You are on page 1of 15

Journal of Learning Disabilities

Volume 42 Number 3
May/June 2009 215-229

Retrieval of Simple Addition Facts


© 2009 Hammill Institute on
Disabilities
10.1177/0022219408331041
http://journaloflearningdisabilities
Complexities Involved in Addressing a Commonly .sagepub.com
hosted at
Identified Mathematical Learning Difficulty http://online.sagepub.com

Sarah Hopkins
Helen Egeberg
University of Western Australia

There exists a substantial number of studies that have identified a subset of low-achieving mathematics students who do not
develop a reliance on retrieval for simple addition but who continue to use a counting strategy to solve these problems. There
are few studies, however, that have focused on how retrieval of simple addition facts may be improved. This study employed
a combined methodological approach to examine the effect extended practice had on increasing a reliance on retrieval for
simple addition. An intervention aimed at improving the efficiency of extended practice was also piloted. Although most
students improved with extended practice, the extent of improvement was not practical for all students and the intervention
did not generally improve the effectiveness of extended practice. The findings emphasize the critical importance of continu-
ing such research and draw attention to the complexities involved in addressing retrieval difficulties for simple addition.

Keywords:  retrieval; mathematical learning difficulties; simple addition; basic facts

I t is well documented in the research literature that


there exists a subgroup of low-achieving students in
mathematics who exhibit retrieval difficulties and con-
components at least once, as in “1, 2, 3 _ 4, 5, 6, 7, 8.”
The count-from-first and min-counting strategies are
sometimes collectively referred to as count-on strategies.
tinue to rely on counting to perform simple (single-digit) If the problem were 3 + 5, the child using a count-from-
addition (Geary, 2004; Geary, Hamson, & Hoard, 2000; first strategy would start at 3 and count on 5, as in “3 _ 4,
Hopkins & Lawson, 2006a; Jordan & Montani, 1997; 5, 6, 7, 8”; the child using the min-counting strategy
Ostad, 1997; Ostad & Sorensen, 2007) but intervention would start at 5 and count on 3, as in “5 _ 6, 7, 8.”
research in this area is sorely lacking. The aims of this By early or middle primary school (usually around
study were to (a) examine the effect that extended prac- Years 2 and 3), practice at successfully applying the min-
tice had on increasing a reliance on retrieval for simple counting strategy leads to an increased reliance on
addition (encompassing direct retrieval and decomposi- retrieval among normally achieving children, and simple
tion strategies) and (b) pilot test an intervention designed addition is no longer a focus of instruction in schools
to improve the effectiveness of extended practice on (Hopkins & Lawson, 2002). Siegler’s strategy choice
increasing a reliance on retrieval, for students in a main- model (Shrager & Siegler, 1998; Siegler & Jenkins,
stream secondary school setting. 1989; Siegler & Shrager, 1984) provides an explanation
Learning to perform simple addition problems by of how direct retrieval begins to dominate simple
directly retrieving the answer (i.e., just knowing the addition performance: practice using a back up strategy
answer) is a complex developmental process (Hopkins & (any strategy other than direct retrieval) for each problem
Lawson, 2002). Before direct retrieval dominates leads to a frequency distribution of answers (correct and
performance, young children gain important conceptual incorrect) that is associated with the problem in memory;
knowledge of numbers and construct increasingly more correct practice leads to a peaked distribution of
efficient counting strategies. The longsum- and sum-
counting strategies are sometimes collectively referred to
Authors’ Note: We gratefully acknowledge the support of the
as count-all strategies and involve one or two rounds of Spencer Foundation for partly funding this research (# 200700173)
counting starting at 1. For example, if the problem were and Professor Michael Lawson for his helpful comments on an earlier
3 + 5, the child using a sum strategy would count all version of the article.

215
216   Journal of Learning Disabilities

associations around the correct answer; the more peaked with special needs to help them acquire certain knowledge
the distribution, the greater the probability the answer and skills associated with arithmetic (including preparatory
will be directly retrieved. Thus, correct practice is understandings, skills performing the four operations,
thought to be the main pedagogical requirement for and problem-solving strategies). Of the 58 interventions
students to move on from counting to direct retrieval. included in the meta-analysis (published between 1985
Even among adults, simple addition problems are not and 2000) and of the 29 studies that focused on skills with
all solved using direct retrieval; generally they are solved the operation of addition and presupposed simple addition
using a combination of direct retrieval and decomposition skill, only 5 intervention studies actually focused on
strategies (Geary & Wiley, 1991; LeFevre, Sadesky, & simple addition skill (i.e., the addition of two single-digit
Bisanz, 1996). Decomposition strategies involve applying numbers). These five studies are mentioned below along
“a known” fact to derive the answer to a problem [e.g., with the most recent intervention research relevant to the
9 + 7 = (10 + 7) – 1 or 3 + 4 = (3 + 3) + 1]. Although in retrieval of basic facts.
certain circumstances, adults may sometimes count on A number of interventions have focused on helping
two or more counts, they more than likely can retrieve students move on from using inefficient counting strategies
the answer and this means that they have at their disposal to using more efficient counting strategies for simple
a body of known facts that can be used to derive other addition, for example, Tournaki (2003). It is important to
facts. Take, for example, the problem 5 + 7: The answer note, however, that these studies did not focus on or assess
could be derived using the approach (5 + 5) + 2, but to if the interventions increased the frequency of retrieval.
successfully use this decomposition strategy, two facts Similarly, Schopman and Van Luit (1996) evaluated the
must be known, 5 + 5 and 5 + 2. effects of an intervention to move children on from using
A considerable body of research has identified a elaborate to more efficient counting strategies, and Miller
group of students who do not develop a reliance on direct and Mercer (1993) evaluated the effects of an instructional
retrieval and decomposition strategies but have to rely on sequence to help children move on from solving simple
counting for simple addition at an age well beyond their addition problems using concrete representations to
peers (Geary et al., 2000; Gray, 1991; Jordan & Montani, solving abstract sums. In neither study did the authors
1997; Ostad, 1997; Ostad & Sorensen, 2007). There is focus on, or assess, the frequency of retrieval.
some debate in the research literature as to whether these A number of interventions have focused on helping
students are delayed in developing simple addition skill, students learn an algorithm to perform addition problems
that is, they use immature counting strategies for an involving two-digit numbers. For example, Simon and
extended period of time but will eventually develop Hanrahan (2004) taught three students with learning
retrieval strategies (Torbeyns, Verschaffel, & Ghesquiere, disabilities (around 10 years old) to use a dot-notation
2004), or whether they have a processing deficit that method known as Touch Math. The children in this study
does not allow them to develop and/or access retrieval all used a combination of count-all and count-on strategies
links (Geary, 2004; Jordan & Montani, 1997). Independent when solving simple addition problems and relied on
of the debate in the literature, teachers are expected to concrete referents such as fingers or tallies. Although
improve procedural fluency of basic skills along with this system offered a method for solving double-digit
conceptual understanding for students with difficulties addition problems using a standard written algorithm, it
learning mathematics (Evans, 2007). was not designed to improve the retrieval of stored facts
While the number of studies that have examined for simple addition. In fact, these researchers assumed
typical and atypical development of simple addition skill that because students were encouraged to repeat their
is extensive (see Hopkins & Lawson, 2002, 2006b, for answers to problems aloud when using the Touch Math
reviews of this research), the number of intervention method, simple addition facts would gradually be
studies that specifically address retrieval difficulties retrieved; however, this assertion was not tested.
associated with simple addition is very limited. To Several interventions intended to increase accuracy
substantiate this point, a review of related intervention and automatic responding to basic number facts have
studies is provided in some detail below. involved memorization; however, memorization techni­
ques, including the Cover, Copy, and Compare (CCC)
procedure (Skinner, Turco, Beatty, & Rasavage, 1989)
A Review of Intervention Research
and the Taped Problems (TP) intervention (McCallum,
Kroesbergen and Van Luit (2003) conducted a meta- Skinner, & Hutchins, 2004), are more commonly
analysis of intervention studies, designed for children associated with learning multiplication and division facts.
Hopkins, Egeberg / Retrieval of Addition Facts   217  

A critical distinction can be made between retrieving the addend increases by 1, the sum increases by 1” (e.g.,
basic addition (and subtraction) facts and retrieving basic “2 + 4 = 6, 2 + 5 = 7, 2 + 6 = 8”). The second method
multiplication (and division) facts. Basic multiplication focused on rote memorization. The tutor would say, “2 +
facts are often taught by memorizing verbal associations 4 = 6; now you try: 2 + 4 = what?” Beirne-Smith reported
between problems and answers, keeping in mind that good that both memorization interventions improved these
practice is based on making sure that children are taught students’ acquisition of simple addition facts. Similarly,
to understand and develop number sense at the same Poncy, Skinner, and Jaspers (2007) found that both the
time as being taught to memorize verbal associations interventions of CCC and TP were effective in helping
(e.g., Gersten & Chard, 1999). Few simple addition facts one student (aged 10 years old), who was functioning at
are taught this way in the classroom. Retrieval links for a cognitive level below that of mild mental retardation,
simple addition are developed and strengthened via correct to acquire simple addition facts.
practice using a back up strategy (Shrager & Siegler, Although both studies report success using memoriza-
1998; Siegler & Jenkins, 1989; Siegler & Shrager, 1984). tion techniques for teaching simple addition facts, this
Memorization techniques may, therefore, not be the most approach may not be the most efficient or effective or the
appropriate way to teach simple addition facts to students most beneficial in allowing students to develop important
with learning difficulties, given that normally achieving conceptual understandings relating to number. Although
students do not learn and are not taught addition facts in this argument cannot be established from existing research,
this way. This argument is further supported by Cohen, it is well documented in the research literature (reviewed
Dehaene, Cohochon, Lehericy, and Naccache (2000), who in Hopkins & Lawson, 2002) that memorization is not
found neurological differences in the way multiplication/ the way addition facts are generally learned by typically
division facts and addition/subtraction facts are stored in achieving children.
memory and suggested that this was due to the different While we acknowledge that students with learning
processes involved in learning these facts. difficulties may well require alternative instruction, it is
At least three intervention studies have incorporated our contention that in the context of learning to retrieve
memorization techniques to teach simple addition facts simple addition facts, pedagogical approaches that more
to students with learning difficulties. A multimodal closely replicate typical development should be thoroughly
program was designed and tested with two students investigated first. For learning to retrieve simple addition
(Jones, Thornton, & Toohey, 1985) that incorporated facts, typical development is based on correct practice at
memorization techniques for learning double facts. It is applying an efficient back up strategy. To date, intervention
important to note that the multimodal program still studies that have been based on practice have produced
required students to count on problems with minimum unclear results.
addends of 1, 2, and 3. This point is highlighted because Two of the five studies included in the meta-analysis
more recent research has indicated that learning to by Kroesbergen and Van Luit (2003) examined the
retrieve answers to problems with minimum addends of effectiveness of extended practice for developing simple
1, 2, or 3 is a critical stage in development (Hopkins & addition skill among elementary school students (aged
Lawson, 2006a), as these facts become the basis for around 8 years old), with and without mild mental
decomposition strategies. It is precisely the failure to handicaps. Lin, Podell, and Tournaki-Rein (1994) and
move on from counting to retrieval for problems with Podell, Tournaki-Rein, and Lin (1992) examined the
smaller addends that appears to be the most distinguishing effectiveness of computer-assisted instruction (CAI)
characteristic of a retrieval difficulty. Thus, the multimodal compared with a more traditional paper-and-pencil
program did not address what might be considered the method. Their findings were reported to indicate that
crux of this problem. both methods for providing opportunities for practice
Two other published studies were found in our review were effective at increasing accuracy and reducing
that focused on students memorizing all the simple reaction times (RTs) for both groups of students but that
addition facts. Beirne-Smith (1991) evaluated the students with mild mental handicaps required more
effectiveness of a rule-based memorization approach and practice than their peers. It is interesting that Podell et al.
a rote-based memorization approach with 20 primary- (1992) noted that the question of how much more
aged students with learning disabilities (ages ranging practice is required for these students could not be
from 6 to 10 years old). For the rule-based memorization addressed as these students did not meet the speed-
technique, the tutor displayed simple addition equations mastery criterion during the study: “Our results indicate
with the same addends and stated the rule, “Each time that the difficulty which students who are mildly mentally
218   Journal of Learning Disabilities

handicapped have in developing automaticity relates addition performance by students with learning difficulties:
specifically to lack of speed, rather than accuracy. What Students with learning difficulties produce variations in
remains unresolved, however, is the extent to which RTs to min-counting trials that are not well explained by a
further practice will be beneficial” (p. 205). It is also variable representing the number of counts made (Hopkins
important to note that although improved accuracy and a & Lawson, 2006b). This unexplained variability has been
reduction in RTs may be indicators of performance attributed to (a) interference in working memory from a
becoming more automated, they do not necessarily possible bottleneck effect relating to strategy selection, (b)
indicate an increased use of retrieval. It is likely that with interference from other sources in working memory, and
practice, the min-counting strategy becomes more (c) losing track of the count (Hopkins & Lawson, 2006b).
automated (i.e., performed more quickly) and so a A combined methodological approach makes it possible to
reduction in RTs by itself may not necessarily indicate an distinguish between a back up strategy that becomes more
increase in the use of retrieval. Similarly, a general automated and a strategy transformation where retrieval
reduction in RTs does not necessarily indicate other replaces the min-counting strategy.
types of strategy transformations, such as using count-all The number of intervention studies found that addressed
strategies to count-on strategies. To distinguish between retrieval difficulties associated with simple addition in our
a strategy becoming more automated and a strategy review of literature tallied only six: three studies employed
being transformed, a combined methodological approach memorization techniques (one study involved a single
is required, whereby trials are first separated by strategy student and another involved two students) and three
use based on self-report and observational data, and RTs studies employed extended practice. Five of the six studies
are then analyzed separately for different strategies involved students who were 10 years old or younger,
(Siegler, 1987). This approach makes it possible to three of the six studies involved students with an intellec­
distinguish between the strategy used and time taken to tual disability, and no study adopted the recommended
execute the strategy. combined methodological approach for studying simple
Goldman, Pellegrino, and Mertz (1988) also examined addition performance. In summary, intervention research
the effects of extended practice for a group of students relevant to improving retrieval of simple addition facts, in
with learning disabilities in a special education setting particular for adolescent students in mainstream education
(aged 7 to 13 years old). They used discriminant settings, is sorely lacking.
analysis to infer strategy use from RT patterns. Their
findings indicated that for two thirds of participants,
The Present Study
practice resulted in a decrease in the mixture of counting
strategies and an increase in the use of direct retrieval. The present study employed a combined methodological
For one third of participants, variations in RTs were not approach to examine the effect that extended practice had
consistent with any typical pattern of strategy use and on increasing a reliance on retrieval for simple addition for
these students did not appear to benefit from extended low-performing adolescent students attending mainstream
practice. They also uncovered puzzling RT patterns mathematics classes. In addition to examining the effect of
and cited similar findings from an earlier study by extended practice, an intervention aimed at reducing RTs
Svenson and Broquist (1975), who reported that perfor­ to min-counting trials was also incorpo­rated into the study
mance on simple addition problems with an addend of design. This decision was based on emerging evidence
2 or 3 produced unsystematic RTs among children with that practice using a min-counting strategy will be less
“poor mathematical talent.” These anomalous RT patterns effective for students with slower RTs. Hopkins and
could not be investigated further as the researchers did Lawson (2006a) found that a group of adolescent students
not employ a combined methodological approach. still reliant on counting for simple addition were not
A combined methodological approach to study simple completing 2, 3, and 4 counts within the span of working
addition performance was first advocated by Siegler (1987) memory (depicted to be between 2 and 3 seconds). Their
and has been adopted in studies examining differences in findings supported the argument put forward by Geary
performance between students with and without learning (1993) that problem-answer associations will only
differences by Geary and his colleagues (Geary, 1990; strengthen with practice if both the problem and answer
Geary & Brown, 1991; Geary, Brown, & Samaranayake, are activated in working memory at the same time. If
1991; Geary et al., 2000). Using a combined methodological practice does not strengthen retrieval links for problems
approach, unusual patterns of RTs noted by Goldman et al. with smaller addends, then a body of known facts cannot
have since become a well-documented feature of simple later be relied on to derive answers to other problems.
Hopkins, Egeberg / Retrieval of Addition Facts   219  

There are different reasons why RTs to min-counting The term extended practice denotes providing more con-
trials may be slower for some students. These students centrated and frequent practice than would typically be
may be slower to make each count (Geary et al., 2000), or encountered. In this study, extended practice encom-
they may lose track of the count and have to redo certain passed 15 minutes practice each day for 20 consecutive
steps in the counting procedure (Hopkins & Lawson, school days, an amount comparable to that reported in
2006a). Students with mathematical disabilities display other intervention studies (Goldman et al., 1988; Lin
less use of private speech while counting to perform et al., 1994; Podell et al., 1992). It was hypothesized that
simple addition (Ostad & Sorensen, 2007) and this factor practice alone would lead to some improvement in the
could explain why some students are more prone to frequency of retrieval (based on Siegler’s distributions of
losing track of the count. It has also been proposed that association model) but that the intervention would
interference from a possible bottleneck effect relating to increase the effectiveness of practice based on findings
strategy selection could produce slower RTs (Hopkins indicating that a slower counting speed impedes the
& Lawson, 2006b). The intervention phase of this study effectiveness that practice has on developing a reliance
attempted to decrease RTs to min-counting trials by on retrieval (Hopkins & Lawson, 2006a).
requiring students to (a) count as quickly as possible,
(b) count aloud to help avoid losing track of the count,
and (c) always use the min-counting strategy to help Method
reduce interference from making a strategy selection.
Given that a combined methodological approach requires Participants
intensive trial-by-trial analysis—where the researcher Eight students identified as having a learning difficulty
observes students individually as they complete each were selected to participate. It should be noted that in
problem and students give a self-report of strategy use after Western Australia, and more generally in Australia,
completing each problem—the research questions were students are not identified as having a learning disability
addressed using a single case study design. Other researchers based on an IQ-achievement discrepancy (Klassen,
have used larger sample sizes to assess the effect of Neufeld, & Munro, 2005; van Kraayenoord & Elkins,
extended practice on simple addition skill (Goldman et al., 2004). The term learning difficulty rather than learning
1988; Lin et al., 1994; Podell et al., 1992); however, these disability is used in the Australian context: A learning
studies were critiqued earlier in this article for not using a difficulty is generally identified using professional
combined methodological approach. A single case study judgments based on indicators of low achievement; the
approach has been used to assess the effectiveness of term learning disability (or specific learning difficulty) is
interventions for simple addition skill with students with generally used to denote a more severe (sometimes
intellectual disabilities (Poncy et al., 2007) and learning unexplained) type of learning difficulty (van Kraayenoord
difficulties (Jones, Thornton, & Toohey, 1985). A single & Elkins, 2004). In Western Australia at least, there is
case study approach has also been used to examine the considerable resistance to IQ testing except for identifying
development of simple addition skill among normally an intellectual disability (usually among young children),
achieving preschool children (Siegler & Jenkins, 1989). to allow parents a choice of special or mainstream
Given that a decomposition strategy relies on directly education services.
retrieving a known fact to then derive the answer, the The eight participants were in their 1st year of
phrase a reliance on retrieval is used to denote the use of secondary school (Year 8), in the same senior high
either direct retrieval or decomposition strategies. Two school located in an area of lower socioeconomic status
research questions were addressed in this study using a in the Perth metropolitan area, Western Australia. These
single case study approach: students attended a regular mathematics class but were
considered to have a mathematical learning difficulty
Research Question 1: To what extent does extended prac-
tice produce an increased reliance on retrieval for simple based on low scores attained in Year 7 on a statewide
addition among low-performing Year 8 mathematics standardized test of numeracy achievement, known as
students? the numeracy component of the Western Australian
Research Question 2: Does an intervention designed to Literacy and Numeracy Assessment (WALNA). Scores
reduce the time taken to execute a min-counting strategy for this assessment (referred to as WAMSES) range from
improve the efficiency of extended practice on develop- 0 to 800. A numeracy benchmark score of 402 in Year 7
ing a reliance on retrieval? approximates the 20th percentile and reflects the
220   Journal of Learning Disabilities

Table 1
Participant Details
Age in Mathematics WMTB-C Component Score
Years/ Achievement SPM
Name Gender Months WAMSES (z score) Raw Score CE PL VS

Abbey female 13/3 380 (–1.136) 46   91 132   75


Amelia female   14/11 330 (–1.753) 28 102 120   82
Caleb male 13/0 405 (–0.827) 44 109 106   94
Henry male   14/10 389 (–1.025) 32   73 110   84
Samuel male 14/8 341 (–1.617) 13   56 110   74
Sue female 14/3 397 (–0.926) 33   61   97   71
Tiana female 13/0 352 (–1.481) 37   78 119   84
Zara female 13/3 389 (–1.025) 43   93 111   88

Note: A cell shaded gray indicates a score below the 10th percentile of population scores recorded for that measure. All mathematics achieve-
ment scores for study participants were around or below the population benchmark score of 402 (around or below the 20th percentile). z scores
for the Western Australian Literacy and Numeracy Assessment test of Mathematics Achievement were calculated using 2007 population figures
(M = 472, SD = 81). The median score for the SPM, standardized for similar aged children in Australia, was reported to be between 45 and 47
(de Lemos, 1989, cited in Raven, 2000a), with a score around 36 representing the 10th percentile. WMTB-C = Working Memory Test Battery
for Children; WAMSES = Western Australia Monitoring Standards in Education scores; SPM = Standard Progressive Matrices Test; CE =
Central Executive; PL = Phonological Loop; VS = Visuo-Spatial Sketchpad.

minimum standard of performance that a student must but was not at the time teaching any Year 8 mathematics
achieve to progress. For ease of interpretation, WAMSE classes. Students who gave their signed consent, along
scores were converted to z scores. with their parents’ signed consent, and whose records
Consent was given from the school (then the parents indicated that they did not exceed the Year 7 numeracy
and students) to administer the Standard Progressive benchmark, were selected for inclusion in the study.
Matrices Test (SPM; Raven, 2000a). Consent was also Students from Aboriginal and Torres Strait Islander
given to administer the Working Memory Test Battery for backgrounds were not included in the study. In total,
Children (WMTB-C; Pickering & Gathercole, 2001). 24 out of 109 Year 8 students in the school had not met
Geary, Hoard, Byrd-Craven, Nugent, and Numtee (2007) the numeracy benchmark in Year 7. Four students were
found that different components of working memory not able to participate because they were not attending
affect different aspects of mathematical performance. It school at the time. Out of the possible 20 participants,
was deemed important to have measures of working 3 students and/or their parents did not give consent
memory components for each participant to explain and 5 consent forms were not returned. Twelve students
possible inconsistencies in findings. elected to participate in the screening stage of the study.
Participant details are displayed in Table 1. All The aim of the screening stage was to identify students
participants were nonindigenous Australians attending who had not developed a reliance on retrieval for simple
mainstream mathematics classes in their 1st year of addition. It involved each of the 12 students meeting
secondary school (Year 8) and all had performed around individually with the second author for 20 minutes each
or below the 20th percentile on a standardized test of day, for 5 consecutive school days (where possible).
mathematics achievement administered toward the end of Illness and various individual timetable constraints were
their last year of primary school (Year 7). reasons for this not occurring on consecutive days. All
meeting times were arranged for the morning. During the
meeting time, individual students sat next to the researcher
Procedure
and completed a set of 36 simple addition problems
Permission was obtained from the university’s Human presented using a specially designed computer program
Research Ethics Committee to commence the study. The (described later in the Instruments section). Thirty-six
principal of the school was approached and informed of problems comprised the problem set, which included all
the study. Permission from the principal was given to single-digit simple addition problems presented in the
send a letter home inviting Year 8 students to partici­pate form m + n, where m ≤ n.
in a study investigating retrieval difficulties in The simple addition performance of each student in
mathematics. The second author worked at the school the screening stage was analyzed using a combined
Hopkins, Egeberg / Retrieval of Addition Facts   221  

methodological approach. The number of problems that associated with Space, Measurement, Chance and Data,
were retrieved or derived was calculated for each of the five Number, Pre-algebra, and Working Mathematically. It
time intervals. Students were screened out if during one or requires approximately 45 minutes for students to com-
more time intervals, the number of answers retrieved or plete and includes multiple-choice questions, questions
derived exceeded 90% (i.e., more than 33 out of 36 requiring short answers, and open-ended questions.
problems). The mean RT taken by each student to solve Results from testing in Years 3, 5, 7, and 9 are calibrated
problems with a minimum addend of 4 or less using a min- onto a scaled measure, based on the Rasch measurement
counting strategy was calculated for each of the 5 days. The model, referred to as the Western Australia Monitoring
grand mean (across the five time intervals) was then Standards in Education (WAMSE) scale. Scores (or
calculated for each student. Students were screened out if WAMSES) range from 0 to 800. The mean numeracy
the grand mean RT to correct min-counting trials for achievement score for the state’s population of Year 7
problems with smaller addends was less than 2.5 seconds. students in 2007 was 472 (SD = 81) (M. Thomson, per-
After the screening phase, eight students were left to sonal communication, June 3, 2008). A benchmark score
participate in the study. Based on their performance across of 402 approximates the 20th percentile.
5 days, all students were counting to perform more than
20% of simple addition problems and all produced RTs to Intelligence. Raven’s Progressive Matrices tests
min-counting trials that suggested that practice would be (including Standard, Colored, and Advanced Progressive
less effective in improving the likelihood of retrieval Matrices tests) are widely used nonverbal tests of general
(based on findings from Hopkins & Lawson, 2006a). cognitive ability and intelligence (Marshalek, Lohman, &
The study spanned 1 month (20 school days), and on Snow, 1983; Raven, 2000b). Items measure the intellec-
each school day, a set of 36 simple addition problems was tual efficiency with which a person can make comparisons
completed individually by each student. The study design between figures and develop a logical method of reason-
consisted of one phase (labeled an extended baseline phase) ing to choose an appropriate missing piece to a pattern.
for two students and three phases for the other six students:
(a) the baseline phase, which comprised 6, 8, or 10 days and Working memory. The WMBT-C provides capacity mea-
included data collected in the screening stage; (b) the sures for three working memory components: the
intervention phase, which comprised 4 days; and (c) the Phonological Loop (comprising three subtests: Digit Recall,
return to baseline phase, which comprised 10, 8, or 6 Word List Recall, and Nonword List Recall), the Central
days, respectively. As indicated in Table 1, four participants Executive (comprising three subtests: Listening Recall,
scored low for the central executive (CE) component of Counting Recall, and Backward Digit Recall), and the
working memory and four scored within the average range. Visuo-Spatial Sketchpad (comprising two subtests: Block
Students were allocated to each phase so that the same Recall and Mazes Memory Task). The test battery is
amount of practice in the baseline or extended baseline designed to be used with students aged between 5 and 15
phase was completed by a student with a low CE score years old. Test-retest reliability was calculated for each sub-
(referred to as having a central executive difficulty) and a test using Pearson’s product moment correlation coefficient
student with an average CE score. and was found to range from 0.42 to 0.83 (Pickering &
The intervention phase involved requiring students to Gathercole, 2001). Construct validity was tested and found
always use the min-counting strategy to perform each to be high for the Central Executive and Phonological Loop
problem in the set and to verbalize each count aloud. components but lower for the Visuo-Spatial Sketchpad
Students were also instructed (and reminded) to count as fast (Gathercole & Pickering, 2000). Based on these tests of
as they could during the intervention phase. The goal of the reliability and validity, the WMTB-C has been found suit-
intervention was to decrease overall RTs to min-counting able for measuring different components of working mem-
trials by increasing counting speed and by minimizing RT ory in a variety of studies, including ones examining
variations. During the return to baseline phase, students were mathematical learning and performance (e.g., Fuchs,
again instructed to use whichever strategy they wanted. Compton, Fuchs, Paulsen, Bryant, & Hamlett, 2005; Geary
et al., 2007; McKenzie, Bull, & Gray, 2003).
Instruments
Simple addition skill. A computer program designed
Mathematics achievement. The WALNA is adminis- specifically for the study was used to assess simple addi-
tered each year to all Year 7 school students in the state. tion skill. The program presented a problem set, which
The numeracy component of the assessment provides included all single-digit addition problems with addends
a measure of mathematics achievement for outcomes greater than 1 (2 + 2 to 9 + 9) written in the form “m + n =.”
222   Journal of Learning Disabilities

Problems were presented with the smaller addend first for use with shorter autocorrelated series, where the number
(e.g., 3 + 5 =) so the experimenter was able to distinguish of data points per phase can be as low as 5, although 10 or
between use of a min-counting strategy (where 3 is more is recommended (Crosbie, 1993). This procedure
counted on) and a less mature count-from-first strategy uses least-squares estimates of intercepts and slopes for
(where 5 is counted on). Problems with an addend of 1 each phase (while controlling for autocorrelation) and then
were not included due to difficulties in self-reporting assesses overall change in intercepts or slopes using an
whether a retrieval strategy was used or a count of 1 was F test supplemented by t tests. The program DMITSA
made. In the baseline and return to baseline phases, and (Crosbie & Sharply, 1991) was used to perform the
the extended baseline phase, students were instructed that ITSACORR procedure in the present study to help assess
accuracy was important but that they could use whatever if the intervention improved the effectiveness of practice at
strategy they wanted to use to perform each problem. developing a reliance on retrieval for each student.
The space bar activated the timer and at the same time The use of inferential statistics with a single case study
the problem were displayed. A bell was also sounded to design is not common but has support in different fields of
alert the student that a problem was being presented. A psychology (e.g., Busse, Kratochwill, & Elliot, 1995;
second press of the space bar stopped the timer and Crawford, Garthwaite, & Gray, 2003; Crawford, Garthwaite,
removed the problem from the screen. The researcher Howell, & Venneri, 2003; Lundervold & Belwood, 2000).
controlled the timer by immediately pressing the space bar The study of change based on a single case study design is
when the student called out the answer. After the student typically evaluated using visual inspection (whereby a
entered the answer into the computer, a screen appeared change is considered significant if it is obvious), but visual
asking the student, “How did you do it?” A number of inspection alone can be lenient and unreliable. Inferential
options were presented to the student: “I counted,” “I just statistics are used in this study to augment visual analyses,
knew it,” “I did something else” (indicating decomposition), an approach recommended by Busse et al. (1995).
and “I don’t know.” If the self-report of strategy use
conflicted with the behavior observed by the researcher,
Setting
the researcher probed further until agreement was reached.
The researcher pressed the space bar and the next problem All phases were carried out in a two-room office
was presented. During the intervention phase, students located within the school. Participants were invited to the
were not prompted to indicate what strategy they had office in small groups and were provided with
used. This approach of identifying strategy use for simple refreshments while one student worked individually with
addition using self-report and observation has been the researcher in the meeting room. The computer on
adopted by Geary and his colleagues (Geary, 1990; Geary which students worked was situated on a table.
& Brown, 1991; Geary et al., 1991; Geary et al., 2000) as
well as others (e.g., Canobi, Reeve, & Pattison, 1998). The
accuracy of using self-reports to identify strategy use for Results
mental addition has been confirmed by Siegler (1987) and
LeFevre et al. (1996). The numbers of problems correctly retrieved (or derived)
during each time interval, for each phase, are displayed in
Figure 1a for students with CE difficulties and Figure 1b for
Data Analysis
students without CE difficulties. Time Intervals 1 through 5
To evaluate the effectiveness of extended practice on occurred (where possible) on consecutive school days and
developing a reliance on retrieval, visual inspection Time Intervals 6 through 20 also occurred (where possible)
using a multiple baseline design was combined with on consecutive school days.
regression analyses, to provide an estimate for each A multiple-baseline-across-participants design for
participant of the increase in the number of problems each student group was chosen to highlight differences
retrieved or derived each time a problem set was practiced. that might be observed in the number of problems
To evaluate if the intervention improved the effectiveness directly retrieved or derived as a result of practice. The
of practice on developing a reliance on retrieval, an results are clouded somewhat by the peculiar performance
interrupted time series analysis (ITSA) of data for each of Henry (Figure 1a) in the extended baseline condition,
participant was conducted. This is a recommended procedure where the number of problems he retrieved each day was
for single case studies and is based on performing a t test to highly variable and not influenced by practice, and by
assess change after controlling for autocorrelation to reduce Amelia (Figure 1b), also in the extended baseline
the probability of Type I error (Crosbie, 1993). The ITSA condition, who displayed the most notable improvement
procedure known as ITSACORR was developed specifically in retrieval with practice among all participants.
Hopkins, Egeberg / Retrieval of Addition Facts   223  

Figure 1a
Number of Problems Retrieved or Derived During Baseline Phase, Intervention Phase, and Return to
Baseline Phase, by Students With Central Executive Difficulties

Note: Only Sue used decomposition strategies in this group so that the numbers represent problems directly retrieved for Tiana, Samuel, and
Henry.

In general, improvements in the number of problems increase by less than one problem for four students (0.2
directly retrieved or derived are visually noticeable in for Tiana, 0.3 for Zara, 0.4 for Sue, and 0.7 for Samuel)
Figures 1a and 1b. No retrieval data are recorded for and between one and two problems for two students (1.1
students in the intervention phase as here they were for Caleb and 1.7 for Amelia).
required to always use the min-counting strategy. A single To assess whether the intervention improved the
regression line was fitted to the data presented in Figures effectiveness of practice for developing a reliance on
1a and 1b, for each student, and the number of problems retrieval, an individual test of significance was applied to the
during the intervention phase was treated as missing data. data displayed in Figures 1a and 1b. This involved fitting a
The intercept value, regression coefficient, and correlation regression line to the baseline data and another to the return
coefficient are displayed in Table 2. The results indicate to baseline data, for each of the six students involved in the
that for six of the eight participants, the amount of practice intervention phase, and testing for significant differences in
completed did explain a significant amount of variation in intercept and slope using the program DMITSA (Crosbie &
the number of problems retrieved (directly or derived). Sharply, 1991). The results indicated that among the three
The coefficient for the variable Time can be interpreted students in the CE difficulties group, the intervention did
to indicate that, for each consecutive school day when a not improve the effectiveness of practice for developing a
set of 36 simple addition problems is practiced, the reliance on retrieval: Sue, F(2, 10) = 1.03, p = .392; Tiana,
number of problems retrieved (directly or derived) will F(2, 10) = 2.732, p = .061; Samuel, F(2, 10) = 0.633,
224   Journal of Learning Disabilities

Figure 1b
Number of Problems Retrieved or Derived by Students Without Central Executive Difficulties

Note: Only Abbey used decomposition strategies in this group so that the counts represent problems directly retrieved for Caleb, Zara, and
Amelia. Caleb and Amelia were absent for consecutive days at the end of the school’s term and therefore could not complete all 20 problem
sets during the intervention phase before a 2-week school holiday. These data were treated as missing data in the following analyses.

Table 2 p = .551. For the group without CE difficulties, the


Summary of Regression Analyses for the Variable intervention did produce a significant improvement in the
(Time) Predicting the Number of Problems effectiveness of practice for one student, Caleb, F(2, 5) =
Retrieved or Derived 7.584, p = .031, but not for the other two students: Abbey,
F(2, 8) = 2.626, p = .133, or Zara, F(2, 10) = 0.898, p =
R F Coeff.
Group Squared Value df Sig. Constant (Time) .438. For Caleb, the difference in intercept approached
significance, T(5) = 2.346, p = .07, but the difference in
Group 1 slope was not significant, T(5) = 0.680, p = .527.
   Sue .524 15.387 15 .002 19.939 0.420
These findings indicate that the intervention did not
   Tiana .266 5.067 15 .041 4.898 0.212
   Samuel .571 18.623 15 .001 17.766 0.680 bring about significant improvements in the effectiveness
   Henry .221 5.116 19 .036 17.325 –0.552 at which correct practice leads to a reliance on retrieval
Group 2 for five of the six students involved in the intervention.
   Caleb .832 44.690 10 .000 19.457 1.104 Although there could be a number of reasons that the
   Abbey .195 2.902 13 .114 29.694 –0.269 intervention did not result in significant changes, the
   Zara .382 8.637 15 .011 20.953 0.267
   Amelia .782 57.498 17 .000 8.575 1.682
most obvious to be considered is whether or not RTs to
min-counting trials were reduced during the intervention
Hopkins, Egeberg / Retrieval of Addition Facts   225  

Figure 2a
Reaction Times to Correct Min-Counting Trials for Students With Central Executive Difficulties

Note: The R Squared Linear values indicate the percentage of variance in reaction times explained by the min variable (i.e., the number of counts
made).

phase. This question was examined by comparing RTs to reliance on retrieval for simple addition among low-
min-counting trials at each of the different phases, for performing, secondary school mathematics students
each student (see Figures 2a and 2b). revealed a complicated pattern of effects. The capacity
By observation, the pattern of RTs displayed in Figures of working memory components, including CE function,
2a and 2b indicates that the time taken to perform simple did not appear to be salient in explaining different effects
addition problems using the min-counting strategy was (e.g., Henry did not have a CE difficulty but did not
generally shorter during the intervention phase than the benefit from extended practice), nor did approximate
other phases, but this reduction in time was due to less measures of IQ (e.g., Samuel scored particularly low on
variation in times (i.e., a smaller range of RTs produced Raven’s Matrices Test but appeared to benefit from
when making the same number of counts) rather than an extended practice). Clearly, further research is needed to
increase in counting speed per se. The finding that RT identify the individual factors that are influencing
variability was reduced during the intervention phase is differential effects of extended practice on retrieval and
supported by an increase in the amount of variance simple addition skill.
explained by the number of counts required (see Figures
2a and 2b). The one exception was Caleb. During the Discussion
baseline phase, Caleb’s pattern of RTs to min-counting
trials was initially well explained by the number of counts This study examined the effect that extended practice
required (approximately 79% of variance was explained). had on developing a reliance of retrieval for simple
Overall, the intervention did not sufficiently decrease RTs addition, for low-achieving Year 8 mathematics students
to min-counting trials requiring 2, 3, or 4 counts, as often attending mainstream schooling, and whether an
these problems were still not being solved within 2.5 intervention designed to decrease the time taken to
seconds (representing the span of working memory) perform the min-counting strategy would improve the
during the intervention phase. effectiveness of extended practice.
Applying a single case study design to examine the The results of the study indicated that, in general,
effectiveness of extended practice on developing a extended practice did lead to a greater reliance on retrieval
226   Journal of Learning Disabilities

Figure 2b
Reaction Times to Correct Min-Counting Trials for Students Without Central Executive Difficulties

for simple addition but supports the argument that extended refute the argument that an increase in counting speed or
practice alone can be inefficient in improving a reliance on overall decrease in RTs to min-counting trials will
retrieval (Hopkins & Lawson, 2006a). Given that the improve the effectiveness of practice because the inter­
participants were in the 1st year of secondary school, and vention did not produce sufficient increases in counting
simple addition skill is generally thought to be mastered in speed or reductions in RTs to test this argument. RTs to
the early years of primary school, the time available or min-counting trials requiring 2, 3, or 4 counts were often
deemed appropriate for practice of simple addition greater than 2.5 seconds (representing the span of
problems is likely to be restricted. For example, working memory) during the intervention phase.
extrapolating the findings for one student (Tiana) indicates Until further evidence suggests otherwise, it seems
that she would need to complete a practice set of 36 simple prudent to persevere with the explanation that counting
addition problems every day for more than 6 months speed influences the effectiveness of correct practice;
before a complete reliance on retrieval for simple addition however, it may not be the only influence. Variability in
would be achieved. Allocating this amount of time to RTs to min-counting trials may also influence the
practicing simple addition problems appears inappropriate. effectiveness of correct practice. Caleb was the only
The findings are not consistent, however, as extended student who during the baseline phase displayed a RT
practice by itself appeared to be very effective in increasing pattern to min-counting trials that was well explained by
a reliance on retrieval for one student (Amelia), even the number of counts required. Therefore, he could have
though she displayed slow and highly variable RTs to min- benefited more than the other students from practice
counting trials. Complicating the picture was the finding during the baseline stage. This would explain why the
that extended practice alone was ineffectual at increasing intervention was effective for him and not the other
a reliance on retrieval for another student (Henry). students. Consistency has been identified as an important
The results also suggest that the intervention did not condition for developing intratask automaticity (Brown
bring about significant improvements in the effective­ness & Carr, 1989). Further research is needed to clarify what
at which correct practice leads to an increased reliance on factors are necessary for correct practice at applying the
retrieval for simple addition (with the exception of one min-counting strategy to efficiently lead to a reliance on
student, Caleb). The results cannot be used to support or retrieval for simple addition.
Hopkins, Egeberg / Retrieval of Addition Facts   227  

Limitations and Further Research assess the effectiveness of practice alone and practice with
an intervention, to address retrieval difficulties for simple
The present study comprised a small sample, thereby
addition. Most important, it encompassed a combined
limiting the potential to find generalized patterns of
methodological approach whereby strategy use was first
performance. It is our assertion, however, that before further
identified by self-report and observation, and RTs were
interven­tions on a larger scale are trialed, further research is
analyzed separately for min-counting trials. This is the first
first needed to (a) better identify students who have a
time, to our knowledge, that a combined approach has
retrieval difficulty with simple addition, (b) pilot test
been applied to study the effects of practice on strategy use
interventions that are capable of sufficiently reducing the
and RT variability involved in the simple addition
time taken to solve smaller simple addition problems using
performance of students with this type of mathematical
the min-counting strategy, and (c) identify the individual
learning difficulty. Moreover, it is unique as it involved
factors that produce differential effects of extended practice.
students attending regular mathematics classes in a
The argument that single case study research that allows for
secondary school setting.
the examination of individual variations should be considered
The findings draw attention to the complexities
first and later followed by larger scale research to substantiate
involved in addressing retrieval difficulties for simple
predicted patterns of generalized behavior (often much later
addition. They highlight the continuing need to evaluate
when more is understood about the topic of investigation)
the effectiveness of interventions among individuals
has been propounded by others (e.g., Cronbach, 1975;
before trying to address retrieval difficulties on a larger
Thorngate, 1986). We suggest that intervention studies for
scale, not only for reasons of efficiency but also to
students with retrieval difficulties are particularly well
examine possible factors that account for poorer responses
suited for this type of bottom-up approach to research.
to intervention, such as processing speed and reaction
The intervention tested here, to improve the
time variability. This approach to research in the field is
effective­ness of extended practice on simple addition
salient given growing support for defining a learning
skill, could be deemed short as it spanned only four
disability in terms of intervention responsiveness (e.g.,
days; however, it should be kept in mind that there
Fuchs et al., 2005; Fuchs, Mock, Morgan, & Young,
were really two interventions, an intervention of
2003).
extended practice covering 20 days and within that,
an intervention designed to improve the effectiveness
of extended practice covering 4 days—this second References
intervention does lend itself to be thought of as a
pilot study. Time was a limitation as 20 consecutive Beirne-Smith, M. (1991). Peer tutoring in arithmetic for children with
school days (or 1 month) was deemed appropriate learning disabilities. Exceptional Children, 57, 330–337.
by the researchers and school involved to spend on Brown, T. L., & Carr, T. H. (1989). Automaticity in skill acquisition:
Mechanisms for reducing interference in concurrent performance.
practicing a skill in secondary school that is a focus of Journal of Experimental Psychology: Human Perception and
instruction in lower primary school. The methodology Performance, 15(4), 686–700.
employed, including a multiple baseline design and Busse, R. T., Kratochwill, T. R., & Elliott, S. N. (1995). Meta-analysis
individual test of significance requiring at least five for single-case consultation outcomes: Applications to research and
data points for the baseline and return to baseline practice. Journal of School Psychology, 33(4), 269–285.
Canobi, K. H., Reeve, R. A., & Pattison, P. E. (1998). The role of
phases, restricted the length of the intervention phase.
conceptual understanding in children’s addition problem solving.
An improvement in the effectiveness of extended Developmental Psychology, 34(5), 882–891.
practice may have been observed if the intervention Crawford, J. R., Garthwaite, P. H., & Gray, C. D. (2003). Wanted:
phase was longer; this warrants further investigation, Fully operational definitions of dissociations in single-case stud-
although the practicality of spending more time in the ies. Cortex, 39, 357–370.
classroom to improve simple addition skill will need Crawford, J. R., Garthwaite, P. H., Howell, D. C., & Venneri, A. (2003).
Intra-individual measures of association in neuropsychology:
to be considered. Inferential methods for comparing a single case with a control or
normative sample. Journal of the International Neuropsychological
Society, 9, 989–1000.
Conclusions Cronbach, L. J. (1975). Beyond the two disciplines of scientific psy-
chology. American Psychologist, 30(2), 116–127.
This study incorporated the use of observation and a Cohen, L., Dehaene, S., Cohochon, F., Lehericy, S., & Naccache,
multiple baseline research design as well as statistics to L. (2000). Language and calculation within the parietal lobe: A
228   Journal of Learning Disabilities

c­ ombined cognitive, anatomical and MRI study. Neuropsychologia, uncovered in the study of simple addition and the moving-on pro-
38, 1426–1440. cess. Educational Psychology Review, 14, 121–154.
Crosbie, J. (1993). Interrupted time-series analysis with brief single- Hopkins, S. L., & Lawson, M. J. (2006a). The effect counting speed
subject data. Journal of Consulting and Clinical Psychology, has on developing a reliance on retrieval in basic addition.
61(6), 966–974. Contemporary Educational Psychology, 31, 208–227.
Crosbie, J., & Sharply, C. (1991). DMITSA 2.0: A statistical program Hopkins, S. L., & Lawson, M. J. (2006b). Mathematical learning dif-
for analyzing data from interrupted time-series. Monash University, ficulties: The influence of working memory limitations on simple
Victoria, Australia. addition performance. In A. V. Mitel (Ed.), Trends in educational
Evans, D. (2007). Developing mathematical proficiency in the psychology research (pp. 73–105). New York: Nova.
Australian context: Implications for students with learning diffi- Jones, G. A., Thornton, C. A., & Toohey, M. A. (1985). A ­multi-option
culties. Journal of Learning Disabilities, 37(4), 420–426. program for learning basic addition facts: Case studies and an
Fuchs, L. S., Compton, D. L., Fuchs, D., Paulsen, K., Bryant, J. D., experimental report. Journal of Learning Disabilities, 18(6),
& Hamlett, C. L. (2005). The prevention, identification, and cog- 319–325.
nitive determinants of math difficulty. Journal of Educational Jordan, N. C., & Montani, T. O. (1997). Cognitive arithmetic and
Psychology, 97(3), 493–513. problem solving: A comparison of children with specific and gen-
Fuchs, D., Mock, D., Morgan, P. L., & Young, C. L. (2003). eral mathematics difficulties. Journal of Learning Disabilities, 30,
Responsiveness-to-intervention: Definitions, evidence, and impli- 624–634.
cations for the learning disabilities construct. Learning Disabilities Klassen, R. M., Neufeld, P., & Munro, F. (2005). When IQ is irrele-
Research & Practice, 18, 157–171. vant to the definition of learning disabilities. School Psychology
Gathercole, A., & Pickering, S. (2000). Assessment of working International, 26, 297–316.
memory in six- and seven-year-old children. Journal of Educational Kroesbergen, E. H., & Van Luit, J.E.H. (2003). Mathematics inter-
Psychology, 92(2), 377–390. ventions for children with special educational needs: A meta-
Geary, D. C. (1990). A componential analysis of early learning deficit analysis. Remedial and Special Education, 24, 97–114.
in mathematics. Journal of Experimental Child Psychology, 49(3), LeFevre, J., Sadesky, G. A., & Bisanz, J. (1996). Selection of proce-
363–383. dures in mental addition: Reassessing the problem size effect in
Geary, D. C. (1993). Mathematical disabilities: Cognitive, neuropsy- adults. Journal of Experimental Psychology: Learning, Memory
chological, and genetic components. Psychological Bulletin, and Cognition, 22(1), 216–230.
114(2), 345–362. Lin, A., Podell, D. M., & Tournaki-Rein, N. (1994). CAI and the
Geary, D. C. (2004). Mathematics and learning disabilities. Journal development of automaticity in mathematics skills in students
of Learning Disabilities, 37(1), 4–15. with and without mild mental handicaps. Computers in the
Geary, D. C., & Brown, S. C. (1991). Cognitive addition: Strategy Schools, 11, 43–58.
choice and speed of processing differences in gifted, normal, and Lundervold, D. A., & Belwood, M. F. (2000). The best kept secret in
mathematically disabled children. Developmental Psychology, counseling: single case (n = 1) experimental designs. Journal of
27(3), 398–406. Counseling and Development, 78(1), 92–102.
Geary, D. C., Brown, S. C., & Samaranayake, V. A. (1991). Cognitive Marshalek, B., Lohman, D. F., & Snow, R. E. (1983). The complexity
addition: A short longitudinal study of strategy choice and speed- continuum in the radex and hierarchical models of intelligence.
of-processing differences in normal and mathematically disabled Intelligence, 7, 107–127.
children. Developmental Psychology, 27(5), 787–797. McCallum, E., Skinner, C. H., & Hutchins, H. (2004). The taped-
Geary, D. C., Hamson, C. O., & Hoard, M. K. (2000). Numerical and problems intervention: Increasing division fact fluency using a
arithmetical cognition: A longitudinal study of process and con- low-tech self-managed time-delay intervention. Journal of Applied
cept deficits in learning disabled children. Journal of Experimental School Psychology, 20(2), 129–147.
Child Psychology, 77, 236–263. McKenzie, B., Bull, R., & Gray, C. (2003). The effects of phonologi-
Geary, D. C., Hoard, M. K., Byrd-Craven, J., Nugent, L., & Numtee, cal and visual-spatial interference on children’s arithmetical per-
C. (2007). Cognitive mechanisms underlying achievement deficits formance. Educational and Child Psychology, 20(3), 93–108.
in children with mathematical learning disability. Child Miller, S. P., & Mercer, C. D. (1993). Using data to learn concrete-
Development, 78(4), 1343–1359. semiconcrete-abstract instruction for students with math disabili-
Geary, D. C., & Wiley, J. G. (1991). Cognitive addition: Strategy ties. Learning Disabilities Research & Practice, 8, 89–96.
choice and speed-of-processing differences in young and elderly Ostad, S. A. (1997). Developmental differences in addition strategies: A
adults. Psychology and Ageing, 6(3), 474–483. comparison of mathematically disabled and mathematically normal
Gersten, R., & Chard, D. (1999). Number sense: Rethinking arithme- children. British Journal of Educational Psychology, 67, 345–357.
tic instruction for students with mathematical disabilities. The Ostad, S. A., & Sorensen, P. M. (2007). Private speech and strategy-
Journal of Special Education, 33(1), 18–28. use patterns: Bidirectional comparisons of children with and
Goldman, S. R., Pellegrino, J. W., & Mertz, D. L. (1988). Extended without mathematical difficulties in a developmental perspective.
practice of basic addition facts: Strategy changes in learning dis- Journal of Learning Disabilities, 40(1), 2–14.
abled students. Cognition and Instruction, 5(3), 223–265. Pickering, S., & Gathercole, S. (2001). The Working Memory Test
Gray, E. M. (1991). An analysis of diverging approaches to simple Battery for Children. United Kingdom: Harcourt Assessment.
arithmetic: Preference and its consequences. Educational Studies Podell, D. M., Tournaki-Rein, N., & Lin, A. (1992). Automatization
in Mathematics, 22(6), 551–574. of mathematics skills via computer-assisted instruction among
Hopkins, S. L., & Lawson, M. J. (2002). Explaining the acquisition students with mild mental handicaps. Education and Training in
of a complex skill: Methodological and theoretical considerations Mental Retardation, 27, 200–206.
Hopkins, Egeberg / Retrieval of Addition Facts   229  

Poncy, B. C., Skinner, C. H., & Jaspers, K. J. (2007). Evaluating and Svenson, O., & Broquist, S. (1975). Strategies for solving simple
comparing interventions designed to enhance math fact accuracy addition problems. Scandinavian Journal of Psychology, 16,
and fluency: Cover, copy, and compare versus taped problems. 143–151.
Journal of Behavioral Education, 16, 27–37. Thorngate, W. (1986). The production, detection, and explanation of
Raven, J. (2000a). Standard Progressive Matrices. United Kingdom: behavioral patterns. In J. Valsiner (Ed.), Individual subject and
OPP Limited. scientific psychology (pp. 71–96). New York: Plenum Press.
Raven, J. (2000b). The Raven’s Progressive Matrices: Change and Torbeyns, J., Verschaffel, L., & Ghesquiere, P. (2004). Strategy
stability over culture and time. Cognitive Psychology, 41, 1–48. development in children with mathematical disabilities: Insights
Schopman, E.A.M., & Van Luit, J.E.H. (1996). Learning and transfer from the choice/no-choice method and the chronological age/
of preparatory arithmetic strategies among young children with a ability/level-match design. Journal of Learning Disabilities, 37,
developmental lag. Journal of Cognitive Education, 5, 117–131. 119–131.
Shrager, J., & Siegler, R. S. (1998). SCADS: A model of children’s Tournaki, T. (2003). The differential effect of teaching addition
strategy choices and strategy discoveries. Psychological Science, through strategy instruction versus drill and practice to students
9(5), 405–410. with and without learning disabilities. Journal of Learning
Siegler, R. S. (1987). The perils of averaging data over strategies: An Disabilities, 36, 449–458.
example from children’s addition. Journal of Experimental van Kraayenoord, C. E., & Elkins, J. (2004). Learning difficulties in
Psychology: General, 116, 250–264. numeracy in Australia. Journal of Learning Disabilities, 37(1),
Siegler, R. S., & Jenkins, E. (1989). How children discover new strat- 32–41
egies. Hillsdale, NJ: Lawrence Erlbaum.
Siegler, R. S., & Shrager, J. (1984). Strategy choices in addition and Sarah Hopkins, PhD, lectures in the Graduate School of
subtraction: How do children know what to do? In C. Sophian Education at the University of Western Australia. She taught
(Ed.), Origins of cognitive skills (pp. 229–293). Hillsdale, NJ: secondary school mathematics for many years before becoming
Lawrence Erlbaum. an academic. Her research interests include the effective teach-
Simon, R., & Hanrahan, J. (2004). An evaluation of the Touch Math
ing of mathematics and mathematical learning difficulties.
method for teaching addition to students with learning disabilities
in mathematics. European Journal of Special Needs Education,
19(2), 191–209. Helen Egeberg is a research student in the Graduate School of
Skinner, C. H., Turco, T. L., Beatty, K. L., & Rasavage, C. (1989). Education at the University of Western Australia. She has
Cover, copy, and compare: An intervention for increasing extensive experience teaching secondary school mathematics
multiplication performance. School Psychology Review, 18, in government schools and now works in private practice as an
212–220. educational consultant.

You might also like