Abstract

Deficits in episodic memory, related to medial temporal lobe dysfunction, are a core component of schizophrenia. Rey's Auditory Verbal Learning Test (RAVLT) is one of the most widely used neuropsychological memory tasks in clinical practice but normative data are limited. Over the last 15 years, the RAVLT has been used to assess verbal episodic memory performance as part of the Western Australian Family Study of Schizophrenia (WAFSS). This study aims to provide RAVLT norms, derived from the WAFSS, for individuals meeting DSM-IV and ICD-10 diagnostic criteria for schizophrenia or schizophrenia-spectrum disorder, for use in clinical settings. Performance on three immediate and one delayed recall trials, and additional measures of encoding and forgetting, is presented for 492 patients and 260 healthy community controls. Results indicated that age and sex (both groups) and IQ (schizophrenia group only) significantly influenced performance. Norms are presented, as means and standard deviations, stratified accordingly. Additional between-groups analysis clearly shows a significant memory deficit in schizophrenia even when patients are matched on IQ, age, and sex with healthy controls.

Introduction

Episodic memory impairment, linked to abnormal activation in the medial temporal lobes (Achim & Lepage, 2005), is one of the most robust and reliable cognitive difficulties experienced by individuals with schizophrenia (Boyer, Phillips, Rousseau, & Ilivitsky, 2007; Danion, Huron, Vidailhet, & Berna, 2007; Toulopoulou & Murray, 2004) and an important predictor of social and community functioning and quality of life (Green, 1996; Green, Kern, Braff, & Minz, 2000). Consequently, the assessment of episodic memory functioning should be an essential component of routine clinical evaluation for individuals with schizophrenia.

Word list learning and recall remains one of the most reliable, valid, and clinically useful outcome measures for the clinical evaluation of new treatments for schizophrenia (Green et al., 2004), and such tasks continue to be widely used in general clinical practice, partly due to their ease of administration. Several different list learning tests are available, but Rey's Auditory Verbal Learning Test (RAVLT; Rey, 1941) is especially popular when there is limited assessment time and when the influence of semantic organization on recall is not a focus of assessment (Lezak, Howieson, & Loring, 2004). The RAVLT provides easily calculated measures of encoding and retrieval of 15 largely unrelated concrete nouns (reviewed in Schmidt, 1996). However, a wide variety of RAVLT word lists and administration procedures have been reported in the literature, including variations in the number of trials presented, rate of presentation of words, use of interference word lists, and in the timing of delayed recall trials (Lezak et al., 2004; Schmidt, 1996). As commonly noted, this makes it difficult to find appropriate norms, especially when testing specific patient populations such as individuals with schizophrenia.

Previous research also suggests that performance on the RAVLT is sensitive to several moderator variables. Decline in performance with increasing age is the most consistently reported finding (Geffen, Moar, O'Hanlon, Clark, & Geffen, 1990; Query & Megran, 1983; Schmidt, 1996; Van der Elst, Van Boxtel, Van Breukelen, & Jolles, 2005). In the healthy adult population (20–60 years), RAVLT scores are more similar to each other than in older adults (aged over 60 years; Strauss, Sherman, & Spreen, 2006). The influence of sex on performance is less consistent in the literature: where differences are reported, women tend to do better than men (Geffen et al., 1990; Strauss et al., 2006; seeMitrushina, Boone, Razani, & D'Elia, 2005, for nonsignificant findings). Similarly, findings concerning the influence of education and intelligence on RAVLT scores are inconclusive (Mitrushina et al., 2005). Bolla-Wilson and Bleecker (1986) noted that education does not account for variance in RAVLT scores over and above that associated with differences in intelligence, and Schmidt (1996) concluded that in healthy individuals neither education nor intelligence appear to have a substantial influence on RAVLT norms.

Despite the importance of examining verbal episodic memory dysfunction in individuals with schizophrenia, clinicians face several significant hurdles when using the RAVLT with this patient population. First, existing normative data are often provided on mixed samples of psychiatric patients (each of which may exhibit different memory impairments; Schoenberg et al., 2006), rather than a specific sample of individuals with schizophrenia. Secondly, the normative data currently available for individuals with psychiatric illness are limited by extremely small sample sizes (Mungas, 1983) and/or have not been stratified by age, sex, or IQ. This may be particularly important when assessing individuals with schizophrenia. For example, recent evidence suggests that women with schizophrenia outperform men on immediate and delayed verbal memory and that this advantage persists over and above the effects of age, illness, and hospitalization (Sota & Heinrichs, 2003). Similarly, recent evidence points to two distinct subtypes of the illness—one characterized by pervasive cognitive deficit, in which memory dysfunction contributes the largest effect size and another, comprising patients with relatively spared cognition (Hallmayer et al., 2005; Jablensky, 2006). Finally, as noted above, the wide variation in administration procedures in use with the RAVLT also makes it difficult to find appropriate normative data.

The purpose of this study, therefore, was to create normative data for the RAVLT for individuals with a confirmed diagnosis of schizophrenia or schizophrenia-spectrum disorder (SZ), together with matching data for healthy, community controls, which might be useful in a clinical setting. The data were derived from the Western Australian Family Study of Schizophrenia (WAFSS; Hallmayer et al., 2005). Owing to the expected influence of demographic factors (age and sex) and the general level of cognitive function (IQ), we first examined the influence of these moderating factors on RAVLT performance and then stratified the data accordingly.

Method

Participants

The participant data used in this study were drawn from the WAFSS. The study was approved by the Human Research Ethics Committee of the University of Western Australia and the North Metropolitan Area Health Service-Mental Health Human Research Ethics Committee, Perth, Western Australia. Written informed consent was obtained from all participants.

Individuals with schizophrenia or a schizophrenia spectrum disorder (SZ) were recruited from inpatient and outpatient services at the Graylands Hospital, Perth. Participants with either schizophrenia or a broader-spectrum disorder, such as schizoaffective disorder, were included since neuropsychological evidence suggests that there are no substantial differences in cognition between these groups (Bora, Yucel, & Pantelis, 2009). Healthy community control (HC) participants were recruited by random sampling from local telephone directories, Red Cross blood donors, and other community group volunteers. All WAFSS participants were aged between 18 and 65, were fluent in English, and had been assessed for premorbid intelligence using the National Adult Reading Test-Revised (Nelson & Willison, 1991). (Although years of education were also available, this measure was not included as a separate variable since it was significantly correlated with NART.) Within this database, 778 participants had been administered the RAVLT. We excluded data for all participants with a history of additional neurological disease (e.g., Parkinson's disease) or brain injury, and self-reported hearing impairment that might influence neuropsychological performance. In addition, healthy controls with a personal or family history of psychotic illness were also excluded. Following this screening process, RAVLT data for 492 individuals with schizophrenia and 260 healthy controls remained for analysis. Table 1 provides basic descriptive data for the two groups.

Table 1.

Demographic characteristics of the study samples

 SZ group (n = 492) HC group (n = 260) t-test/χ2 p-values 
Sex (number of men [%]) 373 (75.8) 159 (61.2) 17.66 <.001 
Age (mean [SD]) 35.6 (10.8) 39.2 (13.1) 3.80 <.001 
NART IQ (mean [SD]) 97.6 (10.3) 106.0 (7.8) 12.58 <.001 
Age at onset of disorder (mean [SD]) 22.9 (6.2) — — — 
Length of illness (yrs) (mean [SD]) 11.5 (9.2)    
Medication, CPZ equivalenta (mean [SD]) 785.5 (535) — — — 
 SZ group (n = 492) HC group (n = 260) t-test/χ2 p-values 
Sex (number of men [%]) 373 (75.8) 159 (61.2) 17.66 <.001 
Age (mean [SD]) 35.6 (10.8) 39.2 (13.1) 3.80 <.001 
NART IQ (mean [SD]) 97.6 (10.3) 106.0 (7.8) 12.58 <.001 
Age at onset of disorder (mean [SD]) 22.9 (6.2) — — — 
Length of illness (yrs) (mean [SD]) 11.5 (9.2)    
Medication, CPZ equivalenta (mean [SD]) 785.5 (535) — — — 

Notes: SZ = schizophrenia-spectrum disorder; HC = healthy community control; NART = National Adult Reading Test; CPZ = chlorpromazine.

aCPZ equivalent data were retrieved for 233 schizophrenia cases.

Diagnostic assessment was based on standardized interviews employing the diagnostic interview for psychosis (DIP; Castle et al., 2006) and scored using the OPCRIT diagnostic algorithm (McGuffin, Farmer, & Harvey, 1991). Video-recorded interviews and clinical charts were independently reviewed by two senior clinicians who assigned consensus research diagnoses. All patients met both ICD-10 and DSM-IV criteria for a lifetime diagnosis of schizophrenia (n = 398, ICD-10 codes: F20.0, F20.1, F20.2, F20.3, F20.5, F20.6, F20.8, and F20.9) or SZ (n = 94, ICD-10 codes: F21, F23.1, F23.2, F25, and F29). SZ participants were taking their usual medication at the time of testing. Data were available for 191 patients on atypical and 42 patients on typical medication. HCs were also screened for psychopathology using the DIP.

Procedure

All participants were tested individually by graduate research assistants with a degree in psychology. All testers received specific additional training in the administration and scoring of the RAVLT by a registered clinical psychologist (JCB). Standard WAFSS quality-control procedures also include double scoring of all psychological test record forms together with ongoing checks on the reliability of administration procedures.

In this project, List A (Lezak et al., 2004; Taylor, 1959) of the RAVLT was presented, comprising 15 words, always in the same order, which the tester read aloud to participants at a rate of one per second. Participant instructions were as follows:

I am going to read a list of words. Listen carefully, for when I stop you are to say back as many words as you can remember. It doesn't matter in what order you repeat them. Just try to remember as many as you can.

After Trial 1 was presented, the tester recorded each word, in the order recalled and then gave the following instructions:

Now, I'm going to read the same list again, and once again when I stop I want you to tell me as many words as you can remember, including words you said the first time. It doesn't matter in what order you say them. Just say as many words as you can remember whether or not you said them before.

This procedure was repeated again, yielding a total of three immediate recall trials, compared with the more common five-trial format. Furthermore, no distractor list was presented. This shorter administration procedure was adopted in order to minimize participant fatigue, since the entire WAFSS test battery required 3–4 h of testing. Following Trial 3 of the RAVLT, participants were engaged in other testing (usually the Continuous Performance Task, Identical Pairs version) for approximately 20 min. After this delay, participants were asked to recall as many words as they could from the original learning trials, in any order (delay recall).

Scoring for each trial (T1, T2, T3, and T-delay) was based on the number of words correctly recalled. A summary score of the total number of words correctly recalled, summed across Trials 1–3, was also calculated (T-total). In addition, the number of repetitions (R-total; number of times a correct word is repeated) and extra-list intrusions (I-total) was recorded (summed across the three learning trials). Finally, indices of encoding and forgetting were calculated, as follows: Encoding (E)—number of new words correctly recalled on each trial, summed over Trials 1–3 (max = 15); forgetting (F)—number of words recalled on Trial 3 minus number of words recalled at T-delay.

Statistical Analysis

A series of linear regression analyses was carried out to investigate the contribution of demographic variables (age and sex; with sex converted into a dummy variable), and general intelligence (NART) to performance on RAVLT measures. Group differences in verbal learning and memory were examined using a mixed design with one within-subjects factor (Trials 1–3 and T-delay) and a between-subjects factor (SZ vs. HC). The magnitude of effects was expressed in terms of generalized eta-squared (i.e., forumla; Bakeman, 2005). No outliers, with scores of 3 SD from the mean, were detected in the study sample. Chi-square and independent samples t-tests were applied to analyze the demographic data. All analyses were carried out using PASW 17.0.

Results

The SZ group represents a fairly representative cross-section of service users with severe mental disorder. Table 1 shows basic descriptive statistics for both groups. The SZ group was significantly younger than the HC group, comprised a higher percentage of men and had a lower estimated premorbid IQ. Of note, the mean estimated IQ of the HC group was similar to that reported for the NART standardization sample (mean = 107.4), though the standard deviation was smaller, suggesting that average IQ performance of the current sample was similar to—though less variable than—the general population. As expected, the total number of words recalled by SZ participants was significantly lower than that of the HC group (SZ T-total: mean = 20.2, SD = 5.9, T-delay: mean = 5.6, SD = 2.9; HC T-total: mean = 27.4, SD = 5.2, T-delay: mean = 8.9, SD = 2.8).

Contribution of Age, Sex, and Intelligence to RAVLT Performance

Multiple linear regression analysis was performed on each group (Tables 2 and 3). Both age and sex (i.e., being men) consistently affected key RAVLT performance measures in each group. In general, RAVLT scores were lower in older than younger adults and in men compared with women. In the SZ group, NART-IQ scores were also significantly related to RAVLT scores.

Table 2.

Multiple linear regression analyses: Contributions of age, sex, and intelligence on RAVLT scores in healthy controls

Score Variable β SE Standardized β t-test p-value r2 
T1 Constant 6.217 1.460  4.259 .000  
 Age −0.044 0.008 −0.327 −5.574 .000  
 Men −0.589 0.213 −0.162 −2.770 .006  
 NART IQ 0.014 0.013 0.060 1.023 .307 .133 
T2 Constant 9.546 1.716  5.562 .000  
 Age −0.061 0.009 −0.380 −6.593 .000  
 Men −0.669 0.250 −0.153 −2.674 .008  
 NART IQ 0.014 0.016 0.053 0.923 .357 .166 
T3 Constant 8.862 1.645  5.386 .000  
 Age −0.057 0.009 −0.364 −6.365 .000  
 Men −0.848 0.240 −0.201 −3.535 .000  
 NART IQ 0.031 0.015 0.117 2.055 .041 .179 
T-total Constant 24.624 4.124  5.970 .000  
 Age −0.162 0.022 −0.408 −7.255 .000  
 Men −2.106 0.601 −0.196 −3.503 .001  
 NART IQ 0.059 0.038 0.088 1.566 .119 .206 
I-total Constant 1.778 1.063  1.673 .096  
 Age 0.004 0.006 0.042 0.669 .504  
 Men 0.142 0.155 0.057 0.914 .362  
 NART IQ −0.008 0.010 −0.050 −0.794 .428 .007 
R-total Constant 3.464 2.054  1.687 .093  
 Age −0.022 0.011 −0.120 −1.934 .054  
 Men −0.509 0.299 −0.105 −1.701 .090  
 NART IQ −0.010 0.019 −0.033 −0.533 .594 .027 
T-delay Constant 8.207 2.285  3.592 .000  
 Age −0.073 0.012 −0.343 −5.918 .000  
 Men −1.182 0.333 −0.205 −3.551 .000  
 NART IQ 0.018 0.021 0.050 0.868 .386 .158 
Constant 12.919 1.796  7.192 .000  
 Age −0.053 0.010 −0.320 −5.398 .000  
 Men −0.814 0.262 −0.184 −3.113 .002  
 NART IQ 0.009 0.016 0.032 0.543 .587 .133 
Constant 0.654 1.627  0.402 .688  
 Age 0.017 0.009 0.117 1.874 .062  
 Men 0.335 0.237 0.088 1.411 .159  
 NART IQ 0.013 0.015 0.054 0.860 .391 .025 
Score Variable β SE Standardized β t-test p-value r2 
T1 Constant 6.217 1.460  4.259 .000  
 Age −0.044 0.008 −0.327 −5.574 .000  
 Men −0.589 0.213 −0.162 −2.770 .006  
 NART IQ 0.014 0.013 0.060 1.023 .307 .133 
T2 Constant 9.546 1.716  5.562 .000  
 Age −0.061 0.009 −0.380 −6.593 .000  
 Men −0.669 0.250 −0.153 −2.674 .008  
 NART IQ 0.014 0.016 0.053 0.923 .357 .166 
T3 Constant 8.862 1.645  5.386 .000  
 Age −0.057 0.009 −0.364 −6.365 .000  
 Men −0.848 0.240 −0.201 −3.535 .000  
 NART IQ 0.031 0.015 0.117 2.055 .041 .179 
T-total Constant 24.624 4.124  5.970 .000  
 Age −0.162 0.022 −0.408 −7.255 .000  
 Men −2.106 0.601 −0.196 −3.503 .001  
 NART IQ 0.059 0.038 0.088 1.566 .119 .206 
I-total Constant 1.778 1.063  1.673 .096  
 Age 0.004 0.006 0.042 0.669 .504  
 Men 0.142 0.155 0.057 0.914 .362  
 NART IQ −0.008 0.010 −0.050 −0.794 .428 .007 
R-total Constant 3.464 2.054  1.687 .093  
 Age −0.022 0.011 −0.120 −1.934 .054  
 Men −0.509 0.299 −0.105 −1.701 .090  
 NART IQ −0.010 0.019 −0.033 −0.533 .594 .027 
T-delay Constant 8.207 2.285  3.592 .000  
 Age −0.073 0.012 −0.343 −5.918 .000  
 Men −1.182 0.333 −0.205 −3.551 .000  
 NART IQ 0.018 0.021 0.050 0.868 .386 .158 
Constant 12.919 1.796  7.192 .000  
 Age −0.053 0.010 −0.320 −5.398 .000  
 Men −0.814 0.262 −0.184 −3.113 .002  
 NART IQ 0.009 0.016 0.032 0.543 .587 .133 
Constant 0.654 1.627  0.402 .688  
 Age 0.017 0.009 0.117 1.874 .062  
 Men 0.335 0.237 0.088 1.411 .159  
 NART IQ 0.013 0.015 0.054 0.860 .391 .025 

Notes: RAVLT = Rey's Auditory Verbal Learning Test; T1 = RAVLT Trial 1; T2 = RAVLT Trial 2; T3 = RAVLT Trial 3; T-total = RAVLT total immediate recall (three trials summed); I-total = RAVLT total immediate intrusions (three trials summed); R-total = RAVLT total immediate repeats (three trials summed); T-delay = RAVLT delayed recall; E = RAVLT encoding; F = RAVLT forgetting (T3-DR); NART = National Adult Reading Test.

Table 3.

Multiple linear regression analyses: Contributions of age, sex, and intelligence on RAVLT scores in schizophrenia participants

Score Variable β SE Standardized β t-test p-value r2 
T1 Constant 0.351 0.805  0.435 .663  
 Age −0.040 0.008 −0.229 −5.095 .000  
 Men −0.378 0.192 −0.086 −1.963 .050  
 NART IQ 0.057 0.008 0.312 7.114 .000 .118 
T2 Constant 0.287 0.960  0.298 .766  
 Age −0.054 0.009 −0.251 −5.728 .000  
 Men −0.446 0.230 −0.083 −1.944 .052  
 NART IQ 0.083 0.010 0.369 8.611 .000 .158 
T3 Constant 0.105 1.084  0.097 .923  
 Age −0.060 0.011 −0.244 −5.613 .000  
 Men −0.560 0.259 −0.092 −2.160 .031  
 NART IQ 0.099 0.011 0.389 9.147 .000 .170 
T-total Constant 0.742 2.413  0.308 .759  
 Age −0.154 0.024 −0.278 −6.500 .000  
 Men −1.38 0.577 −0.100 −2.398 .017  
 NART IQ 0.239 0.024 0.414 9.908 .000 .198 
I-total Constant 4.428 1.118  3.961 .000  
 Age −0.014 0.011 −0.060 −1.272 .204  
 Men −0.013 0.267 −0.002 −0.048 .962  
 NART IQ −0.025 0.011 −0.102 −2.211 .028 .017 
R-total Constant 2.604 1.089  2.390 .017  
 Age −0.020 0.011 −0.087 −1.832 .068  
 Men −0.123 0.260 −0.022 −0.473 .636  
 NART IQ −0.002 0.011 −0.009 −0.191 .848 .008 
T-delay Constant −1.136 1.259  −0.903 .367  
 Age −0.052 0.012 −0.191 −4.206 .000  
 Men −1.03 0.301 −0.151 −3.414 .001  
 NART IQ 0.076 0.013 0.267 6.012 .000 .097 
Constant 3.260 1.122  2.905 .004  
 Age −0.058 0.011 −0.231 −5.182 .000  
 Men −0.439 0.269 −0.071 −1.632 .103  
 NART IQ 0.090 0.011 0.350 8.018 .000 .141 
Constant 1.241 0.926  1.341 .181  
 Age −0.008 0.009 −0.040 −0.853 .394  
 Men 0.468 0.221 0.097 2.113 .035  
 NART IQ 0.023 0.009 0.117 2.534 .012 .024 
Score Variable β SE Standardized β t-test p-value r2 
T1 Constant 0.351 0.805  0.435 .663  
 Age −0.040 0.008 −0.229 −5.095 .000  
 Men −0.378 0.192 −0.086 −1.963 .050  
 NART IQ 0.057 0.008 0.312 7.114 .000 .118 
T2 Constant 0.287 0.960  0.298 .766  
 Age −0.054 0.009 −0.251 −5.728 .000  
 Men −0.446 0.230 −0.083 −1.944 .052  
 NART IQ 0.083 0.010 0.369 8.611 .000 .158 
T3 Constant 0.105 1.084  0.097 .923  
 Age −0.060 0.011 −0.244 −5.613 .000  
 Men −0.560 0.259 −0.092 −2.160 .031  
 NART IQ 0.099 0.011 0.389 9.147 .000 .170 
T-total Constant 0.742 2.413  0.308 .759  
 Age −0.154 0.024 −0.278 −6.500 .000  
 Men −1.38 0.577 −0.100 −2.398 .017  
 NART IQ 0.239 0.024 0.414 9.908 .000 .198 
I-total Constant 4.428 1.118  3.961 .000  
 Age −0.014 0.011 −0.060 −1.272 .204  
 Men −0.013 0.267 −0.002 −0.048 .962  
 NART IQ −0.025 0.011 −0.102 −2.211 .028 .017 
R-total Constant 2.604 1.089  2.390 .017  
 Age −0.020 0.011 −0.087 −1.832 .068  
 Men −0.123 0.260 −0.022 −0.473 .636  
 NART IQ −0.002 0.011 −0.009 −0.191 .848 .008 
T-delay Constant −1.136 1.259  −0.903 .367  
 Age −0.052 0.012 −0.191 −4.206 .000  
 Men −1.03 0.301 −0.151 −3.414 .001  
 NART IQ 0.076 0.013 0.267 6.012 .000 .097 
Constant 3.260 1.122  2.905 .004  
 Age −0.058 0.011 −0.231 −5.182 .000  
 Men −0.439 0.269 −0.071 −1.632 .103  
 NART IQ 0.090 0.011 0.350 8.018 .000 .141 
Constant 1.241 0.926  1.341 .181  
 Age −0.008 0.009 −0.040 −0.853 .394  
 Men 0.468 0.221 0.097 2.113 .035  
 NART IQ 0.023 0.009 0.117 2.534 .012 .024 

Notes: RAVLT = Rey's Auditory Verbal Learning Test; T1 = RAVLT Trial 1; T2 = RAVLT Trial 2; T3 = RAVLT Trial 3; T-total = RAVLT total immediate recall (three trials summed); I-total = RAVLT total immediate intrusions (three trials summed); R-total = RAVLT total immediate repeats (three trials summed); T-delay = RAVLT delayed recall; E = RAVLT encoding; F = RAVLT forgetting (T3-DR); NART = National Adult Reading Test.

Normative Data Stratified

As a result of the regression analyses, normative tables were derived for each group, comprising means and standard deviations, stratified by age and sex (Tables 4 and 5). The entire sample was divided into three subgroups of younger (18–30 years), mature (31–42 years), and older (43–65 years) adults. This stratification was guided in part by previous recommendations that such subgroups should aim for cell size of at least n = 50 subjects (Mitrushina et al., 2005, p. 39) to maximize stability in the data. The age cutoffs were also chosen to be maximally beneficial in clinical settings since the younger age group captures the peak age of onset of schizophrenia. For the SZ group only, the normative table also includes subgroups determined by NART-IQ scores (Table 5), since IQ significantly influenced RAVLT scores. The lower IQ subgroup comprised SZ participants with NART scores 1 SD below the mean of the NART standardization sample (i.e., IQ < 90; mean IQ = 83.8, SD = 4.6), whereas SZ participants with IQ ≥ 90 formed the higher IQ subgroup (mean IQ = 102.1, SD = 7.1).

Table 4.

Normative data for healthy controls stratified by age and sex

 n T1 T2 T3 T-total T-delay 
Men (years) 
 18–30 51 7.2 (1.7) 10.3 (2.1) 11.7 (1.7) 29.2 (4.6) 9.8 (2.4) 13.4 (2.1) 2.0 (1.7) 
 31–42 38 6.9 (1.8) 10.0 (1.4) 11.3 (1.9) 28.1 (4.2) 8.9 (2.6) 13.2 (2.1) 2.6 (2.0) 
 43+ 70 5.8 (1.6) 8.3 (1.9) 9.8 (2.2) 23.9 (5.2) 7.2 (2.7) 11.6 (2.6) 2.6 (2.0) 
Women (years) 
 18–30 30 7.7 (1.8) 10.6 (1.9) 12.1 (1.4) 30.5 (4.4) 10.5 (2.7) 14.0 (1.2) 1.6 (1.9) 
 31–42 23 7.5 (1.7) 10.6 (2.4) 12.0 (2.0) 30.0 (5.3) 9.5 (2.6) 13.5 (1.7) 2.5 (1.7) 
 43+ 48 6.5 (1.5) 9.3 (1.9) 11.1 (1.7) 27.0 (4.2) 9.2 (2.5) 13.0 (1.5) 2.0 (1.7) 
 n T1 T2 T3 T-total T-delay 
Men (years) 
 18–30 51 7.2 (1.7) 10.3 (2.1) 11.7 (1.7) 29.2 (4.6) 9.8 (2.4) 13.4 (2.1) 2.0 (1.7) 
 31–42 38 6.9 (1.8) 10.0 (1.4) 11.3 (1.9) 28.1 (4.2) 8.9 (2.6) 13.2 (2.1) 2.6 (2.0) 
 43+ 70 5.8 (1.6) 8.3 (1.9) 9.8 (2.2) 23.9 (5.2) 7.2 (2.7) 11.6 (2.6) 2.6 (2.0) 
Women (years) 
 18–30 30 7.7 (1.8) 10.6 (1.9) 12.1 (1.4) 30.5 (4.4) 10.5 (2.7) 14.0 (1.2) 1.6 (1.9) 
 31–42 23 7.5 (1.7) 10.6 (2.4) 12.0 (2.0) 30.0 (5.3) 9.5 (2.6) 13.5 (1.7) 2.5 (1.7) 
 43+ 48 6.5 (1.5) 9.3 (1.9) 11.1 (1.7) 27.0 (4.2) 9.2 (2.5) 13.0 (1.5) 2.0 (1.7) 

Notes: T1 = RAVLT Trial 1; T2 = RAVLT Trial 2; T3 = RAVLT Trial 3; T-total = RAVLT total immediate recall (three trials summed); T-delay = RAVLT delayed recall; E = RAVLT encoding; F = RAVLT forgetting.

Table 5.

Normative data for schizophrenia participants stratified by age, sex, and premorbid IQ

 n T1 T2 T3 T-total T-delay 
Men 
 18–30 years 
  Lower IQ 48 4.1 (1.6) 5.8 (1.9) 7.4 (2.3) 17.4 (4.9) 4.7 (2.5) 9.6 (2.3) 2.8 (1.9) 
  Higher IQ 113 5.4 (1.9) 7.9 (2.4) 9.2 (2.7) 22.6 (6.1) 6.4 (3.0) 11.4 (2.4) 2.9 (2.1) 
  Total 161 5.0 (1.9) 7.3 (2.4) 8.7 (2.7) 21.0 (6.2) 5.9 (3.0) 10.9 (2.5) 2.8 (2.0) 
 31–42 years 
  Lower IQ 38 4.6 (2.0) 6.2 (1.5) 7.4 (1.7) 18.3 (4.4) 4.7 (2.3) 9.7 (2.3) 2.8 (2.2) 
  Higher IQ 97 5.1 (1.7) 6.9 (2.0) 8.3 (2.5) 20.2 (5.2) 5.5 (2.9) 10.5 (2.6) 2.8 (2.0) 
  Total 135 5.0 (1.8) 6.7 (1.9) 8.0 (2.4) 19.7 (5.1) 5.2 (2.7) 10.2 (2.6) 2.8 (2.0) 
 43–65 years 
  Lower IQ 14 3.9 (2.1) 6.0 (2.0) 7.2 (2.4) 17.1 (5.4) 5.1 (2.6) 9.9 (2.7) 2.1 (2.2) 
  Higher IQ 63 4.7 (1.6) 6.6 (2.1) 8.0 (2.4) 19.3 (5.2) 5.0 (2.4) 10.0 (2.5) 2.9 (1.9) 
  Total 77 4.6 (1.7) 6.5 (2.1) 7.8 (2.4) 18.9 (5.2) 5.0 (2.4) 10.0 (2.5) 2.8 (2.0) 
Women 
 18–30 years 
  Lower IQ 6.0 (3.2) 6.8 (2.9) 7.1 (3.1) 19.9 (8.5) 4.8 (4.0) 9.4 (2.8) 2.4 (2.7) 
  Higher IQ 26 5.7 (2.1) 8.2 (2.8) 9.8 (2.9) 23.7 (7.1) 7.6 (3.0) 12.0 (2.4) 2.2 (1.9) 
  Total 34 5.8 (2.4) 7.8 (2.9) 9.2 (3.1) 22.8 (7.5) 7.0 (3.4) 11.5 (2.6) 2.2 (2.1) 
 31–42 years 
  Lower IQ 4.4 (1.0) 5.9 (2.0) 7.2 (1.3) 17.4 (3.8) 4.7 (2.1) 10.7 (1.6) 2.4 (1.3) 
  Higher IQ 30 5.6 (1.8) 7.5 (2.5) 9.1 (2.6) 22.1 (5.9) 6.5 (3.1) 10.9 (2.6) 2.5 (2.0) 
  Total 37 5.4 (1.8) 7.2 (2.4) 8.7 (2.5) 21.2 (5.8) 6.2 (3.0) 10.7 (2.5) 2.5 (1.9) 
 43–65 years 
  Lower IQ 11 3.1 (1.3) 5.2 (2.0) 5.6 (2.1) 13.9 (4.4) 3.8 (2.0) 7.5 (2.6) 1.8 (2.0) 
  Higher IQ 37 4.8 (1.8) 6.9 (2.5) 8.5 (2.4) 20.1 (5.7) 6.3 (3.1) 10.6 (3.0) 2.3 (2.5) 
  Total 48 4.4 (1.8) 6.5 (2.4) 7.9 (2.6) 18.7 (6.0) 5.7 (3.0) 9.9 (3.2) 2.2 (2.4) 
 n T1 T2 T3 T-total T-delay 
Men 
 18–30 years 
  Lower IQ 48 4.1 (1.6) 5.8 (1.9) 7.4 (2.3) 17.4 (4.9) 4.7 (2.5) 9.6 (2.3) 2.8 (1.9) 
  Higher IQ 113 5.4 (1.9) 7.9 (2.4) 9.2 (2.7) 22.6 (6.1) 6.4 (3.0) 11.4 (2.4) 2.9 (2.1) 
  Total 161 5.0 (1.9) 7.3 (2.4) 8.7 (2.7) 21.0 (6.2) 5.9 (3.0) 10.9 (2.5) 2.8 (2.0) 
 31–42 years 
  Lower IQ 38 4.6 (2.0) 6.2 (1.5) 7.4 (1.7) 18.3 (4.4) 4.7 (2.3) 9.7 (2.3) 2.8 (2.2) 
  Higher IQ 97 5.1 (1.7) 6.9 (2.0) 8.3 (2.5) 20.2 (5.2) 5.5 (2.9) 10.5 (2.6) 2.8 (2.0) 
  Total 135 5.0 (1.8) 6.7 (1.9) 8.0 (2.4) 19.7 (5.1) 5.2 (2.7) 10.2 (2.6) 2.8 (2.0) 
 43–65 years 
  Lower IQ 14 3.9 (2.1) 6.0 (2.0) 7.2 (2.4) 17.1 (5.4) 5.1 (2.6) 9.9 (2.7) 2.1 (2.2) 
  Higher IQ 63 4.7 (1.6) 6.6 (2.1) 8.0 (2.4) 19.3 (5.2) 5.0 (2.4) 10.0 (2.5) 2.9 (1.9) 
  Total 77 4.6 (1.7) 6.5 (2.1) 7.8 (2.4) 18.9 (5.2) 5.0 (2.4) 10.0 (2.5) 2.8 (2.0) 
Women 
 18–30 years 
  Lower IQ 6.0 (3.2) 6.8 (2.9) 7.1 (3.1) 19.9 (8.5) 4.8 (4.0) 9.4 (2.8) 2.4 (2.7) 
  Higher IQ 26 5.7 (2.1) 8.2 (2.8) 9.8 (2.9) 23.7 (7.1) 7.6 (3.0) 12.0 (2.4) 2.2 (1.9) 
  Total 34 5.8 (2.4) 7.8 (2.9) 9.2 (3.1) 22.8 (7.5) 7.0 (3.4) 11.5 (2.6) 2.2 (2.1) 
 31–42 years 
  Lower IQ 4.4 (1.0) 5.9 (2.0) 7.2 (1.3) 17.4 (3.8) 4.7 (2.1) 10.7 (1.6) 2.4 (1.3) 
  Higher IQ 30 5.6 (1.8) 7.5 (2.5) 9.1 (2.6) 22.1 (5.9) 6.5 (3.1) 10.9 (2.6) 2.5 (2.0) 
  Total 37 5.4 (1.8) 7.2 (2.4) 8.7 (2.5) 21.2 (5.8) 6.2 (3.0) 10.7 (2.5) 2.5 (1.9) 
 43–65 years 
  Lower IQ 11 3.1 (1.3) 5.2 (2.0) 5.6 (2.1) 13.9 (4.4) 3.8 (2.0) 7.5 (2.6) 1.8 (2.0) 
  Higher IQ 37 4.8 (1.8) 6.9 (2.5) 8.5 (2.4) 20.1 (5.7) 6.3 (3.1) 10.6 (3.0) 2.3 (2.5) 
  Total 48 4.4 (1.8) 6.5 (2.4) 7.9 (2.6) 18.7 (6.0) 5.7 (3.0) 9.9 (3.2) 2.2 (2.4) 

Notes: T1 = RAVLT Trial 1; T2 = RAVLT Trial 2; T3 = RAVLT Trial 3; T-total = RAVLT total immediate recall (three trials summed); T-delay = RAVLT delayed recall; E = RAVLT encoding; F = RAVLT forgetting; low IQ = NART IQ ≤ 90; higher IQ = NART IQ > 91.

Correlation Analysis

Correlations between duration of illness, and age of illness onset, and memory scores in the SZ group were also examined. Pearson's r and partial correlation coefficients were calculated; after adjustment for age, duration of illness showed no independent effect on RAVLT measures (all p values > .05). Similarly, there was no significant association between age of illness onset and RAVLT performance (all p values > .05). Small but significant negative correlations were noted, however, between the SZ participants level of medication (CPZ equivalents) and all RAVLT scores except forgetting (T1, r = −.20; T2, r = −.26; T3, r = −.18; T-total, r = −.25; T-delay, r = −.21; E, r = −.23; F, r = .05).

Between-Groups Comparisons: IQ, Age, and Sex-Matched Subsamples

Since the need for lower IQ and higher IQ SZ subsamples in the normative tables is clear from the results above, a follow-up analysis was conducted to examine RAVLT performance between a subset of higher IQ SZ participants and HC participants matched on IQ, age, and sex in order to determine whether a memory deficit is specifically linked to schizophrenia. Participants’ memory performance was analyzed using a 2 (matched subsamples—Group: higher IQ SZ, n = 238 and HC, n = 235) × 4(Trial: T1, T2, T3, T-delay) repeated-measures analysis of variance (ANOVA) design. Subsamples were closely matched on age (p = .929) and premorbid IQ (p = .844) and did not differ significantly in sex distribution (p = .054). The main effect of Trial was statistically significant, F(3,1413) = 648.99, p < .001, forumla, indicating a marked increase in verbal learning during the three immediate recall trials followed by delayed recall. As expected, the main effect of Group was also significant, F(1,471) = 106.92, p <.001, forumla, confirming that RAVLT performance of the higher IQ SZ participants was worse than that of HC group on all trials. Furthermore, the interaction between Group and Trial was also significant, F(3,1413) = 13.28, p < .001, forumla, indicating markedly different learning curves in the matched subsamples of HC and SZ participants (Figure 1). Post hoc pairwise comparisons between subsamples at each trial showed that the mean difference in RAVLT recall increased from 1.4 words on T1 to a 2.6-word difference on T-delay. Additional repeated-measures ANOVA, with within-subjects factor (encoding and forgetting) and between-subjects factor (matched subsamples Group: higher SZ and HC) was used to further clarify the nature of the performance differences observed. Prior to this analysis, all individual scores were transformed into standardized z-scores to avoid obvious differences in absolute values between encoding and forgetting. The results showed a significant main effect of the within-subject factor, F(1,461) = 11.53, p =.001, forumla, and an almost significant main effect of Group F(1,461) = 3.62, p = .058, forumla. Importantly, the interaction was highly significant, F(1,461) = 64.32, p < .000, forumla. Post hoc comparisons revealed a striking between-group difference in encoding between the HC (M = 0.58, SD = 0.83) and higher IQ SZ (M = −0.04, SD = 0.97, t(461) = 7.41, p < .001) matched subsamples with a large effect size (forumla). The mean difference in forgetting was also statistically significant, HC mean = −0.14 (SD = 0.92) versus SZ = 0.23 (SD = 1.03), t(461) = 4.11, p = .001, with notably smaller effect size (forumla).

Fig. 1.

Mean RAVLT scores across four trials (T1, T2, T3, and T-delay) for higher IQ SZ and HC subsamples (matched on IQ, age, and sex) and the lower IQ SZ subgroup. Data for the lower IQ SZ subgroup are provided here for reference purposes only and were not included in the between-groups analysis.

Fig. 1.

Mean RAVLT scores across four trials (T1, T2, T3, and T-delay) for higher IQ SZ and HC subsamples (matched on IQ, age, and sex) and the lower IQ SZ subgroup. Data for the lower IQ SZ subgroup are provided here for reference purposes only and were not included in the between-groups analysis.

Discussion

The aim of this research was to provide norms for the assessment of episodic memory—based on RAVLT performance—in individuals with schizophrenia or SZ, for use in clinical practice. Norms for healthy controls recruited from the same population as individuals with schizophrenia are also presented. The abbreviated administration format of the RAVLT used in the WAFSS—involving only three immediate recall trials, no interference list, and a delay period filled with other cognitive tasks—is likely to be maximally useful in clinical settings where limited testing time may be available. SZ participants exhibited a robust impairment in simple, summary measures of immediate (T-total) and delayed (T-delay) episodic memory, with verbal recall approximately 1.5 SD below that of healthy controls.

As expected, based on previous reports, both age and sex influenced RAVLT performance (Geffen et al., 1990; Schmidt, 1996; Strauss et al., 2006). This pattern was evident in both the SZ and the HC groups. Women tended to recall more words than men, (Sota & Heinrichs, 2003), whereas younger adults tended to outperform older adults. General intelligence also significantly influenced memory performance in the schizophrenia participants only; individuals with lower premorbid IQ recalled fewer words than individuals with IQ's in the normal or above the range. Consequently, normative tables were generated for RAVLT performance for both SZ and HC groups, stratified accordingly. Age and IQ cutoffs were selected to maximize clinical utility and relevance while attempting to maintain a reasonable cell size (∼50) where possible, to increase the likely stability in the data. In particular, the younger adult (18–30 years) subgroup captures the peak age of onset of schizophrenia, whereas lower (IQ < 90) and higher (IQ ≥ 90) IQ subgroups mirror previous evidence of cognitively compromised and (relatively) cognitively spared schizophrenia subtypes, respectively (Hallmayer et al., 2005; Weickert et al., 2000). Importantly, however, between-groups comparisons also confirmed a significant deficit in RAVLT learning and recall in schizophrenia compared with healthy controls, even when subsamples were matched on general intelligence, age, and sex, suggesting that poor verbal memory is specifically linked to schizophrenia. These findings underscore the need for clinicians' to assess specific cognitive functions in individuals with schizophrenia, especially verbal memory, even when general intelligence appears to be relatively well preserved. Further analysis suggested that the observed impairment in schizophrenia is related more to difficulties in encoding rather than forgetting, a finding consistent with previous research (Danion et al., 2007).

Several limitations should be noted with the current research. First, the schizophrenia group was mostly medicated, aged in the mid-1930s, and had lived with the illness on average for 10 years. Consequently, the current norms may not accurately reflect performance for markedly younger or older individuals, such as those experiencing their first episode of illness during late adolescence. However, it must be noted that neither younger age of illness onset nor longer duration of illness was significantly associated with poorer recall. Second, many participants with schizophrenia were receiving their usual medication (typicals and atypicals) at the time of testing, which means that the norms presented may not accurately reflect performance of un-medicated individuals. In addition, while the current data show a small, but significant negative correlation between CPZ equivalents and RAVLT performance, it is important to note that evidence concerning the influence of antipsychotic medications on cognitive functioning is mixed and inconclusive. Finally, the number of female participants with schizophrenia was relatively small, and so when stratified by age and IQ the resulting normative subgroups formed are relatively small (n < 50 per group). While this is a potential limitation of the current study, the sample size of women, and men, subgroups is significantly larger than any previously reported normative data for RAVLT. Furthermore, in view of previous evidence of better verbal memory in women with schizophrenia (Sota & Heinrichs, 2003), presenting norms grouped by gender may improve the sensitivity of detecting deficits in either group.

Funding

This work was supported by the National Health and Medical Research Council, Australia (404046, 513874, CIA Prof. A. Jablensky); and infrastructure contributions from the North Metropolitan Area Health Services, Department of Health, Western Australia.

Conflict of Interest

None declared.

Acknowledgements

We thank all the participants in the study for the generous contribution of their time together with the staff of the Centre for Clinical Research in Neuropsychiatry (CCRN), who assisted with participant recruitment, including Alan Bland and Paul Connolly. Thanks also to Prof. Assen Jablensky for his leading role in the Western Australian Family Study of Schizophrenia and for helpful comments on this manuscript.

References

Achim
A. M.
Lepage
M.
Episodic memory-related activation in schizophrenia: Meta-analysis
British Journal of Psychiatry
 , 
2005
, vol. 
187
 
6
(pg. 
500
-
509
)
Bakeman
R.
Recommended effect size statistics for repeated measures designs
Behavior Research Methods
 , 
2005
, vol. 
37
 
3
(pg. 
379
-
384
)
Bolla-Wilson
K.
Bleecker
M. L.
Influence of verbal intelligence, sex, age, and education on the Rey auditory verbal learning test
Developmental Neuropsychology
 , 
1986
, vol. 
2
 (pg. 
203
-
211
)
Bora
E.
Yucel
M.
Pantelis
C.
Cognitive functioning in schizophrenia, schizoaffective disorder, and affective psychosis: Meta-analytic study
British Journal of Psychiatry
 , 
2009
, vol. 
195
 
6
(pg. 
475
-
482
)
Boyer
P.
Phillips
J. L.
Rousseau
F. L.
Ilivitsky
S.
Hippocampal abnormalities and memory deficits: New evidence of a strong pathophysiological link in schizophrenia
Brain Research Reviews
 , 
2007
, vol. 
54
 (pg. 
92
-
112
)
Castle
D. J.
Jablensky
A.
McGrath
J. J.
Carr
V.
Morgan
V.
Waterreus
A.
, et al.  . 
The diagnostic interview for psychosis (DIP): Development, reliability and applications
Psychological Medicine
 , 
2006
, vol. 
36
 (pg. 
69
-
80
)
Danion
J-M.
Huron
C.
Vidailhet
P.
Berna
F.
Functional mechanisms of episodic memory impairment in schizophrenia
The Canadian Journal of Psychiatry
 , 
2007
, vol. 
52
 (pg. 
693
-
701
)
Geffen
G.
Moar
K. J.
O'Hanlon
A. P.
Clark
C. R.
Geffen
L. B.
Performance measures of 16–86 year old males and females on the Auditory Verbal Learning Test
The Clinical Neuropsychologist
 , 
1990
, vol. 
4
 (pg. 
45
-
63
)
Green
M. F.
What are the functional consequences of neurocognitive deficits in schizophrenia?
American Journal of Psychiatry
 , 
1996
, vol. 
153
 
3
(pg. 
321
-
330
)
Green
M. F.
Kern
R. S.
Braff
D. L.
Minz
J.
Neurocognitive deficits and functional outcome in schizophrenia: Are we measuring the ‘right stuff’?
Schizophrenia Bulletin
 , 
2000
, vol. 
26
 
1
(pg. 
119
-
136
)
Green
M. F.
Nuechterlein
K. H.
Gold
J. M.
Barch
D. M.
Cohen
J.
Essock
S.
, et al.  . 
Approaching a consensus cognitive battery for clinical trials in schizophrenia: The NIMH MATRICS conference to select cognitive domains and test criteria
Biological Psychiatry
 , 
2004
, vol. 
56
 
5
(pg. 
301
-
307
)
Hallmayer
J. F.
Kalaydjieva
L.
Badcock
J.
Dragović
M.
Howell
S.
Michie
P. T.
, et al.  . 
Genetic evidence for a distinct subtype of schizophrenia characterized by pervasive cognitive deficit
American Journal of Human Genetics
 , 
2005
, vol. 
77
 (pg. 
468
-
476
)
Jablensky
A.
Subtyping schizophrenia: Implications for genetic research
Molecular Psychiatry
 , 
2006
, vol. 
11
 (pg. 
815
-
836
)
Lezak
M. D.
Howieson
D. B.
Loring
D. W.
Neuropsychological assessment
 , 
2004
4th ed.
New York
Oxford University Press
McGuffin
P.
Farmer
A.
Harvey
I.
A polydiagnostic application of operational criteria in studies of psychotic illness. Development and reliability of the OPCRIT system
Archives of General Psychiatry
 , 
1991
, vol. 
48
 (pg. 
764
-
770
)
Mitrushina
M. N.
Boone
K. B.
Razani
J.
D'Elia
L.
Handbook of normative data for neuropsychological assessment
 , 
2005
2nd ed.
New York
Oxford University Press
Mungas
D.
Differential clinical sensitivity of specific parameters of Rey Auditory-Verbal Learning Test
Journal of Consulting and Clinical Psychology
 , 
1983
, vol. 
51
 (pg. 
848
-
855
)
Nelson
H. E.
Willison
J. R.
The Revised National Adult Reading Test—test manual
 , 
1991
Windsor, UK
NFER-Nelson
Query
W. T.
Megran
J.
Age-related norms for AVLT in a male patient population
Journal of Clinical Psychology
 , 
1983
, vol. 
39
 (pg. 
136
-
138
)
Rey
A.
L'examen psychologique dans les cas d'encéphalopathie traumatique
Archives de Psychologie
 , 
1941
, vol. 
28
 pg. 
21
 
Schmidt
M.
Rey Auditory and Verbal Learning Test: A handbook
 , 
1996
Los Angeles
Western Psychological Services
Schoenberg
M. R.
Dawson
K. A.
Duff
K.
Patton
D.
Scott
J. G.
Adams
R. L.
Test performance and classification statistics for the Rey auditory verbal learning test in selected clinical samples
Archives of Clinical Neuropsychology
 , 
2006
, vol. 
21
 (pg. 
693
-
703
)
Sota
T. L.
Heinrichs
R. W.
Sex differences in verbal memory in schizophrenia patients treated with ‘typical’ neuroleptics
Schizophrenia Research
 , 
2003
, vol. 
62
 (pg. 
175
-
182
)
Strauss
E.
Sherman
E. M. S.
Spreen
O.
A compendium of neuropsychological tests: Administration, norms and commentary
 , 
2006
3rd ed.
New York
Oxford University Press
Taylor
E. M.
Psychological appraisal of children with cerebral deficits
 , 
1959
Cambridge, MA
Harvard University Press
Toulopoulou
T.
Murray
R. M.
Verbal memory deficit in patients with schizophrenia: An important future target for treatment
Expert Reviews in Neurotherapeutics
 , 
2004
, vol. 
4
 
1
(pg. 
43
-
52
)
Van der Elst
W.
Van Boxtel
M. P. J.
Van Breukelen
G. J. P.
Jolles
J.
Rey's verbal learning test: Normative data for 1855 healthy participants aged 24–81 years and the influence of age, sex, education and mode of presentation
Journal of the International Neuropsychological Society
 , 
2005
, vol. 
11
 (pg. 
290
-
302
)
Weickert
T.
Goldberg
T.
Gold
J.
Bigelow
L. B.
Egan
M. F.
Weinberger
D. R.
Cognitive impairment in patients with schizophrenia displaying preserved and compromised intellect
Archives of General Psychiatry
 , 
2000
, vol. 
57
 (pg. 
907
-
913
)