CoChat
Sign Up
Cited by 175Visual and Cognitive Learning ProcessesInnovative Teaching and Learning MethodsIntelligent Tutoring Systems and Adaptive Learning

An expertise reversal effect of segmentation in learning from animated worked-out examples

Open access
Spanjers, Ingrid A.E., Wouters, Pieter, van Gog, Tamara, van Merriënboer, Jeroen J.G.·2011·Computers in Human Behavior
DOI 10.1016/j.chb.2010.05.011
Open this paper in CoChat — free accountOpen PDF

AI snapshot

Written from the paper’s full text.

TL;DR

This experiment found an expertise reversal effect of segmentation in animated worked-out examples: segmented animations were more efficient than continuous animations for students with lower prior knowledge, but not for students with higher prior knowledge. The effect appeared on mental effort and instructional efficiency, not on transfer test performance.

Key findings

  • Segmented animations were more efficient than continuous animations at one standard deviation below the mean prior knowledge: near transfer, b = 0.39, t(71) = 2.60, p = .01; far transfer, b = 0.33, t(71) = 2.20, p = .03. At one standard deviation above the mean, this difference disappeared: near, b = −0.05, t(71) = −0.36, p = .72; far, b = −0.10, t(71) = −0.67, p = .51.
  • Mental effort during instruction showed a significant prior knowledge × condition interaction, b = 0.38, t(71) = 2.06, p = .04. At lower prior knowledge, continuous examples were marginally higher in mental effort than segmented ones, b = −0.32, t(71) = −1.98, p = .05; at higher prior knowledge, continuous examples decreased in mental effort, b = −0.49, t(71) = −2.71, p = .01, while segmented examples remained almost equal, b = −0.02, t(71) = −0.11, p = .91.
  • No significant interaction was found for near or far transfer performance. The regression models were significant for near transfer, F(3, 71) = 8.61, p = .00, adjusted R² = .24, and far transfer, F(3, 71) = 6.01, p = .00, adjusted R² = .17.
  • Efficiency regression models were significant for near transfer, F(3, 71) = 7.13, p = .00, adjusted R² = .20, and far transfer, F(3, 71) = 6.70, p = .00, adjusted R² = .19, with significant interaction coefficients for near, b = −0.35, t(71) = −2.07, p = .04, and far, b = −0.34, t(71) = −2.00, p = .05.

Methods

  • Participants were 76 Dutch secondary education students (53% female; mean age = 16.16, SD = 0.80), randomly assigned to segmented or continuous animated worked-out examples. One participant was excluded for missing values, leaving 37 in the segmented and 38 in the continuous condition.
  • A computer-based Flash learning environment administered a prior knowledge test (4 multiple-choice and 8 open items; max score 12), eight animated worked-out examples on probability calculation (computer-paced, approximately 2 min each), transfer tests (8 open near-transfer and 4 open far-transfer items), and a 9-point mental effort rating after each animated example. Segmented examples were divided into 5–7 segments with 2-s pauses and a slightly darkened screen; continuous examples were shown as one stream.
  • Regression analyses entered prior knowledge (centered), condition (0 = continuous, 1 = segmented), and their interaction simultaneously as predictors. Follow-up tests examined condition effects at one standard deviation below and above the mean prior knowledge. Efficiency was calculated using the Paas and Van Merriënboer formula with test performance and mental effort invested during learning.

Limitations

  • The authors state that the pauses between segments created a difference in learning time between conditions, with segmented students having more time per animated example. They note that no additional information was provided during pauses and that 2 s is probably too short for reflection, but future studies should control time-on-task.
  • The authors state that the study did not reveal the cognitive processes underlying the expertise reversal effect, and suggest future research using cued retrospective reporting combined with eye movements.
  • The authors state that it is unclear whether the results generalize to animations for declarative knowledge about natural processes or mechanical systems. They also suggest that the limited range of prior knowledge—most participants were not complete novices or full experts—may explain why no interaction was found on performance.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1016/j.chb.2010.05.011

Cites

35 of 35 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

0 of 16 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    The interaction between prior knowledge and condition on near transfer test performance was not significant.

    “The regression coefficients for the interaction terms were not significant, however.”
    Quote in the paperSubstantial claim
  • Findingown data

    With higher levels of prior knowledge, the instructional efficiency of segmented animated examples increased more slowly and only marginally significantly on the far transfer test (b = .25, t(71) = 1.85, p = .07).

    “whereas the efficiency of segmented, animated examples increased more slowly and only marginally significant, near: b = 0.24, t(71) = 1.79, p = .08,far: b = .25,t(71) = 1.85, p = .07.”
    Quote in the paperSubstantial claim
  • Findingown data

    The interaction between prior knowledge and condition was significant for instructional efficiency on the near transfer test (b = -0.35, t(71) = -2.07, p = .04).

    “and so were the regression coefficients for the interaction term, near: b = À0.35,t(71) = À2.07,p = .04;far: b = À0.34,t(71) = À2.00,p = .05.”
    Quote in the paperSubstantial claim
  • Findingown data

    The interaction between prior knowledge and condition on far transfer test performance was not significant.

    “The regression coefficients for the interaction terms were not significant, however.”
    Quote in the paperSubstantial claim
  • Findingown data

    At one standard deviation below the mean prior knowledge, the mental effort invested in studying continuous animated examples was marginally higher than in studying segmented ones (b = -0.32, t(71) = -1.98, p = .05).

    “At one standard deviation below the mean the difference in mental effort invested in studying continuous, animated examples was marginally higher than in studying segmented ones, b = À0.32,t(71) = À1.98,p = .05.”
    Quote in the paperSubstantial claim
  • Classified as probably not a substantial claim. 10
  • May combine two claims. 6

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    The regression model for mental effort during instruction was marginally significant (F(3, 71) = 2.57, p = .06, adjusted R2 = .06).

    “The regression model for mental effort during instruction was marginally significant, F(3, 71) = 2.57, p = .06,adjusted R 2 = .06”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductreasoning

    A possible limitation of the study was that the pauses between segments led to a difference in learning time between conditions, with students in the segmented condition having more time for each animated example than students in the continuous condition.

    “A possible limitation of this study was that the pauses between segments led to a difference in learning time between the conditions: students in the segmented condition had more time for each animated example than students in the continuous condition.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    Efficiency on both the near and the far transfer test was calculated using the formula developed by Paas and Van Merriënboer (1993).

    “Efficiency on both the near and the far transfer test was calculated using the formula developed by Paas and Van Merriënboer (1993)”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Paas & Van Merriënboer, 1993 source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Working memory has a limited capacity and time constraints, so only two or three information elements can be simultaneously processed.

    “Working memory has a limited capacity (Baddeley, 2003) and has time constraints (Barrouillet & Camos, 2007), as a consequence of which only two or three information elements can be simultaneously processed.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Baddeley, 2003 source

      Not checked in the free run.

    • cites Barrouillet & Camos, 2007

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductnone given

    During the pauses between segments, no additional information was provided; only the last frame of the previous segment was visible.

    “Note though, that during these pauses no additional information was provided (only the last frame of the previous segment was visible)”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Cognitive Load Theory distinguishes three additive types of cognitive load imposed on working memory: extraneous, germane, and intrinsic.

    “Cognitive Load Theory (Sweller, Van Merriënboer, & Paas, 1998;Van Merriënboer & Sweller, 2005) distinguishes between three additive types of cognitive load imposed on working memory by maintaining and processing information: Extraneous (i.e., ineffective) and germane (i.e., effective) cognitive load imposed by the design of instructional materials, and intrinsic cognitive load, imposed by the”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Sweller, Van Merriënboer, & Paas, 1998 source

      Not checked in the free run.

    • cites Van Merriënboer & Sweller, 2005 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    Only for students with lower prior knowledge, segmentation successfully reduces the high cognitive load imposed by animations and leads to more efficient learning.

    “This suggests that only for students with lower levels of prior knowledge, segmentation successfully reduces the high cognitive load imposed by animations and leads to more efficient learning.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data

    Participants were randomly assigned to either the segmented or the continuous animated worked-out examples condition.

    “They were randomly assigned to either the segmented or the continuous, animated worked-out examples condition.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    A prior knowledge test consisting of four multiple-choice and eight open items about probability calculation was used to measure students' prior knowledge.

    “A prior knowledge test consisting of four multiple-choice and eight open items about probability calculation (see Wouters et al., 2009Wouters et al., , 2010) ) was used to measure students' prior knowledge.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Wouters et al., 2009 source

      Not checked in the free run.

    • cites Wouters et al., 2010 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    Mental effort was measured with a 9-point subjective rating scale ranging from very, very low to very, very high mental effort.

    “Immediately after studying each animated example, students rated how much mental effort they invested in studying it on a 9-point subjective rating scale ranging from (1) very, very low mental effort to ( 9) very, very high mental effort (Paas, 1992).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Paas, 1992 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    The experiment showed an expertise reversal effect of segmentation in terms of efficiency: segmented animated worked-out examples were more efficient than continuous ones for students with lower prior knowledge, but this superiority disappeared for students with higher prior knowledge.

    “This experiment with animated worked-out examples showed an expertise reversal effect of segmentation: segmented animations were more efficient than continuous animations (i.e., equal test performance with lower investment of mental effort during learning) for students with lower levels of prior knowledge, but not for students with higher levels of prior knowledge.”
    • Also stated in: § 4. Discussion · hedged
      “In sum, this study provides some evidence that an expertise reversal effect of segmentation in animated worked-out examples occurs.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    Schnotz (2002) found a positive effect of segmentation of a simulation for students with lower prior knowledge, but no effect for students with higher prior knowledge on one of the tests.

    “He found a positive effect of segmentation of a simulation for students with lower levels of prior knowledge, but no effect for students with higher levels of prior knowledge on one of the tests.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Schnotz, 2002

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    Segmentation has been proposed as a means to reduce the high cognitive load occurring due to the transience of animations.

    “Segmentation has been proposed as a means to reduce the high load occurring due to the transience of animations (e.g., Ayres & Paas, 2007;Mayer & Moreno, 2003;Schnotz & Lowe, 2008).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Ayres & Paas, 2007

      Not checked in the free run.

    • cites Mayer & Moreno, 2003 source

      Not checked in the free run.

    • cites Schnotz & Lowe, 2008

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    Data from one participant were excluded due to too many missing values, resulting in 37 students in the segmented condition and 38 in the continuous condition.

    “Data from one participant had to be excluded due to too many missing values, which resulted in 37 students in the segmented and 38 in the continuous condition.”
    Quote in the paperNot a substantial claim
14 of 14

  • The paper relies on this without support.

    A 2 s pause is too short for processes such as reflection.

    “and that 2 s is probably too short for processes such as reflection.”
1 of 1

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Segmented animations are more efficient than continuous animations for novice learners, but not for learners with higher levels of prior knowledge.

    “The segmentation effect was reported by Spanjers et al. (2011), who found that segmented animations (i.e. segmented in parts with pauses in between) were more efficient than continuous animations for novice learners, but not for learners with higher levels of prior knowledge.”

    From Cognitive Architecture and Instructional Design: 20 Years Later (2019) · cites this paper as Spanjers et al. (2011)

Statistics recomputed

Every reported test result, recomputed from its test statistic and degrees of freedom. A result shows as not adding up only once an independent check has confirmed it.

  • ConsistentF(3, 71) = 8.61reported p = 0 · recomputed p < .001

    “F(3, 71) = 8.61, p = .00”

  • ConsistentF(3, 71) = 6.01reported p = 0 · recomputed p = .001

    “F(3, 71) = 6.01, p = .00”

  • ConsistentF(3, 71) = 2.57reported p = .06 · recomputed p = .061

    “F(3, 71) = 2.57, p = .06”

  • Consistentt(71) = 2.06reported p = .04 · recomputed p = .043

    “t(71) = 2.06, p = .04”

  • Consistentt(71) = -1.98reported p = .05 · recomputed p = .052

    “t(71) = À1.98,p = .05”

  • Consistentt(71) = -2.71reported p = .01 · recomputed p = .008

    “t(71) = À2.71,p = .01”

  • Consistentt(71) = -0.11reported p = .91 · recomputed p = .913

    “t(71) = À0.11,p = .91”

  • Consistentt(71) = 0.98reported p = .33 · recomputed p = .330

    “t(71) = 0.98, p = .33”

  • ConsistentF(3, 71) = 7.13reported p = 0 · recomputed p < .001

    “F(3, 71) = 7.13, p = .00”

  • ConsistentF(3, 71) = 6.7reported p = 0 · recomputed p < .001

    “F(3, 71) = 6.70, p = .00”

  • Consistentt(71) = -2.07reported p = .04 · recomputed p = .042

    “t(71) = À2.07,p = .04”

  • Consistentt(71) = -2reported p = .05 · recomputed p = .049

    “t(71) = À2.00,p = .05”

  • Consistentt(71) = 2.6reported p = .01 · recomputed p = .011

    “t(71) = 2.60, p = .01”

  • Consistentt(71) = 2.2reported p = .03 · recomputed p = .031

    “t(71) = 2.20, p = .03”

  • Consistentt(71) = 4.04reported p = 0 · recomputed p < .001

    “t(71) = 4.04, p = .00”

  • Consistentt(71) = 4reported p = 0 · recomputed p < .001

    “t(71) = 4.00, p = .00”

  • Consistentt(71) = 1.79reported p = .08 · recomputed p = .078

    “t(71) = 1.79, p = .08”

  • Consistentt(71) = 1.85reported p = .07 · recomputed p = .068

    “t(71) = 1.85, p = .07”

  • Consistentt(71) = -0.36reported p = .72 · recomputed p = .720

    “t(71) = À0.36,p = .72”

  • Consistentt(71) = -0.67reported p = .51 · recomputed p = .505

    “t(71) = À0.67,p = .51”

20 of 20

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

  • Mental-effort interaction does not match the condition-specific slopes

    Minor

    With continuous coded 0 and segmented coded 1, the interaction coefficient should equal the segmented prior-knowledge slope minus the continuous slope: −0.02 − (−0.49) = 0.47, not 0.38.

    Reported 0.38 · recomputed 0.47

1 of 1

Cites

Flagged first, then by citations

Research works printed but not linked: 1

  1. Schnotz, W. (2002). Enabling, facilitating and inhibiting effects in learning from animated pictures. Paper presented at the International Workshop on Dynamic Visualizations and Learning, Tübingen, Germany. Downloaded on 31 July 2008 from http://www.iwm-kmrc.de/workshops/visualization/schnotz.pdf.

Cited by

Sorted by citations

Related

Authors
Spanjers, I. A. E., Wouters, P., van Gog, T., & van Merriënboer, J. J. G.
Affiliations
  • Maastricht University (NL)
  • Utrecht University (NL)
Corresponding author
Spanjers, Ingrid A.E.
Venue
Computers in Human Behavior
Publisher
Elsevier BV
Published
2011-01
Volume
27
Issue
1
Pages
46-52
ISSN
0747-5632 (ISSN-L)
DOI
10.1016/j.chb.2010.05.011
OpenAlex
W2059227461
Semantic Scholar
733a638bd03ef6d298932615e83045a9fffc115b
Cited by
175
References
35
Citations by year
2026: 8 · 2025: 12 · 2024: 8 · 2023: 7 · 2022: 8 · 2021: 27 · 2020: 15 · 2019: 10 · 2018: 13 · 2017: 11 · 2016: 9 · 2015: 4 · 2014: 13 · 2013: 10 · 2012: 10
FWCI
12.56
Type
journal article
Language
en
Open access
Licence
Locations
  • Computers in Human Behavior (publishedVersion)
  • Research Publications (Maastricht University) (publishedVersion, other-oa)
  • EUR Research Repository (Erasmus University Rotterdam) (publishedVersion)
  • EUR Research Repository (Erasmus University Rotterdam) (submittedVersion)
  • DSpace (Open University in the Netherlands) (submittedVersion, cc-by-nc-nd)
  • DSpace (Open University in the Netherlands) (submittedVersion, other-oa)
  • Research Publications (Maastricht University) (submittedVersion, other-oa)
Funders
Nederlandse Organisatie voor Wetenschappelijk Onderzoek
Topics
  • Social Sciences › Psychology › Experimental and Cognitive Psychology › Visual and Cognitive Learning Processes
  • Social Sciences › Psychology › Developmental and Educational Psychology › Innovative Teaching and Learning Methods
  • Physical Sciences › Computer Science › Artificial Intelligence › Intelligent Tutoring Systems and Adaptive Learning
Keywords
expertise reversal effect, segmentation, cognitive load, prior knowledge, cognitive effort, instructional design

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-10.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-10.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review