Abstract

This paper reports two studies which investigated the effect of interference on delayed recall scores of the WMS-III and other commonly used memory measures. In Study 1, participants completed the immediate and delayed components of the WMS-III, with or without the introduction of conceptually similar memory tasks between the recall trials. In Study 2, this order of administration was reversed, with the WMS-III subtests used as the interference items. The results indicated that the introduction of interference items during the delay negatively affected delayed recall performance on almost all sub-tests. In addition, equal effects of proactive and retroactive interference were demonstrated. These findings raise concerns regarding the standardization process for memory tasks and highlight the need to consider interference effects in clinical practice, and stand as a caution in the use of memory-related materials during the delay interval in memory testing.

Introduction

One of the most widely supported accounts of forgetting contends that memory loss occurs due to interference from competing information. The proactive interference (PI) account (e.g., Underwood, 1950, 1957) contends that information presented before the target material is acquired causes greater rates of forgetting after a delay. By contrast, retroactive interference (RI) theory suggests that information presented after the target material negatively affects recall (McGeoch, 1932). Each of these theories has received considerable empirical support and both are now widely accepted as the principal factors that induce forgetting (Anderson, Bjork, & Bjork, 1994; Anderson & Neely, 1996; Bauml, 1996; Bower, Thompson-Schill, & Tulving, 1994; Shimamura, Jurica, Mangels, Gershberg, & Knight, 1995).

Although these theories have been widely accepted, little consideration has been given to the affect that interference may have on the standardization of commonly used memory tasks. This issue particularly applies in the development of test batteries, such as the Wechsler Memory Scale–Third Edition (WMS-III; Wechsler, 1997). For example, the published and standardized versions of the WMS-III were different from each other, with the main difference being that the published version contained significantly fewer subtests (Zhu & Tulsky, 2000) than did the standardization version. Thus during the standardization process, four additional tasks (i.e., Dot Locations, Verbal Fluency, Design Location, and Picture Naming) were included in addition to those tasks which now constitute the published version of the Scale. Using theories of PI and RI, one wonders whether this inclusion may have acted to increase the effect of interference, possibly culminating in a negative affect on recall and raising concerns about the applicability of the normative basis of the scale. These concerns particularly relate to the shortening of the test, the re-ordering of the test sequence, the varying fatigue levels between the shorter and longer version, and the shortening of the immediate and delayed recall interval, as well as the possible effects of interference (Doss, Chelune, & Naugle, 2000; Tulsky, 2004; Tulsky, Chelune, & Price, 2004). Research conducted by Zhu and Tulsky (2000) has indicated that test length and fatigue did not negatively affect on WMS-III performance, however the effect of interference has as yet not been examined.

The Administration Manual for the WMS-III (Wechsler, 1997) notes that: “During standardization, the WMS-III optional subtests and additional verbal and nonverbal tests were administered during this delay interval” (p. 27). It is thus possible that including the additional memory tasks during the standardization may have increased the likelihood of interference, thus causing decrements in delayed recall performances as a result. However, these may have been treated as the ‘normal’ pattern of forgetting rather than as forgetting subject to interference. If this were the case, then the WMS-III norms may not accurately represent the normal rates of forgetting, making the determination of what constitutes an actual deficit less clear. This may also account in some part for factor analytic research (Bradley Burton, Ryan, Axelrod, Schellenberger, & Richards, 2002;,Millis, Malina, Bowers, & Ricker, 1999; Price, Tulsky, Millis, & Weiss, 2002), which demonstrates that the WMS-III cannot readily differentiate between the domains of immediate and delayed recall.

The present studies attempted to examine the influence of interference theories of forgetting in more detail, using the WMS-III and other commonly used memory tasks. The study took place in two stages. The first used a number of commonly employed memory tasks as interference for the delayed recall of the WMS-III and in the second stage the WMS-III was used as interference for the other memory tasks. The principal hypothesis of Study One was that the delayed recall performances of individuals who completed the WMS-III as well as the interference items would perform significantly worse in comparison to those who complete the WMS-III under standard conditions.

Materials and Methods

Participants

In Study 1, there were 30 normal participants ranging in age from 18 to 60 years (interference: M = 32.19 years, SD = 12.85 and non-interference: M = 33.45, SD = 12.80). They were recruited from advertisements on university and community notice boards and were randomly assigned to the conditions. There were 15 men (50%) and 15 (50%) women. Suitability for participation was determined on the basis of individual responses to a demographic/screening questionnaire and the IQ scores obtained on the Wechsler Abbreviated Scale of Intelligence (WASI; The Psychological Corporation, 1999). One participant was excluded due to poor English proficiency. The samples for each of the studies were independent of each other. The distribution across each of the conditions with regard to age, gender, years of education, as well as WASI scores revealed no differences (Table 1).

Table 1

Demographic and WASI variables for Study 1

 Non-interference (n = 15) Interference (n = 15) Test value Significance 
Age (SD33.45 (12.80) 32.19 (12.85) t(28) = −.380 .707 
Gender 
 Male (n40% (6) 47.7% (7) x2(28) = −.357 .724 
 Female (n60% (9) 53.3% (8)   
Year of education (SD13.20 (2.65) 13.53 (2.70) t(28)= .341 .735 
WASI     
 VIQ (SD101.33 (11.80) 102.60 (5.84) t(28)= .373 .712 
 PIQ (SD101.80 (9.90) 101.00 (9.83) t(28) = −.222 .826 
 FSIQ (SD101.93 (11.59) 101.47 (7.34) t(28) = −.132 .896 
 Non-interference (n = 15) Interference (n = 15) Test value Significance 
Age (SD33.45 (12.80) 32.19 (12.85) t(28) = −.380 .707 
Gender 
 Male (n40% (6) 47.7% (7) x2(28) = −.357 .724 
 Female (n60% (9) 53.3% (8)   
Year of education (SD13.20 (2.65) 13.53 (2.70) t(28)= .341 .735 
WASI     
 VIQ (SD101.33 (11.80) 102.60 (5.84) t(28)= .373 .712 
 PIQ (SD101.80 (9.90) 101.00 (9.83) t(28) = −.222 .826 
 FSIQ (SD101.93 (11.59) 101.47 (7.34) t(28) = −.132 .896 

Note: WASI = Wechsler Abbreviated Scale of Intelligence; VIQ = Verbal Intelligence Quotient; PIQ = Perceptual Intelligence Quotient; FSIQ = Full Scale Intelligence Quotient.

Procedures and Materials

Participants attended two sessions. In the first session, participants completed the demographic and screening information as well as the WASI. In the second session, those in the non-interference group completed the immediate and delayed components of the primary and supplementary tasks from the WMS-III with the 25–35-min delay prescribed by the manual. This delay was unfilled and during this time, the participant being provided with water or coffee. No additional tasks were administered between the immediate and delayed administrations and the length of delay was determined from the time at which the first memory task (i.e., Logical Memory) was administered. Participants in the interference group completed the immediate and delayed components of the WMS-III subtests with conceptually similar memory tasks employed in the time between these components. These tasks included the Logical Memory–Story II from the WMS-Revised (LM-II-R; WMS-R: Wechsler, 1987), the Warrington Recognition Memory for Faces (WR), the Rey Auditory Verbal Learning Test (RAVLT), and the Visual Spatial Learning Test (VSLT) in this set order. These participants were also provided with water or coffee during the interference tasks.

Results

WMS-III Performance

Initially, a repeated measures multivariate analysis of variance (MANOVA) was performed on the WMS-III subtest scaled scores as well as on the raw scores. It was found that there were significant main (F = 1.390, p = <.001, η2 = .996) and time × condition interaction effects (F = 2.214, p = .027, η2 = .371), where condition refers to interference versus non-interference groups. The condition and time effects were also significant (F = 2.601, p = .010, η2 = .409 and F = 88.977, p < .001, η2 = .960, respectively). To establish which individual subtests were significantly different across time, 12 separate repeated measures analysis of variance (ANOVAs; six for scaled scores and six for raw scores) were undertaken. Whereas it is acknowledged that the use of 12 separate ANOVAs is likely to inflate the Type I error rate, the large effect sizes observed in the context of a small sample and the fact that the analysis was post hoc mitigate against the need for correction of the family-wise error. Despite these issues, however, the results should be cautiously interpreted.

Within-subject comparisons indicated that there were significant differences for time alone for the Verbal Paired Associates (F = 13.9, p = .001, η2 = .332) and Family Pictures (F = 8.50, p = .007, η2= .233) subtests. In addition, significant condition differences were demonstrated for the Verbal Paired Associates (F = 8.786, p = .004, η2 = .136), Family Pictures (F = 9.674, p = .003, η2 = .147) and Word Lists (F = 5.941, p = .018, η2 = .096) subtests. Follow-up independent t-test comparisons revealed that the condition differences were only significant at delay, which means that the groups were not significantly different to begin with. Table 2 presents the WMS-III scaled scores and interaction effects of the interference and non-interference conditions. Repeated measures ANOVA revealed significant interaction effects between time and condition on all measures, excluding the Faces subtest.

Table 2

WMS-III scaled score interaction effects for the interference and non-interference groups

 Non-interference (n = 15), M (SDInterference (n = 15), M (SDp η2 
LM I 9.87 (1.46) 9.93 (1.17) <.001 .374 
LM II 10.20 (1.32) 9.33 (1.18) 
Faces I 9.80 (1.32) 9.76 (1.06) .816 .002 
Faces II 9.73 (1.27) 9.07 (1.36) 
VPA I 9.93 (1.28) 9.73 (1.49) <.001 .403 
VPA II 10.07 (1.53) 8.00 (1.65) 
FP I 10.06 (1.67) 9.40 (1.24) .034 .151 
FP II 9.93 (1.10) 8.40 (1.40) 
WL I 8.87 (1.46) 8.67 (1.18) <.001 .452 
WL II 9.73 (1.39) 7.67 (1.23) 
VR I 9.40 (1.55) 9.60 (1.05) .001 .340 
VR II 10.00 (1.55) 8.73 (.960) 
 Non-interference (n = 15), M (SDInterference (n = 15), M (SDp η2 
LM I 9.87 (1.46) 9.93 (1.17) <.001 .374 
LM II 10.20 (1.32) 9.33 (1.18) 
Faces I 9.80 (1.32) 9.76 (1.06) .816 .002 
Faces II 9.73 (1.27) 9.07 (1.36) 
VPA I 9.93 (1.28) 9.73 (1.49) <.001 .403 
VPA II 10.07 (1.53) 8.00 (1.65) 
FP I 10.06 (1.67) 9.40 (1.24) .034 .151 
FP II 9.93 (1.10) 8.40 (1.40) 
WL I 8.87 (1.46) 8.67 (1.18) <.001 .452 
WL II 9.73 (1.39) 7.67 (1.23) 
VR I 9.40 (1.55) 9.60 (1.05) .001 .340 
VR II 10.00 (1.55) 8.73 (.960) 

Note: LM = Logical Memory; VPA = Verbal Paired Associates; FP = Family Pictures; WL = Word Lists; VR = Visual Reproduction.

Repeated measures ANOVA using the WMS-III raw scores indicated that condition differences were significant for the Verbal Paired Associates and Word Lists subtests [F(1,28) = 11.5, p = .002, η2 = .291 and F(1,28) = 4.28, p = .048, η2 = .132, respectively]. Follow-up independent samples t-test comparisons indicated that these differences were only significant for delayed recall. For all the subtests, excluding Faces and Word Lists, there was a significant time × condition interaction effect, indicating that the participants subjected to the interference condition performed more poorly following a delay when compared with the non-interference condition (Table 3). Significant time differences were demonstrated for all subtests except Faces.

Table 3

WMS-III raw score interaction effects for the interference and non-interference groups for Study 1

 Non-interference (n = 15), M (SDInterference (n = 15), M (SDp η2 
LM I 27.40 (2.82) 27.93 (2.96) .009 .217 
LM II 24.67 (1.40) 22.67 (1.33) 
Faces I 38.13 (4.13) 37.80 (4.97) .357 .030 
Faces II 37.33 (2.77) 37.73 (3.34) 
VPA I 7.53 (.644) 7.40 (.630) .003 .291 
VPA II 6.13 (1.13) 4.40 (1.40) 
FP I 42.20 (6.22) 46.73 (3.51) .016 .191 
FP II 47.06 (5.69) 47.93 (6.30) 
WL I 10.73 (1.03) 10.33 (1.35) .094 .097 
WL II 6.67 (1.63) 5.27 (1.62) 
VR I 85.67 (5.31) 86.73 (3.43) <.001 .402 
VR II 65.20 (12.22) 54.30 (6.91) 
 Non-interference (n = 15), M (SDInterference (n = 15), M (SDp η2 
LM I 27.40 (2.82) 27.93 (2.96) .009 .217 
LM II 24.67 (1.40) 22.67 (1.33) 
Faces I 38.13 (4.13) 37.80 (4.97) .357 .030 
Faces II 37.33 (2.77) 37.73 (3.34) 
VPA I 7.53 (.644) 7.40 (.630) .003 .291 
VPA II 6.13 (1.13) 4.40 (1.40) 
FP I 42.20 (6.22) 46.73 (3.51) .016 .191 
FP II 47.06 (5.69) 47.93 (6.30) 
WL I 10.73 (1.03) 10.33 (1.35) .094 .097 
WL II 6.67 (1.63) 5.27 (1.62) 
VR I 85.67 (5.31) 86.73 (3.43) <.001 .402 
VR II 65.20 (12.22) 54.30 (6.91) 

Note: LM = Logical Memory; VPA = Verbal Paired Associates; FP = Family Pictures; WL = Word Lists; VR = Visual Reproduction.

The aim of Study 2 was to compare the effect of introducing conceptually similar information between immediate and delayed administrations when the order of administration was altered, such that the WMS-III subtests now served as the interference items. There were two reasons for this experimental variation. The first was to determine whether similar effects of interference were evident even when the administration order was reversed and also it allowed the results from both studies to be pooled. The primary objective of this approach was to allow for comparisons between the experimental conditions applied in each study.

The specific hypotheses of Study 2 were: (i) that the delayed recall performances of individuals who completed the LM-II-R, WR, RAVLT, and VSLT as well as the interference items would be significantly lower in comparison with those who completed the aforementioned memory tasks under standard conditions, and (ii) that PI and RI effects would not be significantly different across Studies 1 and 2.

Materials and Methods

Participants

Participants in Study 2 were 30 randomly assigned, normal participants, aged from 18 to 60 years (interference: M = 29.73 years, SD = 11.46 and non-interference: M = 32.18, SD = 11.93). They were once again recruited from advertisements on university and community notice boards. There were 15 men (50%) and 15 women (50%). The distribution across each of the conditions with regard to age, gender, years of education, as well as intelligence variables revealed no differences (Table 4).

Table 4

Demographic and WASI variables for Study 2

 Non-interference (n = 15) Interference (n = 15) Test value Significance 
Age (SD32.18 (11.93) 29.73 (11.46) t(28) = −.021 .885 
Gender 
 Male (n46.7% (7) 53.3% (8) x2(28) = −.354 .726 
 Female (n53.3% (8) 46.7% (7)   
Year Education (SD12.40 (2.80) 13.00 (2.51) t(28) = .071 .791 
WASI     
 VIQ (SD101.00 (10.16) 98.60 (8.35) t(28) = .910 .347 
 PIQ (SD99.23 (10.09) 101.73 (4.59) t(28) = −1.258 .219 
 FSIQ (SD101.87 (11.64) 99.00 (8.05) t(28) = 3.17 .086 
 Non-interference (n = 15) Interference (n = 15) Test value Significance 
Age (SD32.18 (11.93) 29.73 (11.46) t(28) = −.021 .885 
Gender 
 Male (n46.7% (7) 53.3% (8) x2(28) = −.354 .726 
 Female (n53.3% (8) 46.7% (7)   
Year Education (SD12.40 (2.80) 13.00 (2.51) t(28) = .071 .791 
WASI     
 VIQ (SD101.00 (10.16) 98.60 (8.35) t(28) = .910 .347 
 PIQ (SD99.23 (10.09) 101.73 (4.59) t(28) = −1.258 .219 
 FSIQ (SD101.87 (11.64) 99.00 (8.05) t(28) = 3.17 .086 

Note: WASI = Wechsler Abbreviated Scale of Intelligence; VIQ = Verbal Intelligence Quotient; PIQ = Perceptual Intelligence Quotient; FSIQ = Full Scale Intelligence Quotient.

Procedures and Materials

The first session was the same as that outlined in Study 1. In the second session, those participants in the non-interference group completed the immediate and delayed recall components of the LM-II-R, WR, RAVLT, and VSLT with the 30-min delay. Again, no additional tasks were administered between the immediate and delayed administrations and the length of delay was determined by the time at which the first memory task (i.e., WMS-R Logical Memory) was administered. Participants in the interference group also completed the immediate and delayed components of these tasks although the time between these components was filled with all the primary and supplementary memory subtests from the WMS-III.

Results

LM-II-R, WRMT, RAVLT, and VSLT Performance

Repeated measures MANOVA results indicated a significant main effect (F = 1.530, p = .032, η2 = .250) and a time × condition interaction effect (F = 3.861, p = .008, η2 = .226). Condition and time effects were also significant (F = 6.520, p < .001, η2 = .330 and F = 16.514, p < .001, η2 = .555, respectively). The individual repeated measures ANOVA interference and non-interference condition results on LM-II-R, WR, RAVLT, and VSLT are presented in Table 5. As the table demonstrates, there was a significant time × condition interaction effect for LM-II-R and RAVLT with those in the interference condition performing more poorly following a delay when compared with those without interference. This was not demonstrated for WR and VSLT, although VSLT approached significance. Differences over time were also demonstrated for RAVLT [F(1,28) = 56.32, p = .000, η2 = .668] as well as for the VLST [F(1,28) = 22.47, p = .000, η2 = .445].

Table 5

Analysis of variance interaction effect results for LM-II-R, WR, RAVLT, and VSLT raw scores

 Non-interference (n = 15), M (SDInterference (n = 15), M (SDp η2 
LM-II-R I 13.93 (2.67) 14.80 (2.37) .017 .187 
LM-II-R II 13.53 (1.45) 11.47 (2.13) 
WR I 46.40 (3.89) 45.20 (2.81) .934 <.001 
WR II 45.60 (2.35) 44.53 (2.80) 
RAVLT I 13.53 (1.30) 12.67 (1.40) .033 .152 
RAVLT II 11.60 (1.88) 8.20 (1.86) 
VSLT I 13.60 (.830) 13.47 (.920) .083 .14 
VSLT II 12.40 (1.35) 10.80 (2.48) 
 Non-interference (n = 15), M (SDInterference (n = 15), M (SDp η2 
LM-II-R I 13.93 (2.67) 14.80 (2.37) .017 .187 
LM-II-R II 13.53 (1.45) 11.47 (2.13) 
WR I 46.40 (3.89) 45.20 (2.81) .934 <.001 
WR II 45.60 (2.35) 44.53 (2.80) 
RAVLT I 13.53 (1.30) 12.67 (1.40) .033 .152 
RAVLT II 11.60 (1.88) 8.20 (1.86) 
VSLT I 13.60 (.830) 13.47 (.920) .083 .14 
VSLT II 12.40 (1.35) 10.80 (2.48) 

Note: LM-II-R = Logical Memory Story II from WMS-R; WR = Warrington Recognition Memory for Faces; RAVLT = Rey Auditory Verbal Learning Test; VSLT = Visual-Spatial Learning Test.

Cross-Study Comparisons

Demographic variables and IQ scores

There were 60 participants in total between the ages of 18 and 60 for the combined sample from both Studies 1 and 2. The gender breakdown was 30 males and 30 females. Analysis of the demographic and screening characteristics indicated that the participants in both Studies 1 and 2 were equivalent in terms of age, gender, education, and intelligence (all p > .05).

WMS-III Performance

Examination of the repeated measures ANOVA results for the WMS-III scaled and raw scores revealed that there was a significant effect for time in the absence of interaction or condition effects for all subtests, excluding Faces, indicating that the delayed recall performances were significantly reduced in comparison with immediate recall in both studies (all p > .05; Table 6).

Table 6

WMS-III scaled score time effects for Studies 1 and 2

 Study 1 (n = 15), M (SDStudy 2 (n = 15), M (SDp η2 
LM I 9.93 (1.17) 9.47 (1.46) .003 .279 
LM II 9.33 (1.18) 9.20 (1.12) 
Faces I 9.76 (1.06) 9.33 (1.05) .139 .077 
Faces II 9.07 (1.36) 9.00 (1.13) 
VPA I 9.73 (1.49) 9.27 (1.28) <.001 .480 
VPA II 8.00 (1.65) 8.33 (1.45) 
FP I 9.40 (1.24) 9.00 (.530) .001 .324 
FP II 8.40 (1.40) 8.53 (.743) 
WL I 8.67 (1.18) 8.93 (1.97) .003 .270 
WL II 7.67 (1.23) 8.40 (1.24) 
VR I 9.60 (1.05) 9.40 (.985) <.001 .437 
VR II 8.73 (.960) 8.40 (.032) 
 Study 1 (n = 15), M (SDStudy 2 (n = 15), M (SDp η2 
LM I 9.93 (1.17) 9.47 (1.46) .003 .279 
LM II 9.33 (1.18) 9.20 (1.12) 
Faces I 9.76 (1.06) 9.33 (1.05) .139 .077 
Faces II 9.07 (1.36) 9.00 (1.13) 
VPA I 9.73 (1.49) 9.27 (1.28) <.001 .480 
VPA II 8.00 (1.65) 8.33 (1.45) 
FP I 9.40 (1.24) 9.00 (.530) .001 .324 
FP II 8.40 (1.40) 8.53 (.743) 
WL I 8.67 (1.18) 8.93 (1.97) .003 .270 
WL II 7.67 (1.23) 8.40 (1.24) 
VR I 9.60 (1.05) 9.40 (.985) <.001 .437 
VR II 8.73 (.960) 8.40 (.032) 

Note: LM = Logical Memory; VPA = Verbal Paired Associates; FP = Family Pictures; WL = Word Lists; VR = Visual Reproduction.

LM-II-R, WRMT, RAVLT, and VSLT Performance

Repeated measures ANOVA using LM-II-R, WR, RAVLT, and VSLT indicated that for all tasks, excluding WR, there was a significant time effect (Table 7), with individuals performing more poorly following a delay in comparison with their immediate recall. Condition and interaction effects were not significant.

Table 7

ANOVA time effect results for LM-II-R, WR, RAVLT, and VSLT raw scores

 Study 1 (n = 15), M (SDStudy 2 (n = 15), M (SDp η2 
LM-II-R I 14.26 (2.22) 14.80 (2.37) <.001 .465 
LM-II-R II 12.80 (1.86) 11.47 (2.13) 
WR I 44.27 (2.25) 45.20 (2.81) .131 .080 
WR II 43.20 (3.05) 44.53 (2.80) 
RAVLT I 12.20 (1.86) 12.67 (1.40) <.001 .833 
RAVLT II 6.27 (3.17) 8.20 (1.86) 
VSLT I 12.80 (1.47) 13.47 (.920) <.001 .557 
VSLT II 9.73 (2.91) 10.80 (2.48) 
 Study 1 (n = 15), M (SDStudy 2 (n = 15), M (SDp η2 
LM-II-R I 14.26 (2.22) 14.80 (2.37) <.001 .465 
LM-II-R II 12.80 (1.86) 11.47 (2.13) 
WR I 44.27 (2.25) 45.20 (2.81) .131 .080 
WR II 43.20 (3.05) 44.53 (2.80) 
RAVLT I 12.20 (1.86) 12.67 (1.40) <.001 .833 
RAVLT II 6.27 (3.17) 8.20 (1.86) 
VSLT I 12.80 (1.47) 13.47 (.920) <.001 .557 
VSLT II 9.73 (2.91) 10.80 (2.48) 

Note: LM-II-R = Logical Memory Story II from WMS-R; WR = Warrington Recognition Memory for Faces; RAVLT = Rey Auditory Verbal Learning Test; VSLT = Visual-Spatial Learning Test.

Discussion

The objective of the current research was to examine the effect of interference on delayed recall performance of the WMS-III and other commonly used memory tasks. In support of the hypotheses proposed, it was found in Study 1 that introducing additional information between immediate and delayed trials on the WMS-III negatively affected recall performance following a delay on most tasks. A similar pattern of performance was evident in Study 2 when other memory tasks were employed. These findings were consistent with research examining the concept of RI and its effect on learning and memory (Bauml, 1996; Briggs, 1957; McGeoch, 1932; Richardson, 1956).

Collating the data from both studies allowed for examination of the effect of PI (see Fig. 1). In accordance with this theory, declines in delayed recall performances were demonstrated when interference was introduced when compared with individuals who completed the required memory task under standard conditions. These studies also supported previous interference research, in that, although it continues to be debated, no one theory of interference has yielded greater empirical support when compared with the others (Anderson & Spellman, 1995). Thus, the findings in these studies indicate that similar patterns of recall were evident irrespective of when the interference was introduced.

In relation the WMS-III as a whole, the results validate the concerns raised by Doss, and coworkers (2000) about the standardization process of the WMS-III and the resultant scaled scores. Whereas the study failed to identify differences between a normal population and epilepsy patients with the two versions of the battery (i.e., standardized and published), it does raise the question whether this may have been because the standard scores used to determine differences have an intrinsic interference bias.

The published version of the WMS-III is shorter in length and as a result is briefer in its administration time (The Psychological Corporation, 2002). This has raised a number of concerns in the literature, some of which were addressed in the study conducted by Zhu and Tulsky (2000). The results of these studies indicate that one of the other concerns raised, the role of interference, is an issue that needs to be given more serious consideration when interpreting the current WMS and in developing revisions of the Scale.

These studies illustrate that introducing other memory tasks in the delay interval serves to interfere with performance following the delay. Thus, it is more than likely that the inclusion of a number of other memory tasks during standardization processes of memory batteries may produce a pattern of results that may lack sensitivity. Let us consider for example, the supplementary memory subtests on the WMS-III; Visual Reproduction and Word Lists. On each of these tasks, individuals can forget between 30% and 40% of the information originally encoded and still receive scores that are not only within the average range, but are often higher than the scaled scores obtained on initial recall. These retention rates are in no way similar to the rates of retention obtained on the other memory tasks employed in these studies, where on average a loss of 20% of the information was considered ‘normal’. This is an issue of considerable concern given that clinicians who may be unaware of this finding may be underestimating a person's memory impairment within clinical samples because they received an ‘average’ scaled score.

One of the major limitations of the present research was the small sample sizes in the experimental and control groups for each study (n = 15 for both Studies 1 and 2). However, it is important to note that the statistically significant findings were further supported by the presence of large to very large effect sizes for the majority of items examined. This indicates that there were large to very large differences between immediate and delayed recall performances. These were observed despite a small sample size and lower statistical power.

Fig. 1.

Statistical comparisons made between Studies 1 and 2.

Fig. 1.

Statistical comparisons made between Studies 1 and 2.

Another limitation is that no measure was employed to control for the effect of the similarity between the interference and test material. Thus, although clear effects of interference were demonstrated on the basis of the study design, it cannot be concluded whether these findings were elevated as a result of the task similarity or whether these results would have been obtained irrespective of similarity. Whereas it is recognized that it would not be standard clinical practice to administered as many memory items (especially those that are conceptually similar) between immediate and delayed trials, these studies were designed to assess interference that may have direct implications for the development of memory batteries. A useful line of future research would be to examine more typical clinical practices, whereby verbal tasks such as fluency or vocabulary assessment are administered during the immediate and delayed trials of a word list, to establish whether there is an interference effect.

Overall, the results of these studies stand as a caution to clinicians with regard to employing similar material to the WMS-III or other memory tests in the period between the immediate and delayed recall trials and highlight the importance of consideration of interference effects when standardizing memory tasks.

Conflict of Interest

None declared.

References

Anderson
M. C.
Bjork
R. A.
Bjork
E. L.
Remembering can cause forgetting: Retrieval dynamics in long-term memory
Journal of Experimental Psychology: Learning, Memory and Cognition
 , 
1994
, vol. 
20
 (pg. 
1063
-
1087
)
Anderson
M. C.
Neely
J. H.
Bjork
E. L.
Bjork
R. A.
Interference and inhibition in memory retrieval
Handbook of perception and cognition
 , 
1996
2nd ed.
San Diego, CA
Academic Press
(pg. 
237
-
313
Memory
Anderson
M. C.
Spellman
B. A.
On the status of inhibitory mechanisms in cognition: Memory retrieval as a model case
Psychological Review
 , 
1995
, vol. 
102
 (pg. 
68
-
100
)
Bauml
K.
Revisiting an old issue: Retroactive interference as a function of the degree of interpolated learning
Psychonomic Bulletin and Review
 , 
1996
, vol. 
3
 (pg. 
380
-
384
)
Bower
G. H.
Thompson-Schill
S.
Tulving
E.
Reducing retroactive interference: An interference analysis
Journal of Experimental Psychology: Learning, Memory and Cognition
 , 
1994
, vol. 
20
 (pg. 
51
-
66
)
Bradley Burton
D.
Ryan
J. J.
Axelrod
B. N.
Schellenberger
T.
Richards
H. M.
A confirmatory factor analysis of the WMS-III in a clinical sample with cross validation in the standardization sample
Archives of Clinical Neuropsychology
 , 
2002
, vol. 
18
 (pg. 
629
-
641
)
Briggs
G. E.
Retroactive inhibition as a function of degree of original and interpolated learning
Journal of Experimental Psychology
 , 
1957
, vol. 
53
 (pg. 
60
-
67
)
Doss
R. C.
Chelune
G. J.
Naugle
R. I.
Comparability of the expanded WMS-III standardisation protocol to the published WMS-III among right and left temporal lobectomy patients
Clinical Neuropsychologist
 , 
2000
, vol. 
14
 (pg. 
468
-
473
)
McGeoch
J. A.
Forgetting and the law of disuse
Psychological Review
 , 
1932
, vol. 
39
 (pg. 
352
-
370
)
Millis
S. R.
Malina
A. C.
Bowers
D. A.
Ricker
J. H.
Confirmatory factor analysis of the Wechsler Memory Scale – III
Journal of Clinical and Experimental Neuropsychology
 , 
1999
, vol. 
21
 (pg. 
87
-
93
)
Price
L. R.
Tulsky
D.
Millis
S.
Weiss
L.
Redefining the factor structure of the Wechsler Memory Scale–III: Confirmatory factor analysis with cross-validation
Journal of Clinical and Experimental Neuropsychology
 , 
2002
, vol. 
24
 (pg. 
574
-
585
)
Richardson
J.
Retention of concepts as a function of degree of original and interpolated learning
Journal of Experimental Psychology
 , 
1956
, vol. 
51
 (pg. 
358
-
364
)
The Psychological Corporation
The WAIS-III–WMS-III technical manual: Update
 , 
2002
San Antonio, TX
Author
Shimamura
A. P.
Jurica
P. J.
Mangels
J. A.
Gershberg
F. B.
Knight
R. T.
Susceptibility to memory interference effects following frontal lobe damage: Findings from tests of paired-associate learning
Journal of Cognitive Neuroscience
 , 
1995
, vol. 
7
 
2
(pg. 
144
-
152
)
Tulsky
D. S.
A new look at the WMS-III: New research to guide clinical practice
Journal of Clinical and Experimental Neuropsychology
 , 
2004
, vol. 
26
 (pg. 
453
-
458
)
Tulsky
D. S.
Chelune
G. J.
Price
L. R.
Development of a new delayed memory index for the WMS-III
Journal of Clinical and Experimental Neuropsychology
 , 
2004
, vol. 
26
 (pg. 
563
-
576
)
Underwood
B. J.
Proactive inhibition with increased recall time
American Journal of Psychology
 , 
1950
, vol. 
63
 (pg. 
594
-
599
)
Underwood
B. J.
Interference and forgetting
Psychological Review
 , 
1957
, vol. 
64
 (pg. 
49
-
60
)
Wechsler
D.
Wechsler Memory Scale-Revised; Manual
 , 
1987
San Antonio, TX
The Psychological Corporation
Wechsler
D.
Wechsler Memory Scale-Third Edition
 , 
1997
San Antonio, TX
The Psychological Corporation
Zhu
J.
Tulsky
D. S.
Co-norming of the WAIS-III and WMS-III: Is there a test-order effect on IQ and memory scores?
Clinical Neuropsychologist
 , 
2000
, vol. 
14
 (pg. 
461
-
467
)