analytic studies of the Halstead-Reitan (Fowler, Zillmer, & Newman, 1988; Yeudall, Reddon, Gill, & Stefanyk, 1987) and Luria-Nebraska Neuropsychological Test Batteries (Boyd & Hooper, 1993). These studies also report correlations between tactile-kinesthetic tests and measures of intelligence [usually one of the Wechsler scales (i.e., WAIS or WISC)]. The results from this body of research are in general agreement with the present findings. Thus, there would appear evidence for a construct best defined by (i.e., having highest loadings from) TP measures (i.e., Variables 13 and 14 of the present study). In turn, these tactile-kinesthetic tests manifest moderate to high correlation with spatial (Gv) and cognitive performance (i.e., Gf) capabilities. It is probably best to conceptualize this TP factor as a Stratum I ability with linkages to the Stratum II factors of Gf and Gv. This proposition was first formalized by Roberts et al. (1997, see Fig. 2b, p. 142), and is presently incorporated in Fig. 1.(12) The derivation of this ability provides some limited support for Gardner's (1983) claims surrounding the existence of bodily kinesthetic intelligence.
The battery of tests employed in the present study contains seven tests that were not used by Roberts et al. (1997), are not part of neuropsychological test batteries, and are generally simpler than those used for assessment purposes. Two factors identified in the present study (KS and TS) have salient loadings from this group of `new' tests. Various conceptual and statistical analyses of these factors indicate that they possess the following qualities:
(a) Both KS and TS are structurally independent of Gf, Gv, and TP. In other words, unlike the tactile-kinesthetic factor identified by Roberts et al. (1997), these factors appear minimally related to the higher stratum of (existing) cognitive abilities.
(b) Both KS and TS tap cognitively simpler processes that are akin to those aspects of performance captured by pure perceptual markers of higher-stratum factors (i.e., Gv and Ga).
(c) Despite sharing features that characterize KS and TS as sensory-perceptual factors, these constructs share relatively little common variance among themselves. This outcome cast doubts on the existence of a broad ability that spans the kinesthetic and tactile domains.
Taken together, the above three points suggest the existence of separate broad kinesthetic (Gk) and tactile (Gh) perceptual organizations at the second stratum of cognitive ability. At present, however, processes captured by KS and TS are rather narrow in comparison to the broad functions represented by other perceptual organizations (i.e., Gv and Ga). Therefore, until further research is conducted, it would appear best to place them among Stratum I abilities (see Carroll, 1993). Fig. 1 illustrates the most salient aspects of this conceptualization.(13)
4.2. Tactile and kinesthetic abilities: future research directions
In addition to exploring the structure of tactile and kinesthetic abilities at the higher order, there would appear several important areas of theoretical and applied focus requiring detailed, systematic evaluation. On the theoretical side, four issues appear worthy of further consideration. Firstly, there is the issue of interaction between modalities, both in the sense of cross-modal matching and in the sense of the relative roles played by tactile and kinesthetic processes in tasks depending on both components (see Welch, 1986). Although the cross-modal matching task of the present study did not correlate with any other test, it would appear important to develop a series of analogous measures. This approach would allow the researcher to explore the possibility that individual differences in cross-modal matching may contribute to an improved understanding of cognition. Secondly, researchers should examine that area of psychological inquiry that combines tactile and kinesthetic information -- dynamic touch (e.g., Pagano, Carello, & Turvey, 1996). The appearance of a KS factor having noteworthy correlation with Gf (together with the greater involvement of KS than TS in dynamic touch) suggests that investigation of individual differences in dynamic touch may prove profitable in the future. Thirdly, there would appear a need to employ valid measures of visual imagery (see Burton, 1998) rather than rely on indirect claims that this ability plays a role in tactile, kinesthetic, or visualization processes. Finally, the work of Li et al. (1998) suggests that it may be useful to distinguish sensory tasks that are based on absolute threshold measurements from tasks that require stimulus discrimination. It remains conceivable that the processes underlying these different tasks will contribute useful information to hierarchical models of cognitive ability.
On the applied side, the study of tactile and kinesthetic abilities may lead to the development of a test battery that may be useful for work with special populations of blind and partially sighted people. Further, it is likely that the processes captured by the tactile and kinesthetic tests of this battery may be useful for the selection of professionals in performing arts (e.g., ballet and playing musical instruments) and sports or other occupations were these senses are clearly vital (e.g., carpentry). Carefully conducted validation studies in this area are presently lacking. The possibility should also not be ruled out that tactile and kinesthetic measures might provide important ancillary information to those instances where psychomotor batteries have added to prediction above and beyond the general intelligence factor (Tirre, 1998).
An Australian Research Council Grant to the first author and a `Departmental Research Grant' from the Psychology Department, University of Sydney to the third author supported this research. Due acknowledgement is given to all contributing institutions. We are indebted to Nick Karadimas for programming the computer tasks used in this investigation, and to Greg Pettersen and Michaela Davies for helping with test administration. Thanks are also due to Scott Chaiken, Patrick Kyllonen, David Livesey, and Gerry Pallier for thoughtful comments on an earlier version of this manuscript. Finally, we are most appreciative of the comments made by Professors Carroll and Hunt, and an anonymous reader who reviewed this paper.
Adevai, G., Silverman, A. J., & McGough, W. E. (1968). Perceptual correlates of the rod-and-flame test. Perceptual and Motor Skills, 26, 1055-1064.
Azar, B. (1998). Why can't this man feel whether or not he's standing up? American Psychological Association Monitor (May).
Boring, E. G. (1942). Human nature vs. sensation: William James and the psychology of the present. American Journal of Psychology, 55, 310-327.
Boyd, T. A., & Hooper, S. R. (1993). WISC-R IQ estimates from the Luria-Nebraska neuropsychological battery. Perceptual and Motor Skills, 77, 683-688.
Browne, M. W., & Mels, G. (1994). RAMONA PC: user's guide. Columbus, OH: Department of Psychology, Ohio State University.
Burton, L. J. (1998). A factorial analysis of visual imagery and spatial abilities. Unpublished PhD Thesis, The University of Southern Queensland, Toowoomba, Australia.
Carroll, J. B. (1993). Human cognitive abilities: a survey of factor-analytic studies. New York: Cambridge Univ. Press.
Cattell, R. B. (1987). Intelligence: its structure, growth, and measurement. Amsterdam: North-Holland.
Clark, F. J., & Horch, K. W. (1986). Kinesthesia. In: K. R. Boff, L. Kaufman, & J. P. Thomas (Eds.), Handbook of perception and human performance (pp. 13-1-13-62). New York: Wiley.
Fleury, M., Bard, C., Teasdale, N., Paillard, J., Jonathon, C., Lajoie, Y., & Lamarre, Y. (1995). Weight judgment: the discrimination capacity of a differentiated subject. Brain, 118, 1149-1156.
Fowler, P. C., Zillmer, E., & Newman, A. C. (1988). A multifactor model of the Halstead-Reitan neuropsychological test battery and its relationship to cognitive status and psychiatric diagnosis. Journal of Clinical Psychology, 44, 898-906.
Gardner, H. (1983). Frames of mind: the theory of multiple intelligence. New York: Basic Books.
Gibson, J. J. (1962). Observations on active touch. Psychological Review, 69, 477-491.
Gibson, J. J. (1966). The senses considered as perceptual systems. London: George Allen & Unwin.
Golden, C. J., Hammeke, T. A., & Purisch, A. D. (1981). The Luria-Nebraska neuropsychological battery. Los Angeles, CA: Western Psychological Services.
Halstead, W. C. (1947). Brain and intelligence: a quantitative study of the frontal lobes. Chicago, IL: University of Chicago Press.
Heller, M. A. (1989). Tactile memory in sighted and blind observers: the influence of orientation and rate of presentation. Perception, 18, 121-133.
Horn, J. L. (1988). Thinking about human abilities. In: J. R. Nesselroade, & R. B. Cattell (Eds.), Handbook of multivariate experimental psychology (pp. 645-648). New York: Plenum.
Horn, J. L. (1998). A basis for research on age differences in cognitive capabilities. In: J. J. McArdle, & R. W. Woodcock (Eds.), Human cognitive abilities in theory and practice (pp. 57-91). Chicago, IL: Riverside.
Horn, J. L., & Hofer, S. M. (1992). Major abilities and development in the adult period. In: R. J. Sternberg, & C. Berg (Eds.), Intellectual development (pp. 44-99). New York: Cambridge Univ. Press.
Horn, J. L., & Noll, J. (1994). A system for understanding cognitive capabilities: a theory and the evidence on which it is based. In: D. K. Detterman (Ed.), Current topics in human intelligence, vol. IV (pp. 151-203). Norwood, NJ: Ablex.
Humphreys, L. G. (1962). The organization of human abilities. American Psychologist, 17, 475-483.
Jacobs, P. I., & Vandeventer, M. (1968). Progressive matrices: an experimental, developmental, nonfactorial analysis. Perceptual and Motor Skills, 27, 759-766.
Jarvis, P. E., & Barth, J. T. (1984). Halstead-Reitan test battery: an interpretive guide. Odessa, FL: Psychological Assessment Resources.
Jastrow, J. (1886). The perception of space by disparate modalities. Mind, 11, 539-554.
Jensen, A. R. (1998). The g factor: the science of mental ability. Westport, CT: Praeger.
Kainthola, S. D., & Singh, T. B. (1992). A test of tactile concentration and short-term memory. Journal of Visual Impairment and Blindness, 86, 219-221 (May).
Klatzky, R. L., Golledge, R. G., Loomis, J. M., Cicinelli, J. G., & Pellegrino, J. W. (1995). Performance of blind and sighted persons on spatial tasks. Journal of Visual Impairment and Blindness, 89, 70-82.
Lajoie, Y., Paillard, J., Teasdale, N., Bard, C., Fleury, M., Forget, R., & Lamarre, Y. (1992). Mirror drawing in a differentiated patient and normal subjects: visuoproprioceptive conflict. Neurology, 42, 1104-1106.
Laszlo, J. J., & Bairstow, P. J. (1985). Perceptual-motor behavior: developmental assessment and therapy. Los Angeles, CA: Praeger.
Li, S.-C., Jordanova, M., & Lindenberger, U. (1998). From good senses to good sense: a link between tactile information processing and intelligence. Intelligence, 26, 99-122.
Lindenberger, U., & Baltes, P. B. (1994). Sensory functioning and intelligence in old age: a strong connection. Psychology and Aging, 9, 339-355.
Lindenberger, U., & Baltes, P. B. (1997). Intellectual functioning in old and very old age: cross sectional results from the Berlin Aging Study. Psychology and Aging, 12, 410-432.
Livesey, D. J., & Intilli, D. (1996). A gender difference in visual-spatial ability in 4-year-old children: effects on performance of a kinesthetic acuity task. Journal of Experimental Child Psychology, 63, 436-446.
Livesey, D. J., & Kangas, H. (1997). The role of visual movement imagery in kinesthetic sensitivity. Unpublished manuscript.
Lohman, D. F. (1987). Spatial abilities as traits, processes, and knowledge. In: R. J. Sternberg (Ed.), Advances in the psychology of human intelligence, vol. 4 (pp. 181-248). Hillsdale, NJ: Lawrence Erlbaum Associates.
Loomis, J. M., & Lederman, S. J. (1986). Tactual perception. In: K. R. Boff, L. Kaufman, & J. P. Thomas (Eds.), Handbook of perception and human performance, vol. 2 (pp. 1-41). New York: Wiley.
Moehle, K. A., Rasmussen, J. L., & Fitzhugh-Bell, K. B. (1990). Factor analysis of neuropsychological tests in an adult sample. International Journal of Neuropsychology, 12, 107-115.
Monty, R. A. (1973). Keeping track of sequential events: implications for design of displays. Ergonomics, 4, 443-454.
Moursy, E. M. (1952). The hierarchical organization of cognitive levels. British Journal of Statistical Psychology, 5, 151-180.
Newby, R. F., Hallenbeck, C. E., & Embertson, S. (1983). Confirmatory factor analysis of four general neuropsychological models with a modified Halstead-Reitan battery. Journal of Clinical Neuropsychology, 5, 115-133.
Pagano, C. C., Carello, C., & Turvey, M. T. (1996). Exteroception and exproprioception by dynamic touch are different functions of the inertia tensor Perception and Psychophysics, 8, 1191-1202.
Pallier, G., Roberts, R. D., & Stankov, L. (2000). Biological vs. psychometric intelligence: Halstead's (1947) distinction revisited. Archives of Clinical Neuropsychology, 15, 205-226.
Roberts, R. D., Goff, G. N., & Kyllonen, P. C. (1998). The Armed Services Vocational Aptitude Battery (ASVAB): not much more than acculturated learning (Gc) ? Unpublished manuscript.
Roberts, R. D., Pallier, G., & Goff, G. N. (1999). Sensory processes within the structure of human cognitive abilities. In: P. L. Ackerman, P. C. Kyllonen, & R. D. Roberts (Eds.), Learning and individual differences: process, trait, and content determinants (pp. 339-368). Washington, DC: American Psychological Association.
Roberts, R. D., & Stankov, L. (1999). Individual differences in speed of mental processing and human cognitive abilities: towards a taxonomic model. Learning and Individual Differences, 11, 1-120.
Roberts, R. D., Stankov, L., Pallier, G., & Dolph, B. (1997). Charting the cognitive sphere: tactile and kinesthetic performance within the structure of intelligence. Intelligence, 25 (2), 111-148.
Royce, J. R., Yeudall, L. T., & Bock, C. (1976). Factor analytic studies of human brain damage: I. First and second-order correlates. Multivariate Behavioral Research, 11, 381-418.
Sacks, O. (1985). The man who mistook his wife for a hat. London: Pan Books.
Sage, G. H. (1977). Introduction to motor behavior: a neuropsychological approach. Reading, MA: Addison-Wesley.
Seizova-Cajic, T. (1998). Size perception by vision and kinesthesia. Perception and Psychophysics, 60, 705-718.
Stankov, L. (1978). Fluid and crystallized intelligence and broad perceptual factors among 11 to 12 year olds. Journal of Educational Psychology, 70 (3), 324-334.
Stankov, L. (1988). Aging, intelligence and attention psychology and aging. Psychology and Aging, 3 (2), 59-74
Stankov, L., & Anstey, K. (1997). Health and cognitive ageing: the emerging role of sensorimotor abilities. Australian Journal on Ageing, 16, 34-39.
Stankov, L., & Crawford, J. (1997). Self-confidence and performance on cognitive tests. Intelligence, 25, 93-109.
Stankov, L., & Horn, J. L. (1980). Human abilities revealed through auditory tests. Journal of Educational Psychology, 72, 19-42.
Stevens, J. C., & Cruz, A. L. (1996). Spatial acuity of touching: ubiquitous decline with aging revealed by repeated threshold testing. Somatosensory and Motor Research, 13, 1-10.
Swiercinski, D. P. (1979). Factorial pattern description of functional abilities in neuropsychological assessment. Perceptual and Motor Skills, 48, 231-241.
Teasdale, N., Forget, R., Bard, C., Paillard, J., Fleury, M., & Lamarre, Y. (1993). The role of proprioreceptive information for the production of isometric forces and for handwriting tasks. Acta Psychologica, 82, 179-191.
Thinus-Blanc, C., & Gaunet, F. (1997). Representation of space in blind persons: vision as a spatial sense? Psychological Bulletin, 121, 20-42.
Thorndike, R. L., Hagen, E. P., & Sattler, J. M. (1985). Stanford-Binet intelligence scale (4th ed.). Chicago, IL: Riverside Publishing.
Tirre, W. C. (1998). Crew selection for uninhabited air vehicles: preliminary investigation of the air vehicle operator. Unpublished manuscript.
Turvey, M. T. (1996). Dynamic touch. American Psychologist, 51, 1134-1152.
Weinstein, S. (1968). Intensive and extensive aspects of tactile sensitivity as a function of body part, sex, and laterality. In: D. R. Kenshalo (Ed.), The skin senses (pp. 195-222). Springfield, IL: Charles C. Thomas Publishers.
Welch, R. B. (1986). Adaptation of space perception In: K. R. Boff, L. Kaufman, & J. P. Thomas (Eds.), Handbook of perception and human performance (pp. 24-1-24-45). New York: Wiley.
Welch, R. B., & Warren, D. H. (1986). Intermodality interactions. In: K. R. Boff, L. Kaufman, & J. P. Thomas (Eds.), Handbook of perception and human performance (pp. 25-1-25-36). New York: Wiley.
Yeudall, L. T., Reddon, J. R., Gill, D. M., & Stefanyk, W. O. (1987). Normative data for the Halstead-Reitan neuropsychological tests stratified by age and sex. Journal of Clinical Psychology, 43, 346-367.
(1) A similar case was also recently described in the American Psychological Association Monitor (see Azar, 1998).
(2) A fifth area of interest might plausibly rest in recent research examining the effects of aging on tactual discrimination (e.g., Stevens & Cruz, 1996). Indeed, Lindenberger and Baltes' (1994, 1997) influential research, isolating the importance of sensory processes in cognitive decline, has arguably not gone far enough since it has focused only on auditory and visual factors (see Roberts, Pallier, & Goff, 1999). A recent study by Stankov and Anstey (1997) points to the importance of all sensory modalities in understanding the influence of aging on cognitive abilities (see also Li, Jordanova, & Lindenberger, 1998). Indeed, there would still appear further areas of investigation into tactile and kinesthetic abilities, which were brought to our attention by one of the Referees of this paper, Professor Hunt. In physiology, the neural bases of tactile perception and kinesthesia are known to differ among themselves, and they are clearly different from other sensory modalities. There also exists a large literature on motor performance that deals with related processes in normal populations.
(3) The stratum on which various cognitive ability constructs lie is thought to be of major theoretical importance (Carroll, 1993). On Stratum I are the various primary mental abilities `discovered' by over 50 years of research within the Thurstonian tradition. On Stratum II are the broad ability organizations encompassed largely by the theory of Gf and Gc. On Stratum III lies the general factor, psychometric G.
(4) Based on this suggestion, Li et al. (1998) recently found tactile pressure sensitivity to be relatively independent of fluid (and general) intelligence. Two other measures (roughness discrimination and part-whole matching) replicated the findings of Roberts et al. (1997) in that they were substantially correlated with Gf. It should be noted, however, that this study included too few tactile tasks (i.e., three) to address structural issues in a compelling fashion.
(5) In studies of individual differences, it is preferable to employ unselected samples of participants. This is because unselected samples, in comparison to selected samples, tend to produce a stronger general factor. There is a general and a specific reason for accepting the present data as valid. First, the overall level of education in our society has increased considerably during the post-WWII years and the so-called Flynn Effect tells us that the increase has been in terms of IQ as well. Much of the argument for the importance of g (see Jensen, 1998) does not consider this aspect since many data sets predate the likely impact of this change. It is possible that current unselected samples would produce a rather diversified factorial structure and that this, rather than a strong general factor, may be the rule today. Furthermore, it is not always the case that selected samples produce a markedly different factorial structure of abilities. For example, our work with auditory abilities (Stankov & Horn, 1980), children (Stankov & Horn, 1978), and aging (Stankov, 1988) employed unselected samples, but the factorial structure remained essentially the same as that obtained with students. Second, the present study employs simple versions of tactile and kinesthetic measures, and it is unlikely that selection with respect to academic achievement will affect the range of performance on such measures.
(6) We have followed Horn (1998) who lists the four tests as measures of Gv. An anonymous reviewer has pointed out that Lohman (1987) has some difficulties with this classification though we note that Carroll's (1993) analysis of extant factor analytic data sets suggests that these are Gv markers.
(7) Roberts et al. (1997) reported the mean over the three conditions of testing in their original paper. This is somewhat idiosyncratic with respect to the way this test's results are more frequently reported (i.e., the sum over the three conditions).
(8) In the Roberts et al. (1997) paper, the same material was used to assess haptic memory, such that the test was entitled Tactual Bead Memory. In the present version of the test, the memory component is clearly absent since the experimental stimulus is always present for visual inspection.
(9) Most of the participants made stable constant errors. (Note that the score we employ disregarded the sign of the error.) In general, participants underestimated shorter lengths and overestimated longer ones (for detailed description and analysis of a similar task, see Seizova-Cajic, 1998). Note also that a larger mismatch does not mean inferior performance in terms of sensitivity. Other analyses that were performed revealed that participants' trial-to-trial consistency (i.e., sensitivity) remained unrelated to the size of mismatch. For example, a participant whose average error was 30% across the trials might be equally (or more) consistent in giving a particular response than a participant whose average error was 5%.
(10) The analysis of the Roberts et al.'s (1997) data is based on 108 participants who took the Gibson's Touch Test in that earlier study.
(11) A brief comment on nomenclature is in order. The astute reader will note that among specialists, as pointed out earlier in the text, haptic perception equates with active touch. However, various dictionaries that we examined defined haptic in such a way as to be inclusive of both touch and kinesthesia. Given appropriate nomenclatures (i.e., Gt -- broad touch or Gs -- broad somatosensory ability) have already been taken up by other cognitive ability factors [i.e., broad mental speed (Roberts & Stankov, 1999) and broad clerical/perceptual speediness (Horn & Hofer, 1992), respectively], we retain the Gh nomenclature.
(12) The label for the corresponding factor in Roberts et al.'s (1997) study was TK rather than TP.
(13) This is also supported by the application of Schmid-Leiman procedure to the exploratory and confirmatory solutions of this paper. KS and TS remain well defined at the first order. Interestingly, exploratory analysis shows that KS is more strongly grounded at the first order than TS, while confirmatory analysis shows the opposite trend.
Department of Psychology, The University of Sydney, Sydney, NSW 2006, Australia
Lazar Stankov, Corresponding author. Tel.: +61-2-351-2157; fax: +61-2-351-2603.
E-mail address: email@example.com (L. Stankov).
Received 11 January 1998; received in revised form 4 June 1999; accepted 5 August 1999
|Printer friendly Cite/link Email Feedback|
|Author:||Stankov, Lazar; Seizova-Cajic, Tatjana; Roberts, Richard D.|
|Date:||Mar 1, 2001|
|Previous Article:||Genetic Influence on the Variance in Coincidence Timing and Its Covariance with IQ: A Twin Study.|
|Next Article:||Genetic and environmental molarity and modularity of cognitive functioning in 2-year-old twins.|