Contemporary Educational Psychology 46 (2016) 128–147

Contents lists available at ScienceDirect

Contemporary Educational Psychology

journal homepage: www.elsevier.com/locate/cedpsych

Empirical study

Mind mapping as a meta-learning strategy: Stimulating pre-adolescents’

text-learning strategies and performance?
Emmelien Merchie ⇑, Hilde Van Keer
Department of Educational Studies, Ghent University, Belgium

a r t i c l e i n f o a b s t r a c t

Article history: This study examined the effectiveness of two instructional approaches of mind mapping used as a meta-
Available online 18 May 2016 learning strategy to stimulate fifth- and sixth-graders’ text-learning strategies and recall performance.
Thirty-five fifth- and sixth-grade teachers and 644 students from 17 different elementary schools partic-
Keywords: ipated. A randomized quasi-experimental repeated measures design was set up with two experimental
Text-learning strategies conditions and one control condition. Students in the experimental conditions received a 10-week
Researcher-provided mind maps teacher-delivered instructional treatment, working with either researcher-provided or student-
Student-generated mind maps
generated mind maps. Multilevel piecewise growth analysis was used to examine the evolution in stu-
Elementary education
Intervention research
dents’ cognitive and metacognitive text-learning strategies, and free recall performance. Results show
Multilevel modeling the greatest gains from pre- to posttest and sustained effects from post- to retention test in observable
cognitive text-learning strategy use for students in the condition with researcher-provided mind maps.
These findings have direct implications for both research and practice. Challenges and facilitating factors
for school-based intervention research are discussed.
Ó 2016 Elsevier Inc. All rights reserved.

1. Introduction well-instructed use of graphic organizers or maps (e.g., Dexter &

Hughes, 2011; Kim, Vaughn, Wanzek, & Wei, 2004; Nesbit &
Effective strategies for text-based learning (i.e., processing, Adesope, 2006; Vekiri, 2002). These two-dimensional spatial
organizing, and acquiring knowledge from texts) are indispensable arrangements of words or word clusters (Stull & Mayer, 2007) carry
to manage the information glut in our 21st century (Alexander & great potential as an organizational learning strategy (Dansereau &
Res, 2012). These text-learning strategies become of particular Simpson, 2009; Farrand, Hussain, & Hennessy, 2002; Vekiri, 2002),
importance in middle and high school where informative texts evoking active knowledge transformation which has shown to be
are increasingly used to reach instructional objectives and the related with deeper-level text processing and learning (Nesbit &
expectations for independent text study increase substantially Adesope, 2006; Slotte & Lonka, 1999). Moreover, some researchers
(Broer, Aarnoutse, Kieviet, & Van Leeuwe, 2002; Meneghetti, De also consider graphic organizers as a meta-learning strategy,
Beni, & Cornoldi, 2007). Establishing a good study method by helping learners to learn meaningfully and independently, inducing
inducing and initiating a broad repertoire of text-learning strate- a larger strategy set (Chiou, 2008; Chmielewski & Dansereau, 1998).
gies arises therefore as an important educational goal in late ele- Although the effects of various types of graphic organizers have
mentary grades (McNamara, Ozuru, Best, & O’Reilly, 2007). In already been extensively studied (e.g., concept maps; Chiou, 2008;
this respect, students can strategically engage in a variety of Chularut & DeBacker, 2004; Novak, 2002), mind maps (Buzan,
observable and less-observable, deep- and surface level cognitive 1974, 2005) have received far less empirical attention in contem-
and metacognitive text-learning strategies during learning from porary educational intervention research. Disregarding the fre-
text to support the essential cognitive processes of text selection, quent use of mind maps in educational practice, their empirical
organization, transformation, and integration (Mayer, 2001; investigation in late elementary grades is underaddressed in the
Schlag & Ploetzner, 2011). current international research literature. Most published mind
To stimulate text-based learning, review studies have map research was executed in secondary or higher education and
specifically illustrated the beneficial effects of well-planned and has focused on mathematics (Brinkmann, 2003), science (Abi-El-
Mona & Adb-El-Khalick, 2010), or economics (Budd, 2004). The
⇑ Corresponding author at: Department of Educational Studies, Ghent University, very limited number of mind map studies in elementary education
Henri Dunantlaan 2, 9000 Ghent, Belgium. Fax: +32 331 03 20. have been mainly explorative by nature or focusing on teachers’
E-mail address: [email protected] (E. Merchie). views (Moi & Lian, 2007; Seyihoglu & Kartal, 2010). Further, only

http://dx.doi.org/10.1016/j.cedpsych.2016.05.005
0361-476X/Ó 2016 Elsevier Inc. All rights reserved.
 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147 129

a limited number of these studies have appeared in high-quality Dansereau, 1998) and mental imagery (Anderson & Hidde, 1971;
indexed journals the past fifteen year. Consequently, studies Leopold, Sumfleth, & Leutner, 2013) into maps. Gestalt principles
empirically addressing mind map effects and their instructional relate to the visual nearness of information by using similar colors
use in elementary school-based intervention research are very and shapes (‘equality’) or the grouping of related elements by
scarce. The present study aims to bridge this gap by explicitly means of their spatial arrangement (‘proximity’) (Wallace et al.,
focusing on mind mapping used as a meta-learning strategy to ini- 1998). Furthermore, certain keywords can be complemented with
tiate autonomous text-learning strategy use. Underneath, this aim a corresponding representation which enhances mental imagery
is elaborated on from a theoretical and empirical perspective. First, processes (Anderson & Hidde, 1971; Leopold et al., 2013). Including
the definition and effectiveness of graphic organizers, such as mind gestalt principles and mental imagery have found to promote fas-
maps, are described and it is clarified how mind mapping can be ter knowledge identification, processing, and retention (Anderson
considered as a meta-learning strategy. Second, we specifically & Hidde, 1971; Nesbit & Adesope, 2006; O’Donnell et al., 2002;
focus on how mind mapping can help students in their indepen- Wallace et al., 1998). Beside these design-related elements, also
dent text processing and learning. Next, after conceptualizing a the way in which the content is rearranged in mind maps might
mind map strategy instruction, two instructional mind map enhance text processing and learning. In this respect, informative
approaches are discussed. texts are succinctly summarized by hierarchically associating
super- and subordinate keywords or sentences to blanket terms
2. Mind mapping and meta-learning (i.e., the main ideas of the text). Various studies underpin the man-
ner of hierarchically associating keywords when learning from text
A mind map (MM) (Buzan, 1974, 2005) is a particular type of a (Hilbert & Renkl, 2008; Nesbit & Adesope, 2006; Stull & Mayer,
higher-level graphic organizer, transforming linear text into a 2007). Due to the incorporation of these specific elements (e.g.,
graphical representation (Fig. 1). As other types of graphic organiz- radial structure, mental imagery, gestalt principles) mind maps
ers (e.g., concept maps, knowledge maps), they are typified as fol- diverge from similar types of graphic organizers such as concept
lows: (a) they contain words or word groups, (b) the spatial maps, which are hierarchically top-down oriented, rely less on
arrangement between these elements indicate text relations, and specific design-principles (e.g., using colors or images), and require
(c) they represent the conceptual organization of a text (Stull & the explicit use of connective terms between concepts (Budd,
Mayer, 2007). A mind map is characterized by a radial structure, 2004; Davies, 2011).
that is, the linear text content is reorganized by placing the text’s Theoretically, the Dual Coding Theory (Paivio, 1991) and the
central theme in the middle of the page, from which several related Conjoint Retention Hypothesis (Kulhavy, Stock, Peterson,
main text ideas radiate out in the shape of colorful thick branches. Pridemore, & Klein, 1992; Robinson & Molina, 2002; Robinson,
Associated to these main branches, other smaller sub-branches Robinson, & Katayama, 1999) underpin the use of working with
represent subordinate text ideas. In this way, by using different graphic organizers or maps. Both theories, extensively discussed
dimensions (i.e., thick and smaller branches), a mind map reflects by Vekiri (2002) and Nesbit and Adesope (2006), emphasize the
the macro structure of the text together with more precise rela- importance of conjointly encoding information both non-verbally
tionships among related text units. Mind maps furthermore allow or spatially and verbally. By the interconnectedness of the verbal
the representations of more complex text relationships (e.g., con- and visual system, associative connections can be made and learn-
trasts, causations, comparisons) by adding numbers, images, ing is facilitated. Empirically, various meta-analyses, covering
arrows, or connectors to the branches (Buzan, 2005). The specific more than hundred studies, consistently confirm the effectiveness
mind map characteristics (e.g., radial structure, use of visual ima- of mapping as an organizational learning strategy, helping students
gery, Gestalt principles), supported by educational and brain to learn (e.g., Dexter & Hughes, 2011; Nesbit & Adesope, 2006;
research (e.g., Anderson & Hidde, 1971; Budd, 2004; Haber, 1970; Vekiri, 2002). Also mind mapping was already studied in higher
O’Donnell, Dansereau, & Hall, 2002), make mind maps significantly education as an organizational strategy to schematically represent
different from traditional temporarily ordered linear summaries. information (Zipp, Maher, & D’ Antoni, 2009) or as a prompted
More particularly, research identified aspects which are included study technique (Farrand et al., 2002). However, some researchers
in the design and content arrangement of mind maps, that can go beyond considering mind mapping as an organizational strategy
enhance text processing and learning. For example, beneficial and point at the potential of mapping methods for meta-learning
effects have been demonstrated for the incorporation of gestalt (Chiou, 2008; Chmielewski & Dansereau, 1998; Okebukola, 1992).
principles (O’Donnell et al., 2002; Wallace, West, Ware, & This term was originally introduced in educational research by

Fig. 1. Informative text paragraph and its graphical transformation into a mind map.
130 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147

Biggs (1985) and further elaborated by Jackson (2004), referring to graders’ spontaneous application of this strategy during autono-
a strategy helping students learn how to learn meaningfully and mous study. Second, working with mind maps can also trigger
independently. Chmielewski and Dansereau (1998), for instance, the use of metacognitive strategies during learning from text
found a positive transfer of text processing skills after having (Hilbert & Renkl, 2008; Nesbit & Adesope, 2006). Hilbert and
trained university students in using and constructing maps, even Renkl (2008) refer, in this respect, explicitly to the metacognitive
when the mapping was not explicitly applied during learning. Also function of mapping, by becoming aware of comprehension prob-
Chiou (2008) refers to concept mapping as a ‘meta-learning strat- lems during explicating or constructing maps. To induce metacog-
egy’, empowering students ‘learning how to learn’. Despite this nitive strategies (i.e., planning, monitoring, and evaluation),
potential, very few studies have investigated mapping as a means metacognitive prompts such as checklists or lesson reviews can
for meta-learning and no researchers specifically addressed the be included, helping learners to implement a systematic regulatory
value of mind mapping as a meta-learning strategy in elementary sequence controlling their performance (Askell-Williams, Lawson,
education. The present study therefore aims to renew the interest & Skrzypiec, 2012; Schraw, 1998). In this respect, mind mapping
on this topic and fill the gap in the current mind map research lit- can be viewed as a meta-learning strategy, helping students to
erature by thoroughly investigating how mind mapping can be implement metacognitive strategies (such as planning, monitoring
used to induce a larger text-learning strategy repertoire during and evaluating) during an independent learning task, wherein no
independent text learning. strategy prompts are provided. To date however, very little
research documents on the transfer of taught metacognitive strate-
gies to independent text-learning. Third, the induced text-learning
3. Using mind mapping as a meta-learning strategy to support strategies (such as actively reorganizing text information) by
learning from text means of the mapping training might be transferred to indepen-
dent learning tasks, affecting students’ text recall (Chmielewski &
A wide range of text-learning strategies, (e.g., highlighting, note Dansereau, 1998). In this respect, research point out that, rather
taking, rereading), have been reported in the literature to stimulate than surface-level or linear learning strategies (e.g., repeatedly
the essential processes of text selection, organization, and transfor- rereading, coping texts, trying to memorize text literally), deep-
mation, leading to text-integration into long-term memory (Mayer, level strategies (e.g., creation of summaries, maps) evoke the gen-
1996, 2001; Schlag & Ploetzner, 2011). These processes respectively eral capacity to analyze, structure, and organize knowledge which
refer to the selection of relevant words, the organization of this in turn promotes deep text processing and learning (Lahtinen et al.,
selected information into coherent mental representations, and 1997; Nesbit & Adesope, 2006; Ponce & Mayer, 2014; Schnotz,
transformation processes to generate a verbal and pictorial mental 2002). In this respect, it might be assumed that, students working
model which is integrated into long-term memory (Mayer, 2001). with mind maps in class, transfer the learnt deep-level strategies to
Students can engage in a variety of text-learning strategies, which an independent learning task, in which their application leads to
can been categorized according to their nature (e.g., cognitive and increased text recall. Unfortunately, students’ retention of text
metacognitive strategies) (Weinstein, Jung, & Acee, 2011), level of information has been mainly studied after being explicitly
depth (e.g., surface-level versus deep-level strategies) (Lahtinen, prompted to use maps as a learning tool (e.g., Chularut &
Lonka, & Lindblom-Ylänne, 1997), or perceptibility (e.g., overt DeBacker, 2004; Nesbit & Adesope, 2006). However, there are indi-
observable strategies or covert non-observable strategies) cations that mapping training can increase the recall of text ideas
(Kardash & Amlund, 1991). Previous research on pre-adolescents’ even without explicitly applying the mapping strategy
text-learning strategy use indeed identified in this respect various (Chmielewski & Dansereau, 1998), and that the use of maps results
overt and covert cognitive (e.g., summarizing, highlighting, para- in a positive transfer in learning situations were no maps are avail-
phrasing) and metacognitive text-learning strategies (e.g., planful able (Schnotz, 2002).
approach, monitoring, and self-evaluation) (Merchie & Van Keer,
2014; Merchie, Van Keer, & Vandevelde, 2014). Effective learning
from texts requires the application of these text-learning strategies 4. Conceptualizing a mind map strategy instruction
in a self-regulated, strategic, and goal-oriented way (Mayer, 2001;
Schlag & Ploetzner, 2011; Wade, Trathen, & Schraw, 1990). How- As mentioned above, the research literature points at the poten-
ever, recent findings show rather inefficient and superficial text- tial of mind mapping as a meta-learning strategy, helping students
learning strategy use in pre-adolescence (Merchie & Van Keer, ‘learn how to learn’. Therefore, a mind mapping strategy instruc-
2014; Merchie et al., 2014) and researchers call for longitudinal tion was conceptualized in the present study (Table 1), to induce
interventions inducing a large and flexible repertoire of diverse specific cognitive (i.e., organization, elaboration, and rehearsal)
strategies (Donker, de Boer, Kostons, Dignath van Ewijk, & van der (Weinstein et al., 2011) and metacognitive strategies (i.e., plan-
Werf, 2014; Simpson & Nist, 2000; Weinstein et al., 2011). Imple- ning, monitoring, and evaluation) (Schraw, 1998), as both cognitive
menting mind mapping as a meta-learning strategy seems promis- and metacognitive strategies should be the focus of instruction
ing to this end because of various reasons. when stimulating strategic learning (Simpson & Nist, 2000). These
First, working with mind maps might prompt cognitive strategy strategies were linked to the essential processes during learning
use during learning from text. When working with mind maps, stu- from text on the one hand (i.e., text selection, organization, trans-
dents are deliberately involved in reorganizing text information by formation, and integration) (Mayer, 2001; Schlag & Ploetzner,
cognitively searching for and interpreting associations among con- 2011) and specific learning techniques (e.g., scanning the text,
cepts in the text (Brinkmann, 2003; Davies, 2011; Farrand et al., highlighting) on the other hand (Askell-Williams et al., 2012;
2002). By means of this cognitive search, students receive practice Schlag & Ploetzner, 2011). As to the latter, students are stimulated
in important cognitive text-learning strategies such as organiza- to get a general overview of the to-be-learned text material in a
tion (e.g., highlighting important information), elaboration (e.g., first step of the strategy instruction. This step focuses on incorpo-
paraphrasing, making links with prior knowledge, and relating text rating specific reading and comprehension activities, such as scan-
to pictures), and rehearsal (e.g., rereading, reviewing notes) ning text, reading and clarifying difficult words or text parts during
(Weinstein & Mayer, 1986; Weinstein et al., 2011). The study of reading (De Corte, Verschaffel, & Van De Ven, 2001; Zimmerman,
Broer et al. (2002) indicated in this respect, that a lesson series Bonner, & Kovach, 2002). These learning techniques touch upon
in making schematic representations positively influenced sixth the metacognitive strategy ‘planning’ (Schraw, 1998), including
 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147 131

Table 1
Conceptualization of the mind map strategy instruction of the present study.

Induced cognitive (c) and metacognitive Essential cognitive processes during Sequentially ordered learning techniques (Askell-Williams et al.,
(mc) text-learning strategies learning from text (Mayer, 2001; 2012; Schlag & Ploetzner, 2011)
(Weinstein, Jung, & Acee, 2011; Schraw, 1998) Schlag & Ploetzner, 2011)
Planning (mc) Text selection (1) Getting an overview:
 Scan and read the text
 Clarify incomprehension
Organization (c) Text organization (2) Identifying key information: highlight relevant key words
Elaboration (c) Text transformation (3) Working with mind maps
Rehearsal (c)
Monitoring (mc)
Evaluation (mc) Text integration (4) Process and product evaluation

for example orientation to the text. In a second and third step, stu- a scaffold for students’ strategic processing (Kirschner et al., 2006;
dents select and organize important information by means of high- Leopold, Sumfleth, & Leutner, 2013; Stull & Mayer, 2007).
lighting, and information transformation is encouraged by the use To our knowledge, there is hardly any research on the influence
of mind maps. The focus here lies on organization, elaboration, and of these instructional approaches on elementary students’ sponta-
rehearsal strategies (cognitive strategy use; Weinstein et al., 2011) neous unprompted strategy use, since previous research has lar-
and on monitoring text comprehension and processing (metacog- gely focused on the performance of secondary and higher
nitive strategy use; Schraw, 1998). Finally, in a fourth step, stu- education students’ learning with either provided or self-
dents are encouraged to review the process followed and constructed maps (e.g., Leopold et al., 2013; Stull & Mayer, 2007).
evaluate their outcomes (metacognitive strategy use; Schraw, Hence, this raises the important question of which instructional
1998). In this respect, mind mapping can be regarded as a meta- approach best promotes strategy transfer to an independent learn-
learning strategy when the text-learning strategies mentioned ing task. In this respect, research on strategy transfer and the pro-
above, and incorporated into the mind map strategy instruction, motion of self-regulated learning seems to favor a learning by
are also spontaneously applied during independent text-learning. doing-approach, indicating that strategy transfer is more likely to
occur when (a) learners are treated as active participants
(Garner, 1990), and (b) direct promotion of strategies is provided
5. Two instructional mind map approaches by explicit instruction in specific strategy components (Frazier,
1993; Kistner et al., 2010). In this line of reasoning, it would also
A rich history of studies supports the conclusion that interven- be those learners who are most actively engaged with the mind
tions with longer-term and more complex instruction are required mapping strategy in class, applying the highest deep-level learning
to teach students how to learn and that essential strategies for strategies during independent text learning and consequently
text-based learning can be taught through a variety of well- attaining the highest text recall.
planned instructional techniques from late elementary grades on
(Dignath, Buettner, & Langfeldt, 2008; Merchie & Van Keer, 2013;
Mok, Ma, Liu, & So, 2005; Wolters, Benzon, & Arroyo-Giner, 6. The current study
2011). In this respect, it is essential to identify the most effective
instructional approach for integrating mind mapping as a meta- Above, we argued that working with mind maps might support
learning strategy into domain-specific content courses wherein students’ autonomous text-based learning. More particularly, we
students are provided with opportunities to apply and practice expect that implementing mind mapping as a meta-learning strat-
the newly acquired text-learning strategies (Dignath et al., 2008). egy in class will induce students’ cognitive and metacognitive text-
With regard to this instructional approach, a well-known dis- learning strategy use and increase their recall performance during
cussion is going on about working with researcher-provided versus independent text-learning. Additionally, it is hypothesized that
student-generated maps (Kirschner, Sweller, & Clark, 2006; Lee & when students work with student-generated mind maps and thus
Nelson, 2005; Stull & Mayer, 2007). On the one hand, researchers learn by doing, more spontaneous deeper-level text-learning strat-
are pleading for working with student-generated maps based upon egy application will occur. Hence, it is expected that greater gains
the activity theory, which states that deep learning involves the in this deeper-level text-learning strategy use and respectively
engagement of learners in productive learning activities recall performance will be found for students working with
(Kirschner et al., 2006; Stull & Mayer, 2007). In this respect, gener- student-generated maps. The present research investigates the fol-
ative processing (i.e., deeper cognitive processing of the material) lowing research questions:
is induced by ‘learning by doing’, challenging learners to think dee-
ply and letting them actively engage in selecting and organizing 1. What is the impact of two instructional approaches of mind
text in relation to their existing knowledge structures (Kirschner mapping (i.e. with researcher-provided and student-generated
et al., 2006; Stull & Mayer, 2007). However, learners’ generative mind maps) on pre-adolescents’ spontaneously applied cogni-
processes might also be inhibited by the extra cognitive demands tive and metacognitive strategy use?
this active engagement entails (Stull & Mayer, 2007). On the other 2. What is the impact of two instructional approaches of mind
hand, researchers mainly inspired by cognitive load research mapping (i.e. with researcher-provided and student-generated
(Sweller & Chandler, 1994) support the use of author- or mind maps) on pre-adolescents’ free recall performance?
researcher-provided maps. By providing students with worked-
examples and letting them ‘learn by viewing’, the level of extrane- 7. Method
ous processing is reduced. More cognitive capacity is available for
generative processing, inducing learners to seek understanding of 7.1. Design
how the linear text is selected, organized, and transformed into a
spatial structure. Texts complemented with worked-examples of A quasi-experimental repeated measures design (i.e., pretest,
maps would thus provide more opportunities to learn, as they offer posttest, retention test) was applied. Elementary schools who
132 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147

agreed to participate in the study were randomly assigned to either df = 2, p = .023) between the three conditions. However, a signifi-
(a) a condition with researcher-provided mind maps (RPMM), (b) a cant difference in distribution was found according to gender
condition with student-generated mind maps (SGMM), or (c) a (v2 = 7.552, df = 2, p = .023), as the control condition included more
control condition. To avoid design contamination effects, teachers boys. From all parents passive informed consent was obtained as
within the same school were assigned to the same condition. they were offered the opportunity to withdraw their child from
Teachers in the experimental conditions embedded a specific participation in the study.
teacher-directed instructional approach of mind mapping once a
week over a 10-week interval in their social study and science 7.3. Intervention
lessons during regular classroom hours. Classes assigned to the
control condition received no explicit strategy instruction in text- 7.3.1. General features of the mind map strategy instructions
learning strategies and teachers followed their usual teaching A previously tested mind map-training (Merchie & Van Keer,
repertoire. Teachers in the control condition were not aware of 2013) was used as the basis for the development of two strategy
the information provided in the experimental conditions. Control instruction programs. Both programs include 10 lessons of 50
classes were provided however with the opportunity to use the min each, spread over 10 consecutive weeks, and share a general
instructional material during the subsequent school year. The con- structure. In the first lesson, students were explicitly introduced
ducted research consisted of five phases: (1) pretest administration to essential mind map characteristics (e.g., radial structure, color
(midterm September 2011); (2) a focused 1.5 h after-school train- use, dimension) and their importance in processing and learning
ing for teachers in the experimental conditions (midterm Septem- informative texts. Lesson 2–9 were devoted to the gradual instruc-
ber 2011); (3) ten-week intervention period (from October until tion, practice, and application of specific text-learning strategies,
midterm December 2011); (4) posttest administration (midterm supporting the essential text-learning processes, by means of four
December 2011); and (5) retention test administration (midterm sequentially ordered learning techniques i.e., (1) scanning the text,
March 2012). Fig. 2 illustrates the design of the study. reading the text and clarifying incomprehension (‘getting an over-
view’), (2) identifying key information by highlighting relevant key
7.2. Participants words, subideas, and supporting details in different colors (‘text
organization’-strategy), (3) active manipulation of the text mate-
Fourteen fifth-grade teachers, fifteen sixth-grade, and six multi- rial by means of mind map assignments (‘text transformation’-str
grade teachers and their 644 students from 17 different elemen- ategy), and (4) reviewing the process and product outcomes (cf.,
tary middle-class schools from the same socio-economic area in Table 1). This strategy instruction can be regarded as a multicom-
Flanders (the Flemish-speaking part of Belgium) participated in ponent instruction (Edmonds et al., 2009), as the intervention
the study. Multigrade teachers in the present study are teaching included instruction in various text-learning strategies. The tenth
multigrade classes comprising students from fifth and sixth grade, and final lesson was spent on explicitly addressing the transfer of
which retain however their grade-specific designation and text- mind mapping in multiple content areas (e.g., writing, mathemat-
books (Mulryan-Kyne, 2007). The average class size was approxi- ics, French).
mately 19 students (SD = 4.68) per class. Teachers were on Informative texts used in both experimental conditions were
average 34.74 years old (SD = 9.49) and had on average 13.77 years identical and derived from students’ grade-specific social studies
(SD = 9.21) of teaching experience. Eight teachers were men and science textbooks. Texts were provided for fifth and sixth
(22.9%). Table 2 presents the characteristics of the teachers in all grade, addressing their grade-specific subject-matter on nature
conditions. Chi-square analyses indicated no significant differences (e.g., animals in fifth grade, ecology in sixth grade), history (e.g.,
in the distribution of gender (v2 = 0.413, df = 2, p = .205) and grade World War I in fifth grade and armistice in sixth grade), and society
(v2 = 0.694, df = 4, p = .311) across conditions. Furthermore, one- (e.g., the town council in fifth grade and country’s governance in
way analysis of variances indicated no significant differences sixth grade). As lessons progressed, structural clues and signaling
between the three conditions regarding teachers’ age (F(2, 34) devices in the informative texts (e.g., subheadings, words printed
= 0.134, p = .875) and teaching experience (F(2, 34) = 0.265, in italics or boldface indicating or emphasizing text structure)
p = .769). (Lorch, 1989) were gradually omitted, to induce students’ indepen-
Students average age was 11.44 (SD = 0.68), with 53.9% boys dent selection and organization processes. On the basis of prior
and 46.1% girls in the sample. 94.5% of the students had Dutch as research (Merchie & Van Keer, 2013) and elementary school teach-
their home language, which is the instructional language in Flan- ers’ suggestions, these texts were previously evaluated and
ders. Table 3 summarizes students’ characteristics (gender, home adjusted as to their length, difficulty, clarity, content, and
language, and grade) in the three research conditions. Chi-square organization.
analyses indicated no significant differences in the distribution of Experimental conditions were provided with all necessary
home language (v2 = 5.004, df = 2, p = .082) and grade (v2 = 7.552, teaching materials to implement the lessons, i.e., learning books

Fig. 2. Design of the study.

 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147 133

Table 2
Teacher characteristics.

Researcher-provided mind Student-generated mind Control group

map condition map condition Experimental
Experimental condition 1 condition 2
N % N % N %
Gender
Male 2 14.3 2 20 4 36.4
Female 12 85.7 8 80 7 63.6
Total 14 100 10 100 11 100
Grade
Fifth grade 5 35.7 4 40 5 45.5
Sixth grade 5 35.7 5 50 5 45.5
Multigrade 4 28.6 1 10 1 9
Age M SD M SD M SD
35.14 8.25 35.20 11.56 37.00 9.75
Years of teaching experience M SD M SD M SD
12.79 8.33 13.30 10.17 15.45 10.04

Table 3
Student characteristics.

Researcher-provided mind Student-generated mind Control group

map condition Experimental map condition Experimental
condition 1 condition 2
N % N % N %
Gender
Male 108 50.9 108 49.3 131 61.5
Female 104 49.1 111 50.7 82 38.5
Total 212 100 219 100 213 100
Home language
Dutch 191 91.8 209 96.8 200 94.8
Other language 17 8.2 7 3.2 11 5.2
Total 208 100 216 100 211 100
Grade
Fifth grade 95 44.8 92 42.0 106 49.8
Sixth grade 117 55.2 127 58.0 107 50.2
Total 212 100 219 100 213 100

Note. Due to missing data of 9 children, participant totals in the home language section slightly differs from participant totals in the gender and grade section.

for the students with all color-print lessons and teacher manuals. searched for particular answers in the mind map by means of elab-
These manuals comprised the ten detailed lesson plans with lesson orative interrogation (i.e., asking why-questions to connect their
objectives and assignment solutions, classroom posters, and a prior knowledge with informative text information) and applied
compilation reader with additional background information. the mind map to search for and deliberately draw inferences, cre-
ate and track analogies or contrasts, and explicitly explain relation-
7.3.2. Two distinct instructional approaches ships between concepts in mind map and text. In the teacher
Regardless of the identical included grade-specific informative training and manual teachers were requested to predominantly
texts, and similar implementation structure of the first, second, apply modeling and guided practice, and to ensure explicit feed-
and fourth learning technique, experimental conditions differ sig- back on each completed exercise. Further, the manual provided
nificantly in the implementation of the third learning technique, elaborated suggestions to incorporate individual work, group
that is the active manipulation of text material by means of mind work, and discussions in the lessons as well.
mapping. As to the first instructional approach, which was imple- As to the second instructional approach, which was imple-
mented in experimental condition 1, researcher-provided mind mented in experimental condition 2, student-generated mind
maps (RPMM) were applied. Mind maps included into the lessons maps (SGMM) were used for initiating text-learning strategy use,
were previously evaluated (Merchie & Van Keer, 2013) and drawn evoking the explicit application of text-learning strategies (i.e.,
by hand by a researcher trained in mind mapping. Paper-and- organization, elaboration, rehearsal and monitoring) by gradually
pencil mind maps were deliberately used as this was most similar teaching students to construct mind maps themselves. In this
to students’ paper and pencil assignments. After scanning, reading, way strategic learning was induced by letting students search for
and structuring the text, various text-learning strategies (i.e., orga- main and sub-branches themselves, and asking them to deliber-
nization, elaboration, rehearsal, monitoring) are induced through ately include arrows and figures into the mind map to portray
different types of exercises students had to complete on the basis important text inferences and relationships (e.g., analogies, con-
of the texts and the accompanying researcher-provided mind trasts). By means of the teacher manuals teachers were requested
map example. To this aim, teachers used techniques to induce to teach students directly how to construct mind maps. Here also,
strategic learning (Simpson & Nist, 2000). For example, students elaborated suggestions were incorporated to use individual work,
134 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147

group work and discussions. Appendix A provides an example les- administrated, asking students to write everything down what
son showing the four main phases of each lesson and the differ- they already knew about the topic of the text, as this might influ-
ences between both instructional approaches. ence students’ text processing and learning (Armand, 2001;
Ausubel, 1968). These student notes were matched to the informa-
7.3.3. Teacher training tion reflected in the text content. More particularly, students’ writ-
One week before the intervention and after pretest administra- ten down information was compared to an expert mind map in
tion, all experimental schools were visited individually, providing which all the ideas units of the text were visualized. The percent-
fifth- and sixth-grade teachers with a researcher-directed 1.5 h age of text information that was already covered in the notes and
after-school training. At the beginning of the training each teacher the expert mind map was applied as the measure of students’ prior
was provided with a grade-specific teacher manual for lesson knowledge. It was found that students had very limited or no prior
preparation and consultancy reference when delivering instruction knowledge about the text topics during the three measurement
(100 pages), and with a compilation reader with background infor- moments (Mpretest = 2.03; SDpretest = 1.78; Mposttest = 0.99; SDposttest =
mation on the use of mind maps (125 pages). The training con- 1.39; Mretention test = 0.66; SDretention test = 1.59). Second, after prior
sisted of the following parts: (1) a detailed description of the knowledge test completion, the learning task was administered.
rationale of the study, (2) the demonstration and exemplification Students were instructed to study a 500-word informative text to
of the specific mind map characteristics and the use of mind map- prepare for a test afterwards without any prompted strategy use
ping as a meta-learning strategy, and (3) a thorough presentation (Veenman, 2011). They received a finite amount of study time
of the instructional materials, including an overview of the phases, (30 min) to elicit efficient regulation of study activities (Leopold
topics, structured activities, classroom assignments, and suggested & Leutner, 2012; Rawson & Dunlosky, 2007). For the three mea-
instructional strategies per lesson. surement occasions, three different informative texts were used,
respectively about the life of seahorses (pretest), meerkats (postt-
7.3.4. Ensuring fidelity to treatment est), and hummingbirds (retention test). The same texts were used
Three main methods were applied to assure and enhance treat- for fifth and sixth grade. All texts were comparably subdivided into
ment fidelity (O’Donnell, 2008; Smith, Daunic, & Taylor, 2007; five text paragraphs accompanied by the following subheadings:
Swanson, Wanzek, Haring, Ciullo, & McCulley, 2013). First, all 24 general information, body parts, living environment, reproduction,
experimental condition teachers were asked to complete a struc- and interesting facts. The first four paragraphs were comple-
tured protocol booklet during the intervention period (1 protocol mented with a picture. During the learning task, students were
per lesson, consisting of three main sections). Herein teachers indi- allowed, but not obligated to use scratch paper. Third, students
cated the specific date and hour of the taught lesson, together with were requested to answer a task-specific self-report inventory con-
the total time spent on each lesson. Further, teachers evaluated per cerning their applied text-learning strategies. Fourth, students
lesson on a 5-point Likert-scale to which degree (a) each lesson were asked on a free recall test to write everything down they still
objective, which reflected critical intervention components and remembered from the text.
were stated in teachers’ manuals, was attained, (b) each assigned
activity was completed, and (c) the instructional materials for both 7.4.2. Cognitive and metacognitive text-learning strategy use
teachers and students were clear. Two open questions concluded As advised in the literature, students’ overt and covert cognitive
the protocol, asking for teachers’ additional remarks or sugges- and metacognitive text-learning strategy use was measured by
tions. Second, teachers’ lessons implementation and protocol com- combining task-specific self-reports with detailed trace analysis
pletion was encouraged by three-weekly phone calls and of the informative texts and the scratch papers (Samuelstuen &
personalized electronic reminders. Schools were given the oppor- Braten, 2007; Scott, 2008; Winne, 2010).
tunity to subscribe to an additional school visit in case of addi-
tional questions or instructional difficulties. Third, after the ten- 7.4.2.1. Self-reported text-learning strategy use. To gain insight into
week intervention, teachers received an additional questionnaire. students’ covert non-observable text-learning strategy use, the
This questionnaire queried which particular instructional strate- task-specific self-report Text-Learning Strategy Inventory (TLSI)
gies were applied per lesson (e.g., whole-class instruction, group was administered (Merchie et al., 2014) immediately after comple-
work, individual work). Further, teachers were asked to answer tion of the learning task. The TLSI comprises nine subscales: sum-
on a 5-point Likert-scale various items related to (a) students’ atti- marizing and schematizing (7 items, apre = .89, apost = .90,
tude during the lessons (e.g., Were students motivated during the aret = .91), highlighting (1 item), rereading (3 items, apre = .74,
lessons?), (b) teachers’ experiences during the project (e.g., Was it apost = .80, aret = .77), paraphrasing (7 items, apre = .76, apost = .82,
difficult for you to integrate mind mapping?), (c) the instructional aret = .82), linking with prior knowledge (3 items, apre = .74,
material (e.g., Was the material attractive?), and (d) the achieve- apost = .80, aret = .78), studying titles and pictures (3 items,
ment of the main lesson objectives (e.g., Is the majority, that is apre = .61, apost = .69, aret = .71), planful approach (3 items,
more than 80%, of the students capable of making specific mind apre = .65, apost = .69, aret = .74), monitoring (5 items, apre = .70,
map assignments?). Also interview information was collected as apost = .71, aret = .71), and self-evaluation (5 items, apre = .72,
background information to help interpret the protocols and ques- apost = .71, aret = .75). Good model fit results were obtained for this
tionnaire information. In this way, treatment fidelity measures nine-factor model in a confirmatory factor analysis (YB
were collected both qualitatively and quantitatively. v = 1014.708, df = 588, p < .001, CFI = .93, TLI = . 92, RMSEA = .03
with a 90% CI [.03, .04], SRM = .05) (Merchie et al., 2014). Table 4
7.4. Dependent measures presents some example items of the subscales. Students were
asked to complete the self-report questionnaire on a 5-point Likert
7.4.1. Overview of the task administration scale, indicating to which degree they applied the strategies during
At pretest, posttest, and retention test, a list of tasks were the abovementioned learning task activity (Braten & Stromso,
administered to investigate the role of mind mapping as a meta- 2011; Leopold & Leutner, 2012). Subscale scores for pretest, postt-
learning strategy and thus whether spontaneous (meta)cognitive est, and retention test were calculated in SPSS 18.
text-learning strategies were used during independent text learn-
ing and to evaluate students’ recall performance. First, before 7.4.2.2. Observed text-learning strategy use. Informative texts and
the onset of the learning task, a prior knowledge test was scratch papers were collected for investigating overt observable
 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147 135

Table 4 Table 5
Example items of the TLSI-subscales (Merchie et al., 2014). Overview of the Krippendorff’s alpha interrater reliability coefficients for the scored
traces on informative text and scratch paper.
Nitems Example item
Summarizing and 7 I repeated the text with my summary or graphic
schematizing organizer on my scratch paper
Highlighting 1 I marked the most important things
Rereading 3 To learn the text, I read the text a lot of times
Paraphrasing 7 I covered up the text information and tried to
recall it
Linking with prior 3 Before learning, I thought about what I already
knowledge knew
Studying titles and 3 I looked at the titles to understand the text
pictures
Planful 3 First, I read the whole text and then I started
approaching learning
Monitoring 5 While learning, I checked what I had already done
and how much I still had to do
Self-evaluation 5 I managed to learn the text in a good way

strategy use on students’ study material by means of trace

methodology (Braten & Samuelstuen, 2007; Winne, 2010). Traces
were examined in detail by a trained team of three researchers
by means of a previously developed scoring rubric (Appendix B).
The scoring rubrics were accompanied with an extensive coding
manual wherein expert mind maps were included representing an expert mind map, consisting out these different idea units,
the main and sub ideas of the three tests. This allowed to study was constructed, also based on a consensus among several text-
both the quantity and quality of the text or scratch paper notes learning experts. Third, students’ written reports were screened
(e.g., Braten & Samuelstuen, 2007; Meier, Rich, & Cady, 2006; on the idea units identified in the text. In a fourth step, the content
Nitko & Brookhart, 2007). First, informative texts traces (IT traces) (i.e., idea units) of students’ free recall was compared to the idea
were scored on a 4-point scale as to the following seven categories: units in the expert mind map. For each idea unit mentioned in both
(sub)-title markings, figure markings, marking sentences or key the free recall of the student and the expert map, the student
words into the paragraph information, quantity of the marked received one point. In this way, in a fifth step, a free recall score
information, used color quantity, color use for distinguishing was calculated, representing the percentage of correctly recalled
between main and subideas, and manner of color use. Based on text information (based on McNamara & Kintsch, 1996). In Appen-
these scores, an overall mean score was calculated, representing dix C, more detailed information and examples are included on this
the overall quality of text selection and organization on the infor- 5-step scoring procedure.
mative text (7 items, apre = .80; apost = .82; aret = .80). This score
was transformed to a score on 10.
Second, for the trace analyses on the scratch paper (SP traces) 7.5. Data analysis
different components of a good graphical summary were coded
on a 4-point scale (e.g., Hilbert & Renkl, 2008; Lee & Nelson, To investigate the short-term (posttest) and relatively longer
2005; Merchie & Van Keer, 2013; Taricani & Clariana, 2006), i.e. term (retention test) effects of both instructional mind map pro-
the hierarchical structure of the written information, the degree grams on students’ text-learning strategy use and free recall per-
of hierarchy, color use, the integration of key words, and content formance, multilevel piece-wise growth analyses was performed
coverage. Similar to the IT trace-score, a mean score was calculated in MlwiN 2.25 as the data under investigation have a clear hierar-
representing the overall quality of text transformation on the chical three-level structure. More particularly, measurement occa-
scratch paper (5 items, apre = .89; apost = .91; aret = .91) and trans- sions (i.e. pre-, post-, and retention test) (level 1) are clustered
formed to a score on 10. Study materials from 225 students within students (level 2), which are in turn nested within classes
(35%) (78 RPMM condition students, 74 SGMM condition students, (level 3). In this respect, the interdependency between students,
and 75 control condition students) were double coded for the three as belonging to the same class and sharing a common history
measurement occasions to check interrater reliability by means of and experiences, was explicitly taken into account (Maas & Hox,
Krippendorff’s alpha (Hayes & Krippendorff, 2007). Table 5 pre- 2005). Multigrade classes were included as one class into the anal-
sents the Krippendorff’s alpha (a) interrater reliability coefficients yses, as fifth- and sixth-grade students share the same teacher. In
of the coded trace categories. Krippendorff’s alpha ranged from .81 view of the analysis, the time span from pretest to retention test
to .99, indicating good to excellent agreement. is split up into a first piece (P1) covering students’ evolution from
pretest to posttest, and a second piece (P2) covering students’ evo-
7.4.3. Recall performance lution from posttest to retention test.
To measure students’ text recall, a free recall test was adminis- For each of the dependent variables (i.e., the nine subscale-
tered after completion of the TLSI. Students were asked to write scores on the TLSI, fourteen trace-scores and free recall score),
everything down they remembered from the text. A 5-step proce- three main steps were taken into the analyses. In the first step of
dure was followed to score students’ free recall performance (see the analysis, a three-level conceptual null model was estimated
Appendix C). First, important idea units were identified for each (model 0), which served as a baseline to compare with more com-
text included in the pretest (about seahorses), posttest (about plex models. This model predicts the overall pretest score on the
meerkats) and retention test (about hummingbirds), based on a dependent variable and the overall change from pretest to posttest
consensus among several text-learning experts. These idea units (P1 = phase 1) and from posttest to retention test (P2 = phase 2) for
could consist out of key words (i.e., ‘fish’) of short key phrases all students across all classes. In a second step, the experimental
(i.e., ‘biggest seahorse is 11 in.’). Second, for each text included conditions were included into the model to investigate the
136 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147

differential scores of the experimental groups contrasted against Table 6

the control group (model 1). In a third step, interaction effects were Numbers of students’ using highlighting and using scratch paper during pretest,
posttest and retention test.
included between condition and P1 and P2 to investigate the differ-
ential progress of the experimental groups contrasted against the RPMM SGMM Control
control group (model 2). By means of chi-squared tests it was Informative text
tested whether the individual parameters in the model are signif- Pretest 176 194 192
icantly different from zero. Posttest 201 196 192
Retention test 203 199 200
Scratch paper
8. Results Pretest 104 109 109
Posttest 146 130 128
8.1. Pre-analysis: assessment of treatment fidelity Retention test 137 98 132

First, analysis of the structured protocols showed that teachers

in both experimental conditions implemented the mind map-
pretest score to obtain the mean pretest score for all students in all
lessons in a consistent way and at the requested frequency of
classes in respectively the RPMM and SGMM condition. In the
one lesson per week. Average time spent on each lesson ranged
phase 1 column, the mean increase or decrease from pre- to postt-
from 45 to 70 min, without significant differences between the
est is shown for all students in all classes for the three different
experimental conditions (p = .116). In both experimental condi-
conditions. These parameters, also differential with respect to the
tions the main objectives and the assigned activities per lesson
control condition, must be added or subtracted from students’
were indicated as largely or fully attained and completed for the
pretest score to obtain the mean posttest score for the students
vast majority of the lessons (that is, more than 80%). Instructional
in respectively the RPMM, SGMM, and control condition. In the
teacher and student materials were found to be clear. Second, three
phase 2 column, the mean increase or decrease from post- to reten-
schools, (i.e., two RPMM condition schools and one SGMM condi-
tion test is shown for all students in all classes for the three differ-
tion school) requested a school visit and were individually visited
ent conditions. To obtain the mean retention test score, these
by the researcher. Third, corresponding to the lesson specifications
parameters must be added or subtracted to students’ mean postt-
in the teachers’ manuals, both conditions reported to have mainly
est score to obtain the mean retention test score in respectively the
used whole-class instruction and modeling during the first lessons
RPMM, SGMM, and control condition. In the last column, the vari-
and applied more varied instructional strategies as the lesson
ances at the different levels are represented, based on the concep-
series progressed (e.g. individual and group work). Further analysis
tual null models, showing for each separate subscale the total
of the teacher questionnaires indicated that (a) students were
variance partitioned into the between-classes variance (level 3),
motivated (MRPMM = 4.30, SDRPMM = .48; MSGMM = 4, SDSGMM = .53,
between-students within-classes variance (level 2) and between-
p = .188) and concentrated (MRPMM = 4.15, SDRPMM = .38; MSGMM =
measurements within-students variance (level 1). As to the first
4.12, SDSGMM = .35, p = .863) during the lessons, (b) teachers did
four subscales (i.e., ‘summarizing and schematizing’, ‘highlighting’,
not find it difficult to implement the lessons (MRPMM = 2.23,
‘rereading’, and ‘paraphrasing’) variances on each of the three
SDRPMM = .93; MSGMM = 2.75, SDSGMM = .89, p = .221) and believed
levels were significantly different from zero. For the other five sub-
that the instructional program had helped them to implement
scales (i.e., ‘linking with prior knowledge’, ‘studying titles and pic-
text-learning strategies into their regular school curriculum
tures’, ‘planful approach’, ‘monitoring’, and ‘self-evaluation’) only
(MRPMM = 4.23, SDRPMM = .44; MSGMM = 4.25, SDSGMM = .46, p =
the variance at level 2 and 1 was significantly different from zero,
.925). Furthermore they (c) underlined the attractiveness of the
indicating that the total variance of those dependent variables can
lesson material (MRPMM = 4.61, SDRPMM = .51; MSGMM = 4.50,
be explained by differences between students and measurement
SDSGMM = .53, p = .625) and reported that (d) they had largely
occasions.
attained the main lesson objectives (MRPMM = 3.56, SDRPMM = .42;
As to phase 1, four main findings draw our attention. First, stu-
MSGMM = 3.83, SDSGMM = .25, p = .117).
dents from both experimental conditions reported significantly
less paraphrasing-activities during learning from text (i.e., covering
8.2. Evolution in self-reported and observed cognitive and meta- texts parts and trying to recall, mentally retelling text information)
cognitive text-learning strategy use (v2RPMM = 15.031, df = 1, p < .001; v2SGMM = 19.820, df = 1, p < .001)
as opposed to the control condition, reporting significantly more
A first aim of this study was to investigate whether the mind use from pretest to posttest (v2 = 8.546, df = 1, p = .003). Second,
mapping intervention programs can be beneficial for students’ students from the control condition reported significantly more
independent cognitive and metacognitive text-learning strategy summarizing and schematizing activities (i.e., using scratch paper
use. Results from the self-reported and observed strategy use anal- for learning) (v2 = 11.298, df = 1, p < .001), but engaged signifi-
yses are represented successively. cantly less in linking prior knowledge (i.e., before learning, think-
ing about what they already know about the topic) (v2 = 16.779,
8.2.1. Self-reported strategy use df = 1, p < .001). Third, students from the SGMM-condition have a
First, the results of the evolution in self-reported text-learning significantly lower score on ‘self-evaluation’ directly after the
strategy use are presented, which allow us to gain more insight intervention period (v2SGMM = 5.803, df = 1, p = .016) and conse-
into the overtly and covertly used text-learning strategies. Table 7 quently seemed to be less self-assured about their text-learning
more particularly summarizes the model 2 estimates for the three- approach. Fourth, when comparing both experimental conditions
level analyses of the nine subscale scores of the TLSI. The intercept in phase 1, no significant differences could be observed.
b0 in the first column represents the mean pretest score for all stu- As to phase 2, regarding the evolution from posttest to retention
dents in all control condition classes. The parameters for the RPMM test, no significant evolution or devolution was found on reported
and SGMM condition in the column representing the pretest scores text-learning strategy use for students in the RPMM-condition
are differential with respect to the control condition and conse- when contrasted with students from the control condition. SGMM
quently must be added to or subtracted from the control condition students, however, reported significantly less rereading-activities
 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147 137

Table 7
Summary of the Model 2 estimatesa for the three-level analysis of the Text-Learning Strategy Inventory subscalesb for the three conditions.c

Pretest score Phase 1 (evolution from pre- to posttest) Phase 2 (evolution from posttest to Variance at
retention test) the
different
RPMM SGMM Control condition RPMM SGMM Control RPMM SGMM Control
levelsd
condition condition intercept b0 condition condition condition condition condition condition
SS +0.065 0.030 2.711 (0.146) +0.015 0.182 +0.299 0.020 0.237 0.081 3 11.79%⁄
(0.199) (0.210) (0.127) (0.125) (0.089)⁄ (0.126) (0.125) (0.089) 2 34.13%⁄
1 54.08%⁄
HL 0.131 0.105 4.476 (0.092) +0.122 0.049 +0.012 0.036 0.178 +0.108 3 2.84%⁄
(0.128) (0.131) (0.131) (0.130) (0.092) (0.131) (0.130) (0.093) 2 25.65%⁄
1 71.51%⁄
RR 0.077 +0.095 3.335 (0.100) 0.167 0.156 0.122 +0.007 0.255 0.005 3 5.02%⁄
(0.138) (0.143) (0.112) (0.112) (0.079) (0.113) (0.112)⁄ (0.080) 2 38.88%⁄
1 56.10%⁄
PAR +0.060 +0.051 3.063 (0.075) 0.339 0.386 +0.180 +0.100 +0.052 0.133 3 3.35%⁄
(0.104) (0.107) (0.087)⁄ (0.087)⁄ (0.062)⁄ (0.087) (0.086) (0.062)⁄ 2 47.58%⁄
1 49.07%⁄
LPK 0.076 0.201 3.496 (0.092) +0.107 +0.116 0.371 0.182 0.273 0.140 3 2.57%
(0.129) (0.131) (0.129) (0.128) (0.091)⁄ (0.129) (0.127)⁄ (0.091) 2 33.36%⁄
1 64.07%⁄
TP +0.104 +0.048 3.474 (0.076) 0.062 +0.004 0.08 0.122 0.280 +0.088 3 1.41%
(0.108) (0.109) (0.104) (0.102) (0.073) (0.104) (0.103)⁄ (0.073) 2 42.83%⁄
1 55.76%⁄
PA +0.212 +0.181 3.663 (0.090) 0.250 0.252 0.113 +0.137 +0.001 0.127 3 1.87%
(0.126) (0.129) (0.135) (0.133) (0.095) (0.135) (0.133) (0.095) 2 25.47%⁄
1 72.66%⁄
MON +0.233⁄ +0.137 3.029 (0.070) 0.082 0.039 0.035 +0.062 0.251 0.077 3 2.62%
(0.099) (0.100) (0.088) (0.087) (0.062) (0.088) (0.087)⁄ (0.062) 2 50.36%⁄
1 47.38%⁄
SE 0.066 0.005 4.130 (0.049) 0.090 0.151 +0.068 +0.060 0.143 +0.050 3 2.90%
(0.069) (0.071) (0.063) (0.063)⁄ (0.045) (0.063) (0.062)⁄ (0.045) 2 43.01%⁄
1 54.09%⁄
a
Significant parameters are indicated with an asterisk (⁄), standard error estimates are placed between brackets.
b
SS = summarizing and schematizing, HL = highlighting, PAR = paraphrasing, LPK = linking with prior knowledge, TP = studying titles and pictures, PA = planful approach,
MON = monitoring, SE = self-evaluation.
c
RPMM-condition = researcher-provided mind map condition, SGMM-condition = student-generated mind map condition.
d
3 = class-level variance, 2 = student-level variance, 3 = measurement occasion-level variance in the fully unconditional three-level null models.

(v2SGMM = 4.062, df = 1, p = .044), linking with prior knowledge- df = 1, p < .001). Students from both experimental conditions
activities (v2SGMM = 4.601, df = 1, p = .031), studying tittles and marked significantly more titles, subtitles, and key sentences, or
pictures-activities (v2SGMM = 7.473, df = 1, p = .006), monitoring words in the paragraph information and differentiated more
(v2SGMM = 8.375, df = 1, p = .004), and self-evaluation activities between main and subideas in the text by means of different col-
(v2SGMM = 5.240, df = 1, p = .022) than the control group. As to the ors. When mutually comparing experimental condition students’
paraphrasing-activities, control condition students’ subscale scores informative text traces, RPMM students outperform SGMM stu-
significantly declined during phase 2 (v2 = 4.631, df = 1, p = .034). dents in phase 1 on the amount of applied colors in the informative
Further, when comparing experimental condition students with text (v2 = 8.211, df = 1, p = .004). Fig. 3 illustrates the difference
each other, RPMM students outperform SGMM students in phase between a highlighted text paragraph of a RPMM-condition stu-
2 on rereading (v2 = 4.303, df = 1, p = 0.038), monitoring dent, a SGMM-condition student, and a control condition student,
(v2 = 13.094, df = 1, p < .001), and self-evaluation strategies receiving the highest overall posttest informative text-scores.
(v2 = 10.634, df = 1, p = .001). RPMM students more particularly As to phase 2, students in the SGMM-condition show a signifi-
reporting a significantly higher use of these strategies. cant decline with regard to title (v2SGMM = 14.094, df = 1, p < .001)
and paragraph markings (v2SGMM = 5.915, df = 1, p = .015), whereas
8.2.2. Observed strategy use RPMM-condition students evolve significantly more in marking
The findings regarding self-reported strategy use must be com- figure information (v2RPMM = 5.729, df = 1, p = .017) when con-
plemented with results from the overt text-learning strategy anal- trasted with control condition students. Further, RPMM-condition
ysis, stemming from students’ observable text-noting behavior on students outperform SGMM-condition students on sub title
their informative text and scratch paper. When first inspecting the (v2 = 5.060, df = 1, p = .024) and figure markings (v2 = 6.566,
level variances (Tables 7 and 8), variances on each of the three df = 1, p = .010).
levels for all dependent variables were significantly different from With regard to the traces on students’ scratch papers, a detailed
zero. examination of Table 9 reveals various significant findings which
With regard to the analyses of the traces on the informative text should be interpreted carefully. Even though students from the
(Table 8), students from both experimental conditions evolve sig- control condition significantly engaged in more scratch paper use
nificantly more than students from the control condition in phase from pre- to posttest, students from both experimental conditions
1 on almost every coded informative text-noting behavior. This is engaged significantly more in active knowledge transformation
also reflected into the overall informative text score, which reflects when compared with the control condition students. Analyses
the overall quality of text selection and organization on the infor- show in this respect the greatest gains for RPMM-condition stu-
mative text (v2RPMM = 38.792, df = 1, p < .001; v2SGMM = 21.969, dents, who significantly evolve on almost every coded category.
138 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147

Table 8
Summary of the Model 2 estimatesa for the three-level analysis of the traces on the informative textsb for the three conditions.c

Pretest score Phase 1 (evolution from pre- to Phase 2 (evolution from posttest to Variance
posttest) retention test) at the
different
RPMM SGMM Control condition RPMM SGMM Control RPMM SGMM Control
levelsd
condition condition intercept b0 condition condition condition condition condition condition
(Sub-)title markings 0.201 +0.085 0.971 (0.201) +0.927 +0.916 0.070 0.238 0.594 +0.131 3 17.55%⁄
(0.272) (0.289) (0.160)⁄ (0.158)⁄ (0.113) (0.160) (0.158)⁄ (0.113) 2 20.79%⁄
1 61.66%⁄
Figure markings 0.070 +0.010 0.444 (0.058) 0.027 +0.033 +0.016 +0.163 0.010 0.065 3 5.10%⁄
(0.079) (0.082) (0.068) (0.067) (0.048) (0.068)⁄ (0.067) (0.048) 2 30.38%⁄
1 64.52%⁄
Disting. main/sub 0.059 +0.154 1.468 (0.089) +0.368 +0.181 +0.088 0.143 0.233 0.069 3 7.82%⁄
ideas (0.122) (0.128) (0.097)⁄ (0.096) (0.068) (0.097) (0.096)⁄ (0.069) 2 26.01%⁄
1 66.17%⁄
Quantity highlight. 0.236 0.019 2.111 (0.143) +0.642 +0.262 +0.239 0.132 0.211 +0.091 3 9.90%⁄
text (0.196) (0.206) (0.134)⁄ (0.132)⁄ (0.094)⁄ (0.134) (0.132) (0.094) 2 30.48%⁄
1 59.62%⁄
Color quantity 0.102 0.139 1.353 (0.116) +0.390 +0.377 +0.089 0.009 0.182 0.014 3 8.53%⁄
(0.159) (0.167) (0.122)⁄ (0.121)⁄ (0.086) (0.122) (0.121) (0.086) 2 22.13%⁄
1 69.34%⁄
Dinsting. main/sub 0.140 0.091 1.223 (0.089) +0.450 +0.274 +0.014 0.060 1.22 +0.014 3 8.43%⁄
ideas (by color) (0.122) (0.128) (0.095)⁄ (0.094)⁄ (0.067) (0.095) (0.094) (0.067) 2 22.09%⁄
1 68.48%⁄
Manner of color use 0.048 0.078 1.300 (0.106) +0.457 +0.337 +0.086 0.093 0.170 0.019 3 6.38%⁄
(0.145) (0.152) (0.129)⁄ (0.128)⁄ (0.091) (0.129) (0.128) (0.091) 2 19.90%⁄
1 73.72%⁄
Overall IT-score 0.329 0.041 3.159 (0.221) 1.174 0.874 0.139 0.195 0.554 0.044 3 13.33%⁄
(0.301) (0.319) (0.189)⁄ (0.187)⁄ (0.133) (0.189) (0.187)⁄ (0.133) 2 27.15%⁄
1 59.52%⁄
a
Significant parameters are indicated with an asterisk (⁄), standard error estimates are placed between brackets.
b
For more information on the coded categories, see Appendix B.
c
RPMM-condition = researcher-provided mind map condition, SGMM-condition = student-generated mind map condition.
d
3 = class-level variance, 2 = student-level variance, 3 = measurement occasion-level variance in the fully unconditional three-level null models.

RPMM-condition students even outperform SGMM-condition stu- 9. Discussion

dents in phase 1 on hierarchical degree (v2 = 7.433, df = 1,
p = .006), color use (v2 = 4.001, df = 1, p = . 045), integration of 9.1. The strength of researcher-provided mind maps in stimulating
key words (v2 = 4.239, df = 1, p = . 039), and the overall scratch spontaneous deep-level text-based learning
paper score (v2 = 6.221, df = 1, p = .013). Although control condi-
tion students reached higher scratch paper scores as well Two different instructional approaches of mind mapping (i.e.,
(v2 = 10.648, df = 1, p = .001), they mostly focused on linearly copy- working with researcher-provided versus student-generated mind
ing numerous key words, whereas experimental condition stu- maps) were tested in an elementary school-based intervention
dents reached higher levels of knowledge transformation by research to evaluate their value as a meta-learning strategy for
including a more hierarchical structure (v2RPMM = 14.013, df = 1, stimulating students’ spontaneously applied cognitive and
p < .001; v2SGMM = 3.944, df = 1, p = .047) and more hierarchical metacognitive strategies during and free recall performance after
relationships (v2RPMM = 40.361, df = 1, p < .001; v2SGMM = 13.578, independent text-learning. Two general lines of results are
df = 1, p < .001). Fig. 4 illustrates the difference in scratch paper discussed.
use between students from the three conditions, receiving the First, as to the evolution in students’ text-learning strategy use,
highest overall posttest scratch paper-scores. experimental condition students made significantly greater pro-
As to phase 2, this degree of hierarchy (v2RPMM = 6.536, df = 1, gress from pre- to posttest in applying overt deep-level strategies
p = .011) declines significantly for students in the RPMM- as compared to students in the control condition. This finding is
condition. However, their positive evolution in phase 1 sustained also corroborated with the self-reported strategy use analyses, as
three months after the intervention period in contrast with the students from both experimental groups engaged less in rather
SGMM-condition scores, which decline significantly in phase 2 surface-level paraphrasing activities (i.e., covering up text informa-
for almost every coded category. Table 6 provides a descriptive tion and trying to recall it). Students in the RPMM-condition show
overview of the number of students applying highlighting and the greatest gains in deep-level observable strategy use regarding
using scratch paper on the different measurement moments. the active knowledge organization in the informative text and
active text transformation on their scratch paper. Additionally, a
more sustained effect was found for RPMM-condition students
8.3. Evolution in free recall performance after the end of the intervention period, whereas the deeper-
level strategy use of the SGMM-condition declined significantly
As to the evolution in free text recall (Table 10), students from from posttest to retention test. These findings are consistent with
the control condition evolve significantly more in phase 1 previous work pleading for the implementation of predefined
(v2 = 40.508, df = 1, p < .001), whereas students from the SGMM worked-examples (e.g., Leopold et al., 2013; Stull & Mayer, 2007).
condition have significantly lower scores (v2 = 7.948, df = 1, Furthermore, it goes beyond showing that students, who have not
p = .004). As to phase 2, no significant changes in free recall score received any specific guidelines and practice in constructing
occur for the three conditions. graphical summaries, transferred the induced strategies to an
 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147 139

Fig. 3. Overt text-learning strategy use in the informative text at posttest. Note. For more information on the coded trace categories, see Appendix B.

independent learning task without any provided mind map or introduction section, SGMM-condition students’ learning in class
prompted strategy use. This differs from previous studies wherein could be hampered by extraneous cognitive load (Sweller &
students are tested after being explicitly asked to execute a task Chandler, 1994; Sweller, van Merriënboer, & Paas, 1998). In this
with a provided graphic organizer. It is possible, however, that study, no insight was provided into students’ cognitive load during
the finite amount of provided time to study the informative text the implementation of the mind map lessons. However, it could be
during data collection has daunted SGMM-condition students to hypothesized that students in the researcher-provided mind map
engage in deeper-level scratch paper use. Possibly they could esti- condition might have experienced less extraneous cognitive load
mate more accurately the time it costs to construct a graphical through their interaction with worked-examples and the SGMM-
summary, fearing no time would be left to study their summary procedure is still too demanding for students in primary school
properly (Merchie & Van Keer, 2012). In addition, in the light (Paas, Tuovinen, Tabbers, & Van Gerven, 2003; Schwamborn,
of the Cognitive Load theory, as already mentioned in the Thillmann, Opfermann, & Leutner, 2011). In turn, this could
140 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147

Table 9
Summary of the Model 2 estimatesa for the three-level analysis of the traces on the scratch papersb for the three conditions.c

Pretest score Phase 1 (evolution from pre- to Phase 2 (evolution from posttest to Variance
posttest) retention test) at the
different
RPMM SGMM Control condition RPMM SGMM Control RPMM SGMM Control
levelsd
condition condition intercept b0 condition condition condition condition condition condition
Hierarchical 0.168 0.087 0.806 (0.158) 0.520 0.273 +0.270 0.223 0.449 0.086 3 15.75%⁄
structure (0.214) (0.227) (0.139)⁄ (0.137)⁄ (0.098)⁄ (0.139) (0.137)⁄ (0.098) 2 21.57%⁄
1 62.68%⁄
Degree of +0.400 +0.040 0.307 (0.190) +1.152 +0.661 +0.201 0.464 0.602 +0.140 3 18.59%⁄
hierarchy (0.258) (0.274) (0.181)⁄ (0.180)⁄ (0.128) (0.181)⁄ (0.180)⁄ (0.128) 2 13.33%⁄
1 68.08%⁄
Color use 0.122 0.022 0.641 (0.128) 0.502 0.238 +0.300 0.211 0.386 0.061 3 11.09%⁄
(0.174) (0.184) (0.133)⁄ (0.132) (0.094)⁄ (0.133) (0.132)⁄ (0.094) 2 18.18%⁄
1 70.73%⁄
Integrating 0.011 0.069 1.193 (0.180) +0.629 +0.272 +0.305 0.267 0.561 +0.145 3 11.47%⁄
key words (0.245) (0.259) (0.175)⁄ (0.173) (0.123)⁄ (0.175) (0.173)⁄ (0.123) 2 21.60%⁄
1 66.93%⁄
Content +0.061 +0.019 0.815 (0.127) 0.175 0.345 +0.436 0.059 0.079 0.004 3 9.01%⁄
coverage (0.174) (0.183) (0.129) (0.127)⁄ (0.091)⁄ (0.129) (0.127) (0.091) 2 27.40%⁄
1 66.59%⁄
Overall score 0.329 0.054 1.868 (0.358) 1.356 0.594 0.708 0.614 1.061 0.185 3 15.84%⁄
(0.487) (0.516) (0.308)⁄ (0.305) (0.217)⁄ (0.308) (0.305)⁄ (0.218) 2 23.26%⁄
1 60.90%⁄
a
Significant parameters are indicated with an asterisk (⁄), standard error estimates are placed between brackets.
b
For more information on the coded categories, see Appendix B.
c
RPMM-condition = researcher-provided mind map condition, SGMM-condition = student-generated mind map condition.
d
3 = class-level variance, 2 = student-level variance, 3 = measurement occasion-level variance in the fully unconditional three-level null models.

possibly explain why higher gains for these students were found in in the light of the Fuzzy Trace Theory, it is possible that experimen-
their spontaneous application of deeper-level text-learning strate- tal condition students learned the text differently, forming differ-
gies during independent text learning. Furthermore, it must be ent mental representations and memory traces (e.g., focusing on
noticed that also control condition students made a significant comprehending the gist of the text by graphically summarizing
growth in self-reported and observed scratch paper use from pre- information) than those required in the free recall test (e.g., asking
to posttest, although less deeply as experimental condition stu- for as much as possible detailed information) (Reyna & Brainerd,
dents did. In this respect, control condition students kept a prefer- 1995; Tzeng, 2010).
ence to the traditional linear outline format. Their growth can be Even though output measures are regularly asked for and used
ascribed to the traditional teaching practices in their classes, as in educational research to measure students’ gain, (Zimmerman
standards from elementary education already stress the impor- et al., 2002) state that students’ success is not how much students
tance of cross-curricular ‘learning to learn’ skills. Therefore, these improve their performance (knowledge acquisition) but rather lies
skills are already stimulated in daily practice to a certain degree in students’ learning processes and how well the new strategies
in the commercially available teaching manuals. (stimulating knowledge organization and transformation) are
Second, as to students’ evolution in recall performance, control applied. In this respect, this study shows that mind mapping is a
condition students attained a significantly higher free recall score promising meta-learning strategy to induce and stimulate deep-
at posttest when contrasted with students from de SGMM- level text-learning strategy use, especially the instructional
condition. No significant gains were found for students in the approach incorporating researcher-provided mind maps.
RPMM-condition, compared to the control condition students.
Although greater gains for experimental condition students could 9.2. Limitations and suggestions for future research
have been expected, as deeper-level text processing (e.g., actively
transforming and elaborating text information) has been found to The findings from this study must be complemented with the
be related with higher performance (Lahtinen et al., 1997; Nesbit discussion of some limitations and constraints inherently related
& Adesope, 2006), many possible intermediating factors might to the nature of large-scale school-based intervention research.
explain this result. First, text acquisition might have been influ- A first constraint relates to the used data-gathering methods. In
enced by the dual task experimental condition students were aspiring to query information on a large group of learners, self-
focusing on, i.e. text learning and actively activating procedures report measures were used, to ensure straightforward data gather-
necessary for text organization and graphical transformation ing and scoring (Braten & Samuelstuen, 2004; Schellings, 2011;
(Griffin, Malone, & Kameenui, 1995). Second, experimental condi- Schellings & Van Hout-Wolters, 2011) Additionally, trace method-
tion students might not yet have reached the final fourth stage of ology provided a more arguably objective set of data on students’
strategy acquisition, wherein the performed strategies also overt strategy use (Braten & Samuelstuen, 2007; Winne, 2010).
improve learning (Bjorklund, Miller, Coyle, & Slawinski, 1997; However, some divergence between the self-report and trace anal-
Malmberg, Järvenoja, & Järvelä, 2013). In the first two stages the yses must be acknowledged. The significant growth in experimen-
new strategy is not performed (stage 1) or not produced com- tal condition students’ traced scratch paper use from pre- to
pletely (stage 2). In the third stage, Bjorklund et al. (1997) refer posttest is not reflected, for example, into their self-reports on
to the ‘utilization deficiency’, were little or no benefits arise from ‘summarizing and schematizing’-activities, which partially cover
the used strategies. Third, the finite amount of study time might aspects of this overt strategy use. It might be possible that students
have impede students to learn from their constructed graphical were unable to reflect in a way necessary for the correct comple-
summary or might have affected its quality (Fox, 2009). Fourth, tion of the questionnaire (Schellings, 2011) and overestimated
 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147 141

Fig. 4. Overt text-learning strategy use on the scratch paper at posttest. Note. For more information on the coded trace categories, see Appendix B.

their scratch paper-strategy use at pretest, revealing no subse- text-learning strategy use. In this respect, it would be fruitful to
quent evolution into their growth analysis. Next to this divergence, apply the think-aloud methodology on smaller sub-samples in
not all significant growth evolutions in covert strategy use were future intervention studies (Schellings, Aarnoutse, & van Leeuwe,
straightforwardly interpretable. Consequently, although the com- 2006; Scott, 2008).
bination of self-reports and traces is a clear asset of this study, A second constraint of this study relates to the tests used and to
some applied (meta)cognitive strategies might not be reported the provided time for testing. It was deliberately opted for to use of
on or elicited in the traces. Therefore, these findings point at the a learning task that acknowledged the context- and domain-
need for further data-triangulation and the use of other data- specificity of learning from text and provided the engagement in
gathering methods to provide more detailed understanding of stu- an authentic learning task, as elementary students might experi-
dents’ strategy use, in particular their more covert (meta)cognitive ence difficulties reflecting upon a hypothetical situation (Braten
142 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147

Table 10
Summary of the Model 2 estimatesa for the three-level analysis of the free recall scores for the three conditions.b

Pretest score Phase 1 (evolution from pre- to posttest) Phase 2 (evolution from posttest to Variance
retention test) at the
different
RPMM SGMM Control condition RPMM SGMM Control RPMM SGMM Control
levelsc
condition condition intercept b0 condition condition condition condition condition condition
Free recall 1.211 1.248 18.102 (1.279) 1.452 2.648 +4.308 1.043 0.834 +1.010 3 13.74%⁄
score (1.738) (1.841) (0.951) (0.939)⁄ (0.677)⁄ (0.948) (0.934) (0.672) 2 43.06%⁄
1 43.20%⁄
a
Significant parameters are indicated with an asterisk (⁄), standard error estimates are placed between brackets.
b
RPMM-condition = researcher-provided mind map condition, SGMM-condition = student-generated mind map condition.
c
3 = class-level variance, 2 = student-level variance, 3 = measurement occasion-level variance in the fully unconditional three-level null models.

& Samuelstuen, 2004, 2007). Furthermore, students were provided incidents. However, truthfully acknowledging the value of direct
with a recall test afterwards, ensuring they were orientated observations to obtain treatment fidelity measures, future research
towards learning from text. As a result, teachers were asked to pro- should keep on searching for valid means to obtain valid treatment
vide two lesson periods for test administration and students had to fidelity data in relation to these budgetary considerations.
be concentrated for a considerable amount of time. For this reason,
the inclusion of a more comprehensive range of tests for assessing
performance (e.g. transfer tests, comprehension tests) or more 9.3. Contributions and implications
individual characteristics (e.g., verbal and spatial ability tests)
was difficult, as this would be additionally even more time- Today’s educational practice requires effective teaching
consuming. Nevertheless, the shortcomings in the use of approaches for stimulating text-based learning from early adoles-
immediate and free recall tests (e.g., assessment of mainly factual cence on (Alexander & Res, 2012; Schlag & Ploetzner, 2011). In
knowledge) must be acknowledged (Robinson, 1998) and future response to this call, this study focused on the instructional use
research should include other tests. In particular, test developers of mind maps, which are already frequently employed in class-
are encouraged to compose a standardized test. In contrast with rooms. However their instructional use in current educational
standardized measurements for reading comprehension largely practice misses a clear scientific and evidence-based underpinning.
focusing on narrative texts, similar tests measuring students’ text- In this respect, this study extends prior work and pioneers in some
learning strategies in informative texts for late elementary grades important ways. First, in contrast with the previously rather short-
are not forehand. Furthermore, only free-recall testing was applied time investigated modest sample sizes in higher educational set-
to gain insight into students’ prior knowledge of the text topic. How- tings (e.g., Abi-El-Mona & Adb-El-Khalick, 2010; Dhindsa, Kasim,
ever, in best cases, multiple forms of assessments should be used for & Anderson, 2011; Farrand et al., 2002), this study is executed lon-
a complete characterization of prior knowledge, as this might hinder gitudinally on a large scale, studying students from thirty-five
or influence the learning process (Dochy, Segers, & Buehl, 1999). classes from 17 different elementary schools. Second, the study
Therefore, future research should also consider other forms of prior includes a randomized assignment of two experimental and one
knowledge tests, such as open questions, multiple choice tests, cloze control condition. Multilevel piece-wise growth modeling was
tests, completion tests, and recognition tests, which also provide applied to analyze students’ growth in strategy use, explicitly
valid means of assessment (Dochy et al., 1999). taking into account their hierarchical nesting in classes. Third,
A third constraint relates to the assessment of fidelity to treat- two different instructional approaches of mind mapping were
ment. Although this assessment was executed through the evalua- implemented during a whole semester by students’ regular class-
tion of structured protocols and questionnaires, two general room teachers into students’ intact classroom settings. Mind map-
remarks can be made. First, it might have been very valuable to ping is herein conceived as more than a simple associative tool or a
assess teachers’ specific competencies required for successful single strategy (e.g., Brinkmann, 2003; Davies, 2011; Moi & Lian,
intervention implementation, i.e., teachers mind mapping skills 2007). It is broadly conceptualized as a meta-learning strategy
in the SGMM condition. Although this was largely offset by teach- comprising sequentially ordered phases to induce a more compre-
ers’ training, the offering of background information, and worked- hensive range of text-learning strategies.
example mind maps included into their manuals’ correction key, Some important implications resulting from this study can be
teachers’ mind map skills might have been of influence. Further- translated to educational practice and research. First, this study
more, teachers were not explicitly questioned about balancing wants to encourage elementary school teachers to implement
teaching content and modeling strategy instruction, which might mind mapping as a meta-learning strategy into their content
have been difficult to do (Vaughn et al., 2013). Future mind map courses. In particular, incorporating provided mind maps provide
research, should take these possible influencing factors into a promising means to stimulate students’ autonomous text-based
account. Second, no systematic direct observations were executed learning. The present study shows that practice matters. In this
to compare actual implementation with the established criteria respect, it is of highest importance that this strategy instruction
(Smith et al., 2007) and to determine the amount of overlap does not occur occasionally, but takes place in a consistent way
between the conventional teaching strategies in the control condi- on a regular basis. Only in this way, it can be assured that the
tion and in the induced strategies in the experimental conditions induced text-learning strategies by teachers’ modeling are system-
(Lemons, Fuchs, Gilbert, & Fuchs, 2014). Various factors were pro- atically initiated, recapitulated, and practiced towards a more self-
hibitive to carry out direct observations in this study, such as the regulated, autonomous, and deeper-level strategy use. In relation
expensive form of this data collection (i.e., costs related to the to this, this study also hopes to inspire instructional designers to
employment and training of personal attending the intervention go beyond the inclusion of a single mind map-lesson or chapter
sessions and the time-consuming coding) (Swanson et al., 2013) into their existing teaching methods. In this respect, they are
and the desire to minimize the impact of researchers’ attendance encouraged to implement worked-out mind maps with an
in the authentic classroom settings with their naturally occurring increasing level of difficulty and accompanied with student
 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147 143

activities simulating authentic learning situations, in dialogue with the class’ study results and some general overall findings. In this
mind map- and subject matter-experts. way, good relationships with schools are maintained.
As to the implications for educational research, a final concrete To conclude, the present study proposes an effective educa-
research suggestion, in addition to those already mentioned above, tional teaching practice and shows that the instructional use of
is to investigate an extra hierarchical level (i.e., teachers having the mind maps in late elementary education is a powerful technique,
same principals) into multilevel analyses. As this might have influ- equipping students with essential text-learning strategies neces-
enced teachers’ project-commitment and sustainability (Mishna, sary for school and future learning. It is believed that the added
Muskat, & Cook, 2012), it seems worthwhile to take this into value of intervention-research in early adolescents (i.e., investigat-
account in future research. Next to these concrete suggestions, ing students in an ecologically valid school context) exceeds the
we would like to emphasize some actions that were undertaken potential challenges and educational researchers should prioritize
to address and anticipate on potential setbacks and which were school-based intervention research to provide valid scientific
highly appreciated by the participants. They could be accounted underpinnings of contemporary educational practices.
for in future research, to keep schools motivated to participate in
intervention-research. First, principals and teachers valued the
Acknowledgment
individual school visits for both providing personalized project-
information and teacher training into their daily staffroom. These
This research was supported by a grant from the Research Foun-
visits enhanced project approval and cooperation and made project
dation – Flanders (FWO).
preparation as accessible as possible. Furthermore, the provision of
all necessary school material and the incorporation of the lesson-
series into the regular courses and the general school calendar Appendix A
enhanced participation. To thank the participants for their
project-participation each of the principals, teachers and students Example of a classroom assignment on children’s’ rights
received a personal appreciation. Finally, teachers were informed (lesson 6) in both strategy programs illustrating the instructional
about students’ progress by a feedback report afterwards, outlining approaches in experimental condition 1 and 2.

Instructional approach 1: Working with researcher-provided mind maps

Informative text Classroom Mind map-worksheet with Classroom assignment 2
assignment 1 researcher-provided MM
(1) Getting an over- (3) Working with mind maps: Active
view: Scan and read knowledge transformation
the text, clarify – Search for three groups of chil-
incomprehension dren’s rights in the mind map
(2) Identifying key and write them down in the
information: high- table
light relevant key – Which rights are the pictures
words illustrating? Write down
– Match the numbered rights with
the corresponding group in the
table
(4) Process and product evaluation

Instructional approach 2: Working with student-generated mind maps

Informative text Classroom Mind map-worksheet for Classroom assignment 2
assignment 1 student-generated MM
(1) Getting an over- (3) Working with mind maps: Active
view: Scan and read knowledge transformation
the text, clarify Challenge: Can you complete the follow-
incomprehension ing tasks within the next 35 min?
(2) Identifying key – Find all main branches
information: high- – Elaborate at least one main
light relevant key branch with related subbranches
words using most important key words
– Include 1 cluster, 1 arrow, or
connector
– Respect all specific mind map-
characteristics
Indicate whether or not you think the
challenge will succeed
(4) Process and product evaluation
144 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147

Appendix B

Scoring rubric for the traces on the informative text and scratch
paper.

Informative text
trace
0 1 2 3 4
(Sub-)title Not Only the title is Marked some or all Marked title and some Marked title and subtitles
markings applicable marked subtitles subtitles
(no text
markings)
Figure markings Not Markings in one Markings in several Markings and/or annotation A combination of markings
applicable figure figures in several figures and annotations in one or
(no figure more figures
markings)
Distinguishing Not There are There are There are predominantly Throughout the text
main and applicable predominantly predominantly key short key phrases and key relevant key words are
subideas (no text complete phrases highlighted, words marked. However, the marked, reflecting the main
markings) sentences or with now and then student still has some ideas of the text and most
fairly long text separate keywords difficulty in selecting the important text information
parts marked most important information
Quantity of the Not Relatively few Half of the text Relatively much Almost everything
markings applicable
(no text
markings)
Color quantity Not 1 color used 2 colors used 3 colors used More than 4 colors used
applicable
(no text
markings)
Distinguishing Not Only one color is There is more than Different colors are used, There is a consistent use of
main and sub applicable used during one color used, but however the distinction different colors throughout
ideas by (no text highlighting the distinction between main and sub ideas the text, reflecting the
means of markings) between main and is sometimes made still differences between main
different sub ideas is not clear inconsistently and sub ideas
colors Or: Every paragraph is Or: Every paragraph is
marked with a different color, consistently marked with a
but not always consistently different color
Manner of color Not Throughout the Different color use, Every text paragraph is Combination of marking
use applicable text, the same dependent on title, marked with a different color methods (e.g., marking text
(no text color is used subtitle and and underlining)
markings) paragraph
information

Scratch
paper trace
0 1 2 3 4
Hierarchical Not Linear unstructured summary Linear structured summary Graphical Graphical
structure applicable unstructured structured
(no scratch summary summary
paper use)
Hierarchical Not Columns scheme Sequential structure Tree structure Hierarchical and
degree applicable without hierarchical radial mind map
(no scratch structure structure
paper use)
Color use Not 1 color used 2 colors used 3 colors used More than 4 colors
applicable used
(no scratch
paper use)
 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147 145

Appendix B (continued)
Scratch
paper trace
0 1 2 3 4
Integrating Not There are predominantly There are predominantly There are There are
key words applicable complete sentences or fairly phrases copied on the predominantly short predominantly key
(no scratch long text parts copied on the scratch paper, key phrases copied words copied on
paper use) scratch paper reformulated in own words on the scratch paper the scratch paper
Content Not Little content coverage Average content coverage Good content Almost complete
coverage applicable coverage content coverage
(no scratch
paper use)

Appendix C

Detailed example of the 5-step scoring procedure of students’

free recall (for reasons of clearness illustrated on one text para-
graph of the 500-word informative text).

Original material (in Dutch) Translation in English

(1) Idea unit identification in the informative text (based on a consensus among experts)

(2) Expert mind map construction (based on a concensus among experts)

(3) Students’ free recall: idea unit identification

Eerste tekstparagraaf in stippenlijn: Assignment (in text balloon):
‘Write everything down you still remember about the text ‘The
wonderful life of the seahorse’.

Student’s free recall of the first text paragraph (within the dotted
) ‘Theirr head resembles
lines): resemb
m les that of a horse. They breathe through
gills.
gills.’

Student’s free recall of the rest of the 500-word informative text:

‘Male seahorses have a smooth chest and females a rough chest.
They live in oceans. Most of them live at the equator. They also live
in the North Sea. The males are pregnant. They are faithful their
whole lives. 100 of them can be born.’

(4) Comparison of students’ free recalled idea units (step 3) and the idea units in the expert map

(5) Free recall score calculation

Totale score: 2 idee-units van de 16 zijn onthouden: Total score: 2 idea units out of 16 units are recalled:
2/16 = 0.125% (d.i., studenten onthielden 0.125% van de 2/16 = 0.125% (i.e., students recalled 0.125% of the information
informatie uit de eerste tekstparagraaf) in the first text paragraph)
146 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147

References comprehension outcomes for older struggling readers. Review of Educational

Research, 79(1), 262–300. http://dx.doi.org/10.3102/0034654308325998.
Farrand, P., Hussain, F., & Hennessy, E. (2002). The efficacy of the ‘mind map’ study
Abi-El-Mona, I., & Adb-El-Khalick, F. (2010). The influence of Mind Mapping on eight
technique. Medical Education, 36, 426–431. http://dx.doi.org/10.1046/j.1365-
graders’ science achievement. School Science and Mathematics, 108(7), 298–312.
2923.2002.01205.x.
http://dx.doi.org/10.1111/j.1949-8594.2008.tb17843.x.
Fox, E. (2009). The role of reader characteristics in processing and learning from
Alexander, P. A., & Res, D. R. L. (2012). Reading into the future: Competence for the
informational text. Review of Educational Research, 79(1), 197–261. http://dx.
21st century. Educational Psychologist, 47(4), 259–280. http://dx.doi.org/
doi.org/10.3102/0034654308324654.
10.1080/00461520.2012.722511.
Frazier, D. W. (1993). Transfer of college developmental reading students
Anderson, R. C., & Hidde, J. L. (1971). Imagery and sentence learning. Journal of
textmarking strategies. Journal of Reading Behavior, 25(1), 17–41. http://dx.doi.
Educational Psychology, 62(6), 526–530. http://dx.doi.org/10.1037/h0031824.
org/10.1080/10862969309547800.
Armand, F. (2001). Learning from expository texts: Effects of the interaction of prior
Garner, R. (1990). When children and adults do not use learning-strategies –
knowledge and text structure on responses to different question types.
Toward a theory of settings. Review of Educational Research, 60(4), 517–529.
European Journal of Psychology of Education, 16, 67–86. http://dx.doi.org/
http://dx.doi.org/10.3102/00346543060004517.
10.1007/BF03172995.
Griffin, C. C., Malone, L. D., & Kameenui, E. J. (1995). Effects of graphic organizer
Askell-Williams, H., Lawson, M. J., & Skrzypiec, G. (2012). Scaffolding cognitive and
instruction on fifth-grade students. The Journal of Educational Research, 89(2),
metacognitive strategy instruction in regular class lessons. Instructional Science,
98–107. http://dx.doi.org/10.1080/00220671.1995.9941200.
40, 413–443. http://dx.doi.org/10.1007/s11251-011-9182-5.
Haber, R. N. (1970). How we remember what we see. Scientific American, 222(5),
Ausubel, D. P. (1968). Educational psychology. A cognitive view.New York: Holt,
104–112. http://dx.doi.org/10.1038/scientificamerican0570-104.
Rinehart and Winston.
Hayes, A. F., & Krippendorff, K. (2007). Answering the call for a standard reliability
Biggs, J. B. (1985). The role of metalearning in study processes. British Journal of
measure for coding data. Communication Methods and Measures, 1(1), 77–89.
Educational Psychology, 55, 185–212.
http://dx.doi.org/10.1080/19312450709336664.
Bjorklund, D. F., Miller, P. H., Coyle, T. R., & Slawinski, J. L. (1997). Instructing
Hilbert, T. S., & Renkl, A. (2008). Concept mapping as a follow-up strategy to
children to use memory strategies: Evidence of utilization deficiencies in
learning from texts: What characterizes good and poor mappers? Instructional
memory training studies. Developmental Review, 17(4), 411–441. http://dx.doi.
Science, 36, 53–73. http://dx.doi.org/10.1007/s11251-007-9022-9.
org/10.1006/drev.1997.0440.
Jackson, N. (2004). Developing the concept of metalearning. Innovations in Education
Braten, I., & Samuelstuen, M. S. (2004). Does the influence of reading purpose on
and Teaching International, 41(4), 391–403. http://dx.doi.org/10.1080/
reports of strategic text processing depend on students’ topic knowledge?
1470329042000276995.
Journal of Educational Psychology, 96(2), 324–336. http://dx.doi.org/10.1037/
Kardash, C. M., & Amlund, J. T. (1991). Self-reported learning strategies and learning
0022-0663.96.2.324.
from expository text. Contemporary Educational Psychology, 16, 117–138. http://
Braten, I., & Samuelstuen, M. S. (2007). Measuring strategic processing: Comparing
dx.doi.org/10.1016/0361-476X(91)90032.
task-specific self-reports to traces. Metacognition Learning, 2, 1–20. http://dx.
Kim, A.-H., Vaughn, S., Wanzek, J., & Wei, S. (2004). Graphic organizers and their
doi.org/10.1007/s11409-007-9004-y.
effects on the reading comprehension of students with LD: A synthesis of
Braten, I., & Stromso, H. I. (2011). Measuring strategic processing when students
research. Journal of Learning Disabilities, 37(2), 105–118.
read multiple texts. Metacognition Learning, 6, 111–130. http://dx.doi.org/
Kirschner, P. A., Sweller, J., & Clark, R. E. (2006). Why minimal guidance during
10.1007/s11409-011-9075-7.
instruction does not work: An analysis of the failure of constructivist, discovery,
Brinkmann, A. (2003). Graphical knowledge display – Mind mapping and concept
problem-based, experiential, and inquiry-based teaching. Educational
mapping as efficient tools in mathematics education. Mathematics Education
Psychologist, 41(2), 75–86. http://dx.doi.org/10.1207/s15326985ep4102_1.
Review, 16, 35–48.
Kistner, S., Rakoczy, K., Otto, B., Dignath-van Ewijk, C., Buttner, G., & Klieme, E.
Broer, N. A., Aarnoutse, C. A. J., Kieviet, F. K., & Van Leeuwe, J. F. J. (2002). The effect
(2010). Promotion of self-regulated learning in classrooms:
of instructing the structural aspect of texts. Educational Studies, 28(3), 213–238.
Investigating frequency, quality, and consequences for student performance.
http://dx.doi.org/10.1080/0305569022000003681.
Metacognition and Learning, 5(2), 157–171. http://dx.doi.org/10.1007/s11409-
Budd, J. W. (2004). Mind Maps as classroom exercises. Journal of Economic Education,
010-9055-3.
35(1), 35–46. http://dx.doi.org/10.3200/JECE.35.1.35-46.
Kulhavy, R. W., Stock, W. A., Peterson, S. E., Pridemore, D. R., & Klein, J. D. (1992).
Buzan, T. (1974). Use your head.London: BBC Books.
Using maps to retrieve text – A test of conjoint retention. Contemporary
Buzan, T. (2005). The ultimate book of Mind Maps.London: Thorsons.
Educational Psychology, 17(1), 56–70. http://dx.doi.org/10.1016/0361-476x(92)
Chiou, C. (2008). The effects of concept mapping on students’ learning
90046-2.
achievements and interests. Innovations in Education and Teaching
Lahtinen, V., Lonka, K., & Lindblom-Ylänne, S. (1997). Spontaneous study strategies
International, 45(4), 375–387. http://dx.doi.org/10.1080/14703290802377240.
and the quality of knowledge construction. British Journal of Educational
Chmielewski, T. L., & Dansereau, D. F. (1998). Enhancing the recall of text:
Psychology, 67, 13–24. http://dx.doi.org/10.1111/j.2044-8279.1997.tb01223.x.
Knowledge mapping training promotes implicit transfer. Journal of Educational
Lee, Y., & Nelson, D. W. (2005). Viewing or visualizing-which concept map strategy
Psychology, 90(3), 407–413. http://dx.doi.org/10.1037/0022-0663.90.3.407.
works best on problem-solving performance? British Journal of Educational
Chularut, P., & DeBacker, T. K. (2004). The influence of concept mapping on
Technology, 36(2), 193–203. http://dx.doi.org/10.1111/j.1467-8535.2005.00452.x.
achievement, self-regulation, and self-efficacy in students of English as a second
Lemons, C. J., Fuchs, D., Gilbert, J. K., & Fuchs, L. S. (2014). Evidence-based practices
language. Contemporary Educational Psychology, 29, 248–263. http://dx.doi.org/
in a changing world: Reconsidering the counterfactual in education research.
10.1016/j.cedpsych.2003.09.001.
Educational Researcher, 43(5), 242–252.
Dansereau, D. F., & Simpson, D. D. (2009). A picture is worth a thousand words: The
Leopold, L., & Leutner, D. (2012). Science text comprehension: Drawing, main idea
case for graphic representations. Professional Psychology – Research and Practice,
selection, and summarizing as learning strategies. Learning and Instruction, 22,
40(1), 104–110. http://dx.doi.org/10.1037/A0011827.
16–26. http://dx.doi.org/10.1016/j.learninstruc.2011.05.005.
Davies, M. (2011). Concept mapping, mind mapping and argument mapping: What
Leopold, C., Sumfleth, E., & Leutner, D. (2013). Learning with summaries: Effects of
are the differences and do they matter? Higher Education, 62, 279–301. http://
representation mode and type of learning activity on comprehension and
dx.doi.org/10.1007/s10734-010-9387-6.
transfer. Learning and Instruction, 27, 40–49. http://dx.doi.org/10.1016/j.
De Corte, E., Verschaffel, L., & Van De Ven, A. (2001). Improving text comprehension
learninstruc.2013.02.003.
strategies in upper primary school children: A design experiment. British Journal
Lorch, J. R. F. (1989). Text-signaling devices and their effects on reading and
of Educational Psychology, 71, 531–559. http://dx.doi.org/10.1348/
memory processes. Educational Psychology Review, 1(3), 209–234. http://dx.doi.
000709901158668.
org/10.1007/BF01320135.
Dexter, D. D., & Hughes, C. A. (2011). Graphic organizers and student with learning
Maas, C. J., & Hox, J. J. (2005). Sufficient sample sizes for multilevel modeling.
disabilities: A meta-analysis. Learning Disability Quarterly, 34(1), 51–72.
Methodology, 1(3), 86–92. http://dx.doi.org/10.1027/1614-2241.1.3.86.
Dhindsa, H. S., Kasim, M., & Anderson, O. R. (2011). Constructivist-Visual Mind Map
Malmberg, J., Järvenoja, H., & Järvelä, S. (2013). Patterns in elementary school
teaching approach and the quality of students’ cognitive structures. Journal of
students’ strategic actions in varying learning situations. Instructional Science.
Science Education and Technology, 20, 186–200. http://dx.doi.org/10.1007/
http://dx.doi.org/10.1007/s11251-012-9262-1.
s10956-010-9245-4.
Mayer, R. E. (1996). Learning strategies for making sense out of expository text: The
Dignath, C., Buettner, G., & Langfeldt, H. (2008). How can primary school students
SOI model for guiding three cognitive processes in knowledge construction.
learn self-regulated learning strategies most effectively? A meta-analysis on
Educational Psychology Review, 8(4), 357–371. http://dx.doi.org/10.1007/
self-regulation training programmes. Educational Research Review, 3, 101–129.
BF01463939.
http://dx.doi.org/10.1016/j.edurev.2008.02.003.
Mayer, R. E. (2001). Multimedia learning.New York: Cambridge University Press.
Dochy, F., Segers, M., & Buehl, M. M. (1999). The relation between assessment
McNamara, D. S., & Kintsch, W. (1996). Learning from texts: Effects of prior
practices and outcomes of studies: The case of research on prior knowledge.
knowledge and text coherence. Discourse Processes, 22, 247–288. http://dx.doi.
Review of Educational Research, 69(2), 145–186. http://dx.doi.org/10.3102/
org/10.1080/01638539609544975.
00346543069002145.
McNamara, D. S., Ozuru, Y., Best, R., & O’Reilly, T. (2007). The 4-pronged
Donker, A. S., de Boer, H., Kostons, D., Dignath van Ewijk, C. C., & van der Werf, M. P.
comprehension strategy framework. In D. S. McNamara (Ed.), Reading
C. (2014). Effectiveness of learning strategy instruction on academic
comprehension strategies: Theories, interventions, and technologies
performance: A meta-analysis. Educational Research Review, 11, 1–26. http://
(pp. 465–496). New York: Erlbaum.
dx.doi.org/10.1016/j.edurev.2013.11.002.
Meier, S. L., Rich, B. S., & Cady, J. (2006). Teachers’ use of rubrics to score non-
Edmonds, M. S., Vaughn, S., Wexler, J., Reutebuch, C., Cable, A., Tackett, K. K., et al.
traditional tasks: Factors related to discrepancies in scoring. Assessment in
(2009). A synthesis of reading interventions and effects on reading
Education, 13(1), 69–95. http://dx.doi.org/10.1080/09695940600563512.
 E. Merchie, H. Van Keer / Contemporary Educational Psychology 46 (2016) 128–147 147

Meneghetti, C., De Beni, R., & Cornoldi, C. (2007). Strategic knowledge and Schellings, G., & Van Hout-Wolters, B. (2011). Measuring strategy use with self-
consistency in students with good and poor study skills. European Journal of report instruments: Theoretical and empirical considerations. Metacognition
Cognitive Psychology, 19(4/5), 628–649. http://dx.doi.org/10.1080/ Learning, 6, 83–90. http://dx.doi.org/10.1007/s11409-011-9081-9.
09541440701325990. Schlag, S., & Ploetzner, R. (2011). Supporting learning from illustrated texts:
Merchie, E., & Van Keer, H. (2012). Spontaneous mind map use and learning from Conceptualizing and evaluating a learning strategy. Instructional Science, 39,
texts: The role of instruction and student characteristics. In Z. Bekirogullari 921–937. http://dx.doi.org/10.1007/s11251-010-9160-3.
(Ed.). Procedia social and behavioral sciences (Vol. 69, pp. 1387–1394). http:// Schnotz, W. (2002). Towards an integrated view of learning from text and visual
dx.doi.org/10.1016/j.sbspro.2012.12.077. displays. Educational Psychology Review, 14(1), 101–120. http://dx.doi.org/
Merchie, E., & Van Keer, H. (2013). Schematizing and processing informational texts 10.1023/A:1013136727916.
with mind maps in fifth and sixth grade. Middle Grades Research Journal, (8)(3), Schwamborn, A., Thillmann, H., Opfermann, M., & Leutner, D. (2011). Cognitive load
61–81 (pp. 61–81). and instructionally supported learning with provided and learner-generated
Merchie, E., & Van Keer, H. (2014). Using on-line and off-line measures to explore visualizations. Computers in Human Behavior, 27(1), 89–93. http://dx.doi.org/
fifth and sixth graders’ text-learning strategies and schematizing skills. Learning 10.1016/j.chb.2010.05.028.
and Individual Differences, 32, 193–203. http://dx.doi.org/10.1016/j. Scott, D. B. (2008). Assessing text processing: A comparison of four methods. Journal
lindif.2014.03.012. of Literacy Research, 40(3), 290–316. http://dx.doi.org/10.1080/
Merchie, E., Van Keer, H., & Vandevelde, S. (2014). Development of the Text- 10862960802502162.
Learning Strategies Inventory: Assessing and profiling learning from texts in Seyihoglu, A., & Kartal, A. (2010). The views of the teachers about mind mapping
fifth and sixth grade. Journal of Psychoeducational Assessment, 32(6), 533–547. technique in the elementary life and social studies lessons based on
http://dx.doi.org/10.1177/0734282914525155. constructivist method. Educational Sciences: Theory and Practice, 10(3),
Mishna, F., Muskat, B., & Cook, C. (2012). Anticipating challenges: School-based 1637–1656.
social work intervention research. Children & Schools, 34(3), 135–144. http://dx. Simpson, M. L., & Nist, S. L. (2000). An update on strategic learning: It’s more than
doi.org/10.1093/cs/cds002. textbook reading strategies. Journal of Adolescent & Adult Literacy, 43(6),
Moi, W.-A. G., & Lian, O. L. (2007). Introducing mind map in comprehension. Paper 528–541.
presented at the Educational Research Association, Singapore, May 2007. Slotte, V., & Lonka, K. (1999). Review and process effects of spontaneous note-taking
Mok, M. M. C., Ma, H. S., Liu, F. Y. F., & So, E. Y. P. (2005). Multilevel analysis of on text comprehension. Contemporary Educational Psychology, 24, 1–20. http://
primary students’ perceptions and deployment of self-learning strategies. dx.doi.org/10.1006/ceps.1998.0980.
Educational Psychology, 25(1), 129–148. http://dx.doi.org/10.1080/ Smith, S. W., Daunic, A. P., & Taylor, G. G. (2007). Treatment fidelity in applied
0144341042000294930. educational research: Expanding the adoption and application of measures to
Mulryan-Kyne, C. (2007). The preparation of teachers for multigrade teaching. ensure evidence-based practice. Education and Treatment of Children, 30(4),
Teaching and Teacher Education, 23(4), 501–514. http://dx.doi.org/10.1016/ 121–134. http://dx.doi.org/10.1353/etc.2007.0033.
j.tate.2006.12.003. Stull, A. T., & Mayer, R. E. (2007). Learning by doing versus learning by viewing:
Nesbit, J. C., & Adesope, O. O. (2006). Learning with concept and knowledge maps: A Three experimental comparisons of learner-generated versus author-provided
meta-analysis. Review of Educational Research, 76(3), 413–448. http://dx.doi.org/ graphic organizers. Journal of Educational Psychology, 99(4), 808–820. http://dx.
10.3102/00346543076003413. doi.org/10.1037/0022-0663.99.4.808.
Nitko, A. J., & Brookhart, S. M. (2007). Educational assessment of students (5th ed.) Swanson, E., Wanzek, J., Haring, C., Ciullo, S., & McCulley, L. (2013). Intervention
New Jersey: Pearson Education Inc. fidelity in special and general education research journals. Journal of Special
Novak, J. D. (2002). Meaningful learning: The essential factor for conceptual change Education, 47(1), 3–13. http://dx.doi.org/10.1177/0022466911419516.
in limited or inappropriate propositional hierarchies leading to empowerment Sweller, J., & Chandler, P. (1994). Why some material is difficult to learn. Cognition
of learners. Science Education, 86(4), 548–571. http://dx.doi.org/10.1002/ and Instruction, 12(3), 185–233. http://dx.doi.org/10.1207/s1532690xci1203_1.
sce.10032. Sweller, J., van Merriënboer, J. J. G., & Paas, F. G. W. C. (1998). Cognitive architecture
O’Donnell, A. M., Dansereau, D. F., & Hall, R. H. (2002). Knowledge maps as scaffolds and instructional design. Educational Psychology Review, 10(3), 251–296. http://
for cognitive processing. Educational Psychology Review, 14(1), 71–86. http://dx. dx.doi.org/10.1023/A:1022193728205.
doi.org/10.1023/A:1013132527007. Taricani, E. M., & Clariana, R. B. (2006). A technique for automatically scoring open-
O’Donnell, C. L. (2008). Defining, conceptualizing, and measuring fidelity of ended concept maps. Educational Technology Research & Development, 54(1),
implementation and its relationship to outcomes in K-12 curriculum 65–82. http://dx.doi.org/10.1007/s11423-006-6497-z.
intervention research. Review of Educational Research, 78(1), 33–84. http://dx. Tzeng, J.-Y. (2010). Design of concept maps and their impacts on readers’
doi.org/10.3102/0034654307313793. performance in memory and reasoning while reading. Journal of Research in
Okebukola, P. A. (1992). Attitude of teachers towards concept mapping and vee Reading, 33(2), 128–147. http://dx.doi.org/10.1111/j.1467-9817.2009.01404.x.
diagramming as metalearning tools in science and mathematics. Educational Vaughn, S., Swanson, E. A., Roberts, G., Wanzek, J., Stillman-Spisak, S. J., Solis, M.,
Research and Evaluation, 34(3), 201–213. et al. (2013). Improving reading comprehension and social studies knowledge in
Paas, F., Tuovinen, J. E., Tabbers, H., & Van Gerven, P. W. M. (2003). Cognitive load middle school. Reading Research Quarterly, 48(1), 77–93. http://dx.doi.org/
measurement as a means to advance cognitive load theory. Educational 10.1002/Rrq.039.
Psychologist, 38(1), 63–71. http://dx.doi.org/10.1207/S15326985ep3801_8. Veenman, M. V. J. (2011). Alternative assessment of strategy use with self-report
Paivio, A. (1991). Dual coding theory: Retrospect and current status. Canadian instruments: A discussion. Metacognition Learning, 6, 605–2011. http://dx.doi.
Journal of Psychology, 45(3), 255–287. http://dx.doi.org/10.1037/h0084295. org/10.1007/s11409-011-9080-x.
Ponce, H. R., & Mayer, R. E. (2014). Qualitatively different cognitive processing Vekiri, I. (2002). What is the value of graphical displays in learning? Educational
during online reading primed by different study activities. Computers in Human Psychology Review, 14(3), 261–312. http://dx.doi.org/10.1023/
Behavior, 30, 121–130. http://dx.doi.org/10.1016/j.chb.2013.07.054. A:1016064429161.
Rawson, K. A., & Dunlosky, J. (2007). Improving students’ self-evaluation of learning Wade, S. E., Trathen, W., & Schraw, G. (1990). An analysis of spontaneous study
for key concepts in textbook materials. European Journal of Cognitive Psychology, strategies. Reading Research Quarterly, 25(2), 147–166.
19, 559–579. http://dx.doi.org/10.1080/09541440701326022. Wallace, D. S., West, S. C., Ware, A., & Dansereau, D. F. (1998). The effect of
Reyna, V. F., & Brainerd, C. J. (1995). Fuzzy-trace theory: An interim synthesis. knowledge maps that incorporate gestalt principles on learning. Journal of
Learning and Individual Differences, 7, 1–75. http://dx.doi.org/10.1016/1041- Experimental Education, 67(5), 5–16. http://dx.doi.org/10.1080/
6080(95)90031-4. 00220979809598341.
Robinson, D. H. (1998). Graphic organizers as aids to text learning. Reading Research Weinstein, C. E., & Mayer, R. F. (1986). The teaching of learning strategies. In M. C.
and Instruction, 37(2), 85–105. http://dx.doi.org/10.1080/19388079809558257. Wittrock (Ed.), Handbook of research on teaching (3rd ed., pp. 315–327). New
Robinson, D. H., & Molina, E. (2002). The relative involvement of visual and auditory York: Macmillan.
working memory when studying adjunct displays. Contemporary Educational Weinstein, C. E., Jung, J., & Acee, T. W. (2011). Learning strategies. In V. G. Aukrust
Psychology, 27(1), 118–131. http://dx.doi.org/10.1006/ceps.2001.1081. (Ed.), Learning and cognition in education (pp. 137–143). Oxford: Elsevier
Robinson, D. H., Robinson, S. L., & Katayama, A. D. (1999). When words are Limited.
represented in memory like pictures: Evidence for spatial encoding of study Winne, P. H. (2010). Improving measurements of self-regulated learning.
materials. Contemporary Educational Psychology, 24(1), 38–54. http://dx.doi.org/ Educational Psychologist, 45(4), 267–276. http://dx.doi.org/10.1080/
10.1006/ceps.1998.0979. 00461520.2010.517150.
Samuelstuen, M. S., & Braten, I. (2007). Examinig the validity of self-reports on Wolters, C. A., Benzon, M. B., & Arroyo-Giner, C. (2011). Assessing strategies for the
scales measuring students’ strategic processing. British Journal of Educational self-regulation of motivation. In B. J. Zimmerman & D. H. Schunk (Eds.),
Psychology, 77, 351–378. http://dx.doi.org/10.1348/000709906X106147. Handbook of self-regulation of learning and performance (pp. 298–312). New
Schellings, G. (2011). Applying learning strategy questionnaires: Problems and York: Routledge.
possibilities. Metacognition Learning, 6, 91–109. http://dx.doi.org/10.1007/ Zimmerman, B. J., Bonner, S., & Kovach, R. (2002). Developing self-regulated learners.
s11409-011-9069-5. Beyond achievement to self-efficacy.Washington, DC 20002: American
Schellings, G., Aarnoutse, C. A. J., & van Leeuwe, J. (2006). Third-grader’s think- Psychological Association.
aloud protocols: Types of reading activities in reading an expository text. Zipp, G. P., Maher, C., & D’ Antoni, A. (2009). Mind Maps: Useful schematic tool for
Learning and Instruction, 16, 549–568. http://dx.doi.org/10.1016/j. organizing and integrating concepts of complex patient care in the clinic and
learninstruc.2006.10.004. classroom. Journal of College Teaching & Learning, 6(2), 59–68.