ERIC ED612455: Using Cluster Analysis to Explore Differences in the Reader Self-Perceptions of Adolescent Struggling Readers

2018·0.33 MB·English

by  ERIC

USING CLUSTER ANALYSIS 
TO EXPLORE DIFFERENCES
IN THE READER SELF-PERCEPTIONS
OF ADOLESCENT STRUGGLING READERS
S
F
O
Lindsey A. Chapman, Mary Beth Calhoon, and Jennifer Krawec
University of Miami O
R
8
Adolescent struggling readers (ASRs) with Pand without disabilities are oft1en reported uniformly in the litera-
ture as disengaged from reading and lacking the motivation required to read. The present study tested, at th e
0
beginning of the academic year, rea der self-perception of sixth grade ASRs (n = 71) enrolled in remedial read-
ing courses using the Reader SPelf-Perception Scale (RSPS2). Application of a hierarchical cluster analysi s
using the 4 RSPS domains yielded 3 statistically significantly distinct clusters of ASRs based on varianc e
across each domain of Athe RSPS.
©
I
There is little question that learning to read addressing the areas of motivation, attitude ,
proficiently is a complex process beyond sim- and self-perception in particular has been th e
ply an acquisition of skills. In fact, researc h lack of consensus in defining and differentiat-
over the past 2 decades has shifted from a sole ing them from each other. In both research an d
focus on the cognitive factors associated wit h practice, this inconsistency in terminology ha s
reading proficiency to a more comprehensiv e convoluted the findings and led to a blurrin g
examination of reading that includes affectiv e across both application and measuremen t
constructs such as motivation, attitude, an d (Unrau & Quirk, 2014).
self-perception. Though related, these con- Thus, blanket generalizations about th e
structs are distinct from one another, and eac h motivation and engagement of adolescent
significantly influences reading behaviors, an d struggling readers (ASRs) with and withou t
by extension, academic outcomes (Melnick , disabilities are frequently made without suffi-
Henk, & Marinak, 2009). One challenge i n cient empirical backing or contextual distinc-
• Correspondence concerning this article should be addressed to: Lindsey A. Chapman, [email protected]
Middle Grades Research Journal, Volume 12(1), 2018, pp. 39–49 ISSN 1937-0814
Copyright © 2019 Information Age Publishing, Inc. All rights of reproduction in any form reserved.
40 Middle Grades Research Journal Vol. 12, No. 1, 2018
tion. At times, ASRs are uniformly depicted as (e.g., reading). Though self-perception is a
having poor self-efficacy, being disengage d “work in progress” (Conradi et al., 2013, p.
from learning, and lacking the necessary moti- 155) rather than a fixed point of inquiry ,
vation to read (Guthrie & Davis, 2003). Other research is beginning to investigate how stu-
studies of this population have questioned th e dents see themselves generally and in specifi c
capacity of adolescent struggling readers to academic areas such as reading. To a larg e
accurately calibrate their ability to rea d extent, history of reinforcement and past aca-
(Klassen, 2002). However, it is unclear demic achievement largely influence one’ s
whether this is really the case for all adoles- academic self-perception (Bong & Skaalvic ,
cents who struggle in reading. 2003). Conversely, self-perception may influ-
Given that ASRs are largely heterogeneous ence academic achievement and the ways in
in terms of demographics, school-based expe- which students attempt a gSiven task (Montague
rience, disability status, and cognitive abilit y & van Garderen, 2003).
(Calhoon & Petscher, 2013), it seems likel y F
that there would also be marked variatio n
Reader SelOf-Perception
among ASRs in affective constructs. Althoug h
multiple affective constructs potentially TOhe multifaceted construct of reader self-
impact reading in ASRs, the affective con- perception incorporates students’ beliefs about
struct of interest in this study is reader self- R their ability to read, the processes they choos e
8
perception. According to Bong and Skaalvi k in learning to read, and whether they have pos-
(2003), the ways in which individuals conPstru e itive experi1ences with reading (Baştug &
themselves result in varying self-perceptions Celick,0 2015; Henk & Melnick, 1995). For th e

and likewise different coursesP of action and purposes of this study, reader self-perceptio n
2
behavior in the classroom. For purposes of th e is broadly defined as the way students see
present study, we disAtinguish reader self- themselves as readers (Henk, Marinak, &
©
perception profiles of ASRs as a way to inves- Melnick, 2013). Aligned to this view, reade r
I
tigate the “different courses of action” appro- self-perception can best be understood b y
priate to students to improve reading examining its underlying domains (Henk &
outcomes. Melnick, 1995). Based on Bandura’s (1977 )
model of self-efficacy, the four domains o f
reader self-perception are (a) observationa l
THEORETICAL FOUNDATION
comparison (how students see themselves in
relation to their peers), (b) social feedback (th e
The underlying factors that shape students ’ way students interpret messages received fro m
motivation for reading include students’ goals, their peers and teachers), (c) progress (thei r
dispositions, and beliefs about themselves view of their growth and learning), an d
(Conradi, Jang, & McKenna, 2013). These fac- (d) physiological state (how reading makes
tors are hierarchical and comprise an evolvin g them feel). Together with reading behavio r
area of inquiry in educational research. Self- (Chapman & Tunmer, 2003; Melnick et al. ,
perception—to some, synonymous with self- 2009), these four self-perception domain s
concept—is the foundational construct that serve as the core of the theoretical mode l
relates to how individuals see themselves underlying the present study (see Figure 1).
within a particular context; goals and disposi- Notably, the present study’s theoretica l
tions are then established upon this foundatio n model also posits that reader self-perception i s
(Rahmani, 2011). Self-perception may refer t o an important element in understanding the aca-
individual’s view of himself or herself as a stu- demic progress and outcomes of students who
dent, or more specifically focused on his or her struggle in reading, including the sometime s
self-perceptions related to a particular skil l discrepant relationship between how ASRs
Using Cluster Analysis to Explore Differences in the Reader Self-Perceptions of Adolescent Struggling Readers 41
S
F
O
O
FIGURE 1
R
Theoretical Model Showing Relationship Between Self-Perception and Aca8demic Progress in Reading
P 1
0
perceive their reading ability and their actua l sive reading courses in three middle school s
P
performance growth and outcomes. In additio n w2ithin a very large urban school district in th e
to the reciprocal relationship between both southeastern United States. Pursuant to stat e
A
perceived and actual academic progress an© d and district guidelines, intensive readin g
reader self-perceIption, previous research sug- courses are aimed to remediate identified defi-
gests that students’ reader self-perception ciencies in reading for ASRs who have previ-
influences their reading behaviors (Chapman ously failed to reach proficiency on the stat e
& Tunmer, 2003; Henk & Melnick, 1992). Fo r standardized assessment. The sample include d
example, self-perception may influenc e students with specific learning disabilities,
whether a student seeks out or avoids readin g other health impairments, and speech and lan-
tasks (Henk & Melnick, 1992). Likewise, self- guage impairments as well as general educa-
perception may impact the level of persistenc e
tion students. Table 1 provides additiona l
and effort applied to comprehending difficul t
demographic information. The study’s sampl e
text (Henk & Melnick, 1995). Consequently,
was largely Hispanic, reflecting a demo-
in our theoretical model we hypothesize tha t
graphic characteristic consistent with th e
reader self-perception indirectly affects aca-
school district. A moderate percentage of the
demic performance in reading and is therefore
sample was currently enrolled in or exited
a major factor in understanding chronic under-
from the district’s English for Speakers of
performance in reading among adolescents.
Other Languages (ESOL) program. It i s
important to note that students enrolled i n
METHOD ESOL programs were only excluded from eli-
gibility in the present study if they had not ye t
met proficiency in English according to distric t
Participants
ESOL guidelines.
The participants for the present study wer e Because ASRs for this study were a subset
sixth-grade students (n = 71) enrolled in inten- of a larger intervention study, they adhered to
42 Middle Grades Research Journal Vol. 12, No. 1, 2018
TABLE 1
Demographic Information for Student Participants
Gender Ethnicity Percentages Disability
N M Age Male Female White Hispanic Black IEP No IEP
School 1 33 12.03 61% 39% 3% 91% 6% 70% 30%
School 2 14 12.28 43% 57% — 93% 7% 29% 71%
School 3 24 11.90 50% 50% — 100% — 29% 71%
Total 71 12.03 54% 46% 1% 95% 4% 48% 52%
S
an academic-based inclusionary criteria. ASRs sustained deficits Fin word level reading an d
were deemed eligible based on a distinguish- decoding.
O
able pattern of underachievement on stat e
assessments and enrollment in a sixth-grade
MeaOsures
intensive reading course at their public middl e
school. Students were also administered tw o R Reader Self-Perception Scale (RSPS) .
8
additional assessments used for screening by Henk and Melnick (1995) developed the RSP S
the study team. Students were included iPf the y to better un1derstand the construct of reader
scored under a predetermined threshold on th e self-pe0rception. Based on an exploratory facto r

letter word identification (LWPI) and wor d analysis conducted by the instrument’ s
2
attack (WA) subtests of the Woodcock- authors, the RSPS consists of 33 total items ,
Johnson Tests of AchieAvement, Third Edition beginning with one general and 32 subsequen t
©
(WJIII; Woodcock, McGrew, & Mather , items representing four domains on whic h
I
2001). reader self-perception is measured (observa-
tional comparison, social feedback, progress,
Sample means for the age-based W scores
and physiological states). The RSPS uses a 5-
for LWI (M = 491.91, SD = 17.19) and WA ( M
point Likert-type scale with a variation i n
= 486.07, SD = 13.74) highlight students ’
number of items per factor. Notably, thi s
underperformance in reading given that th e
results in differences in maximum possible
reference W score—the median scores o f
scores which is an important consideration i n
same-age peers in the norming sample— on
analysis. Alpha reliabilities were calculated b y
the two subtests are, respectively, 515 and 51 2
the instrument authors to analyze the internal
(Jaffe, 2009). A W difference score (the differ-
consistency for scores on each of the four fac-
ence between the sample mean and the refer-
tors, which were found to be highly reliabl e
ence W) on LWI of approximately –23 shows
(32 items; α ranging from .81–.84). Table 2
these students generally have limited profi-
provides sample items for the four factors of
ciency in word reading and are likely to find
the RSPS as well as further reliability an d
grade-level word reading tasks very difficult descriptive statistics.
(Jaffe, 2009). Similarly, a W difference score
Woodcock Johnson III Tests of Achieve-
of nearly –26 on WA demonstrates limite d ment. In addition to the two screening assess-
proficiency in phonic decoding such that these ments (LWI and WA), the passag e
students would likely the find grade-leve l comprehension (PC) and spelling (SP) subtests
decoding tasks very difficult (Jaffe, 2009). Th e of the WJIII Tests of Achievement (Woodcoc k
rationale for the inclusionary criteria was t o et al., 2001) were individually administered to
ensure that participants had significant and students to assess reading achievement at the
Using Cluster Analysis to Explore Differences in the Reader Self-Perceptions of Adolescent Struggling Readers 43
TABLE 2
Sample Items and Reliability Calculations for the RSPS Subscale Scores
Factor Sample Items Items (n) α
I. Observational comparison • I read faster than other kids. 9 .84
• I seem to know more words than other kids when I read.
II. Progress • Reading is easier for me than it used to be. 6 .82
• When I read, I don’t have to try as hard as I used to.
III. Social feedback • My teacher thinks I am a good reader. 9 .81
• My classmates think I read pretty well.
IV. Psychological states • I feel calm when I read. 8 .84
• Reading makes me feel good.
S
F
beginning of the sixth grade. The WJIII is a dents, survey items were read to student s
O
widely used battery of assessments (Schrank , orally and students recorded their response s
McGrew, & Woodcock, 2001). Each assess- using the 5-point Likert-type scale. Studen t
O
ment takes about 5 to 7 minutes to complet e responses to each item were then coded, an d
and was administered per the standardized total scores for each of the four domains of the
R
administration protocol consistent. To deter- RSPS were tabula8ted electronically. All item
mine students’ proficiency in identifyingP rea l responses an1d domain scores were input into a
words in isolation, students were administered SPSS database by a study team member. Accu-
0
the LWI subtest. The WA subtest of the WJII I racy of data input was assessed for 20% of th e
P
requires students to read aloud nonsense words d2ata, and agreement was found to be 99.8%.
that are phonically correct in order to measure
A
students’ ability to decode. These two subtes©ts
Analysis and Findings
were also utilizIed as screening assessments.
The PC subtest assesses reading comprehen-
Cluster analysis is an exploratory statistica l
sion by asking students to insert a missing key
technique that maximizes the differences
word that makes contextual sense in the tex t
between groups and similarities within group s
provided. The fourth subtest, SP, is adminis-
(Hair, Black, Babin, Anderson, & Tatham,
tered orally and has students spell orally pre-
2010). Groups, or clusters, are formed around
sented words. Reliability of scores on thes e
predetermined variables selected by th e
four subtests is strong at .94, .87, .88, and .90 ,
researcher. Because the main research ques-
respectively (Schrank et al., 2001).
tion aimed to differentiate between groups o f
ASRs based on their reader self-perception ,
Procedures cluster analysis provided the optimal means to
accomplish this task. As detailed in this sec-
In accordance with university institutiona l tion, the investigation began first with multipl e
review board and district research guidelines , methods of cluster analysis to identify cluster s
parent consent and student assent were followed by testing of between- and within-
obtained prior to the start of data collection an d group differences.
analysis. Participants were administered the Identification of Student Clusters. Spe-
RSPS within the first month of the school yea r cifically, to answer the first research question,
using paper and pencil. Members of the study summative scores on each of the four domains
team administered the assessment to students of the RSPS (social feedback, progress, obser-
in small groups. Following administration o f vational comparison, and physiological states)
sample items to clarify procedures for stu- were used as the grouping variables. However ,
44 Middle Grades Research Journal Vol. 12, No. 1, 2018
because the measurement scales of the fou r 20.451, p < .001), social feedback (F(2,69) =
domains differed, scores were standardized t o 39.868, p < .001), and physiological state s
z scores during the clustering process. Due t o (F(2,69) = 29.075, p < .001). Results from
the fact that there was no a priori theory as t o pairwise comparison in post hoc analysis
the number of clusters that existed within the adjusted using the Tukey method indicate d
sample, hierarchical clustering using Ward’s that Profile 1 and Profile 3, Profile 1 and Pro-
method to maximize significance betwee n file 2, and Profile 2 and Profile 3 were signifi-
clusters was utilized first to determine cluster s cantly different in each of the four analyses, p
of students based on standardized domai n < .05. Table 3 presents the three clusters
scores on the RSPS. Both three-cluster an d (herein referred to as student profiles) with a
four-cluster solutions were initially explored brief description of each.
based on analysis of the resulting dendrogram. Figure 2 shows the mSean scores for each
Using the three-cluster solution, K-Means profile on the four domains assessed in the
F
clustering was employed to confirm cluster RSPS in addition to the sixth grade norm data
membership. Cohen’s κ was calculated to provided foOr comparative analysis by the
determine the level of agreement in cluster instrument’s developers (i.e., Henk & Mel-
assignment between the K-Means and Hierar- nickO, 1995). Notably, the norm data provide d
chical methods. There was moderate and sta- by the developers were not exclusively drawn
tistically significant agreement between th e R from a sample of A8SRs, though these data pro-
clustering of the two techniques (κ = .536, p < vide an important comparative indicator whe n
P 1
.001) according to Altman’s (1999) bench- examining reader self-perception in the study
marks. The three cluster solution emerged as sample0. The ordinal nature of reader self-per-

the best fit given the limited sPample size and ception among the study sample is evident i n
2
the statistically significant differences between that mean scores across all four domains sho w
A
groups on each of the four clustering variables. the progression from extremely low scores i n
©
Two-way analysis of variance (ANOVA) indi- Profile 3 to the average scores associated wit h
I
cated that the three clusters were significantl y Profile 1. In other words, there were no
different on progress (F(2,69) = 38.954, p < domains in which the mean score for Profile 1
.001), observational comparison (F(2,69) = was not the highest and, conversely, the mea n
TABLE 3
Descriptive Characteristics of the Three Clusters of ASRs
Profile n (%) Description
Average reader self-perception 19 (27%) Despite the norm data being gathered from students of a wide
range of abilities, Profile 1 data across all four factors closely
mirrored the norm. Scores in each factor were within or slightly
below the “average” scoring category (Henk & Melnick,
1995).
Generally low reader self-perception 44 (62%) Mean scores for these students across all factors were in the
“low” scoring category according to Henk and Melnick’s
(1995) guidelines and were likely the most accurate indication
of academic performance. The majority of students fell into
this profile.
Very low reader self-perception 8 (20%) Most alarmingly, the smallest cluster rated themselves
extremely low in each of the four factors. Scores in all four
factors were significantly lower than Profile 2 despite scores in
that group also being rated in the low range.
Using Cluster Analysis to Explore Differences in the Reader Self-Perceptions of Adolescent Struggling Readers 45
S
F
O
O
FIGRURE 2 8
Reader Self-Perception Mean Scores by Profile
P 1
0

P
score for Profile 3 was not the lowest. As note d (2Χ2(2) = 3.171, p = 0.205), disability statu s
previously, these differeAnces were statistically (Χ2(2) = .134, p = 0.935), or age (F(2,69) =
significant on all four domains. © .987, p = .381).
Profile CoImparisons. The second
research question asked the extent to which
DISCUSSION
profile groups differed in terms of academi c
outcomes and demographic variables. Interest-
ingly, although statistically significant differ- The results from the present descriptive study
ences were found for each domain of reade r provide further insight into the heterogeneous
self-perception, the profiles did not differ sig- nature of the ASR population with regards t o
nificantly on mean academic outcomes. Th e reader self. The three distinct cluster profile s
mean differences between profiles on the four showed that ASRs exhibit differing patterns of
standardized academic measures (WJIII: LWI, reader self-perception across all four domain s
WA, PC, and SP) were each tested using two- of the RSPS irrespective of their actual aca-
way ANOVA. Age-based W scores for each of demic outcomes. This is noteworthy given that
the four subtests were used as they are prefera- the academic outcomes for these ASRs are rel-
ble for statistical analyses and compariso n atively homogeneous based on the inclusio n
among groups (Jaffe, 2009). Results indicate d criteria for placement into the intensive read-
that there were no statistically significant dif- ing programs and results from the baselin e
ferences between profiles on LWI (F(2,69) = academic assessments administered. The fact
.331, p = .720), WA (F(2,69) = .401, p = .672) , that three distinct cluster profiles were foun d
PC (F(2,69) = 1.377, p = .263), or SP (F(2,69 ) within such an academically similar group o f
= .588, p = .560). In terms of student demo- ASRs shows the heterogeneity of reader self-
graphic variables, no association was foun d perception among these students with and
between profile membership and gender without disabilities, and may highlight oppor-
46 Middle Grades Research Journal Vol. 12, No. 1, 2018
tunities for differentiated support given these are able to judge their own ability and/or effi-
differences. At the least, this finding warrants cacy. Findings showed the most frequent mis-
further exploration as to why these ASRs, wit h calibration among students with disabilities i s
relatively similar reading abilities, have differ- in the form of overestimation. Related to thi s
ent perceptions of their reading ability. notion, Pajares (1996) questioned, “How much
confidence is too much confidence, when can
confidence be characterized as excessive and
Profiles
maladaptive in an academic enterprise, an d
what factors help create inaccurate self-per-
Interestingly, a majority of the study’s sam-
ceptions?” (p. 565). These important questions
ple (n = 44) clustered into Profile 2, character-
highlight the complexity of self-perception ;
ized by a generally low rating across each o f
that is, whereas higher self-perception is asso-
the four domains of the RSPS. However, S
ciated with persistence, effort, and engage-
despite low ratings, students in Profile 2
ment, more is notF necessarily better becaus e
viewed their reader self-perception signifi-
accuracy of perception is also a relevant factor.
cantly higher than their peers in Profile 3 an d O
significantly lower that Profile 1. It is encour- Therefore, teachers should be cautioned from
aging that students in Profile 2 appear to hav e relyOing solely on self-reports of student prog-
a more accurate view of their reading abilities ress and needs. Some ASRs may be unable to
than students in Profiles 1 and 3. This is the R effectively advocate for themselves in terms o f
8
largest profile and contradicts studies tha t needing further assistance and support give n
show students with disabilities have an iPncor- their inability1 to accurately judge that need.
rect perception of their academic abilities an d Prof0ile 3 was the smallest group (n = 8) and

tend to see themselves as acadePmically compe- demonstrated alarmingly low reader self-per-
2
tent (Meltzer, Katzir-Cohen, Miller, & Roditi, ception. Sound consideration must be given by
2001; Meltzer, Miller, KAatzir-Cohen, & Roditi, both researchers and practitioners to the under-
©
2000; Meltzer, Roditi, Houser, & Perlman , lying causes of the negative self-perception s
1998). The demIographic makeup of ASRs i n exhibited by these students. Just as overesti-
this profile varied. In other words, profile mating abilities can be damaging, underesti-
membership was not significantly associate d mating abilities can feed negative feelings tha t
with a particular school, gender, age, or dis- contribute to decreased motivation in not onl y
ability status. Similarly, the demographi c reading, but all other academic areas (Baker &
makeup of Profile 2 did not differ significantl y Wigfield, 1999). The persistence of ongoing
from that of Profiles 1 or 3. difficulties in reading are often reflected i n
Remarkably, ASRs in Profile 1 rated them- self-perceptions that are intimately linked t o
selves similar to the normative data of typi- reading experiences (Chapman & Tunmer ,
cally achieving students (Henk & Melnick; 2003). Students with low reader self-percep-
1995). This is an important insight that corrob- tion may need additional support in the affec-
orates previous research suggesting that there tive domain in order to enhance their academi c
are large discrepancies between the way some performance. Although their performance o n
struggling readers perceive their reading abil- academic outcomes may not differ signifi-
ity and their actual performance, an idea som e cantly from their peers, the socioemotiona l
refer to as an inflated self-concept (Kloomo k support required for these ASRs may differ
& Cosden, 1994). ASRs both with and without from ASRs in other profiles. Teachers should
disabilities were clustered in Profile 1, high- provide these students with strategies that
lighting the fact that inflated self-concept is enhance their opportunities to succeed and to
not specific to students with disabilities. This feel successful (Chapman & Tunmer, 2003).
finding corroborates previous research (i.e., The results of the present study underscor e
Klassen, 2002) on the extent to which students the idea that even if ASRs are similar academ-
Using Cluster Analysis to Explore Differences in the Reader Self-Perceptions of Adolescent Struggling Readers 47
ically, they can be markedly different in thei r administering the RSPS longitudinally. Under-
reader self-perception. It is important not t o standing the factors that both positively an d
look at middle schools students through a negatively impact the way ASRs see them-
homogeneous lens, but instead from a flexible selves as readers is another important area for
lens that recognizes the complexities of ado- future research. Qualitative data collection an d
lescence and reading difficulties. Most of th e methods of analyses could provide a mor e
ASRs in this sample have struggled with read- nuanced account of these changes and further
ing for 3 to 5 years without making appreciabl e explain underlying mechanisms that influenc e
gains. The three profiles demonstrated b y reader self-perception among different group s
these ASRs sheds light on the differences that of ASRs.
exist within reader self-perceptions despite the
commonness of limited achievement. S
CONCLUSION
F
Limitations In examining the complexities underlying th e
O
process of reading, both cognitive and affec-
There are two limitations in this study tha t
tive factors need to be considered as both pla y
must be acknowledged when considering thes e O
a significant role in reading behavior and aca-
findings in light of the broader population of
demic outcomes. Across the three profile s
ASRs and the inherent complexity in assessin g R identified in the p8resent study, it is clear that
reader self-perception. First, the study sampl e
P many ASRs1 are internalizing their negative
was relatively small and restricted to 6th grad-
perceptions of their reading progress, the ways
ers in just three schools within one district. in whi0ch others view their reading, their ow n
Second, though self-report is Pthe most com- c2omparisons to peers, and their feeling s
mon method of measuring academic self-per- toward reading. Whereas one cluster of ASR s
A
ception (Bong & Skaalvik, 2003), there a©r e showed evidence of an inflated self-concept,
inherent limitatiIons in self-report measures, another demonstrated extremely low reader
particularly with student participants. The self-perception. This suggests that ASRs
inclusion of semistructured interviews in addi- should be characterized as heterogeneous i n
tion to the self-report measure would have sub- terms of their reader self-perception. Ulti-
stantiated and further strengthened th e mately, understanding the differing ways in
findings. Therefore, whereas the present stud y which these ASRs construe themselves as
provides an important contribution to reade r readers has important implications for th e
self-perception research, its findings should b e nature of support provided to them in the class-
interpreted cautiously in light of these limita- room.
tions.
Acknowledgment: Preparation of this man-
uscript was supported by a grant from the Insti-
Future Directions
tute of Education Sciences (R324A140003).
It is important to further investigate the dif-
ferences in reader self-perception displayed b y
REFERENCES
older ASRs with additional quantitative and
qualitative research endeavors. Future research
Altman, D. G. (1999). Practical statistics fo r
might explore the extent to which RSPS dat a
research. New York, NY: Chapman & Hall/
are related directly to the reading achievemen t
CRC Press.
outcomes of ASRs. Additional work might
Baker, L., & Wigfield, A. (1999). Dimensions o f
also investigate the extent to which the reader children’s motivation for reading and their rela-
self-perception of ASRs changes over time b y tions to reading activity and reading achieve-
48 Middle Grades Research Journal Vol. 12, No. 1, 2018
ment. Reading Research Quarterly, 34, 452– measuring how children feel about themselve s
477. doi:10.1598/RRQ.34.4.4 as readers. The Reading Teacher, 48, 470–482.
Bandura, A. (1977). Self-efficacy: Toward a unify- Retrieved from http://literacyreferencelem
ing theory of behavioral change. Psychologica l .pbworks.com/f/Henk___Melnick_1995_RSPS
Review, 84(2), 191–215. .pdf
Baştug, M., & Çelik, T. (2015). An examination o f Jaffe, L. E. (2009). Development, interpretation,
secondary school students’ levels of reader self- and application of the W score and the relativ e
perception in terms of gender, grade, reading proficiency index (Woodcock-Johnson II I
environment and frequency. Journal of Human Assessment Service Bulletin No. 11). Rolling
Sciences, 12, 903–919. doi:10.14687/ijhs Meadows, IL: Riverside.
.v12i1.3120 Klassen, R. (2002). A question of calibration: A
Bong, M., & Skaalvik, E. M. (2003). Academic self- review of the self-efficacy beliefs of students
concept and self-efficacy: How different are they with learning disabilitiesS, Learning Disabilitie s
really?. Educational Psychology Review, 15, 1– Quarterly, 25, 88–102. doi:10.2307/1511276
40. doi:10.1023/A:1021302408382 Kloomok, S., & CosdFen, M. (1994). Self-concept in
Calhoon, M. B., & Petscher, Y. (2013). Individual
children with learning disabilities: The relation-
and group sensitivity to remedial reading pro- O
ship between global self-concept, academic “dis-
gram design: Examining reading gains across
counting,” nonacademic self-concept, and
three middle school reading projects. Reading O
perceived social support. Learning Disabilit y
and Writing, 26, 565–592. doi:10.1007/s11145-
Quarterly, 17(2), 140–153.
013-9426-7 R Melnick, S. A., He8nk, W. A., & Marinak, B. A .
Chapman, J. W., & Tunmer, W. E. (2003). Reading
(2009). Validation of a Reader Self-Perception
difficulties, reading-related self-perceptionPs, and 1
Scale (RSPS2) for use in Grades 7 and above. In
strategies for overcoming negative self-
NER0A Conference Proceedings 2009 (Paper 11).
beliefs. Reading & Writing Quarte rly, 19, 5–24 .
P East Hampton, CT: Northeastern Educationa l
doi:10.1080/10573560308205 2
Research Association. Retrieved from
Conradi, K., Jang, B. G., & McKenna, M. C. (2014).
A digitalcommons.uconn.edu/ nera_2009/11
Motivation terminology in reading research: A
© Meltzer, L., Katzir-Cohen, T., Miller, L., & Roditi ,
conceptual review. Educational Psychology
I B. (2001). The impact of effort and strategy us e
Review, 26, 127–164. doi:10.1007/s10648-013-
on academic performance: Student and teacher
9245-z
perceptions. Learning Disability Quarterly, 24,
Guthrie, J. T., & Davis, M. H. (2003). Motivating
85–98. doi:10.2307/1511065
struggling readers in middle school through an
Meltzer, L. J., Miller, L., Katzir-Cohen, T., &
engagement model of classroom practice. Read-
Roditi, B. (2000). Effort, strategy use, and aca-
ing & Writing Quarterly, 19, 59–85. doi:10.1080
demic performance: Student and teacher per-
/10573560308203
spectives. Paper presented at the Internationa l
Hair, J. F., Black, W. C., Babin, B. J., & Anderson,
Academy for Research in Learning Disabilities ,
R. E. (2010). Multivariate data analysis (7th
British Columbia, Canada.
ed.). Upper Saddle River, NJ: Pearson Prentice-
Hall. Meltzer, L., Roditi, B., Houser, R. F., Jr., & Perl-
Henk, W. A., Marinak, B. A., & Melnick, S. A. man, M. (1998). Perceptions of academic strate-
(2013). Measuring the reader self‐perceptions of gies and competence in students with learning
adolescents: Introducing the RSPS2. Journal o f disabilities. Journal of Learning Disabilities, 31,
Adolescent & Adult Literacy, 56, 311–320 . 437–451. doi:10.1177/002221949803100503
doi:10.1002/JAAL.00144/ Montague, M., & van Garderen, D. (2003). A cross-
Henk, W. A., & Melnick, S. A. (1992). The initial sectional study of mathematics achievement,
development of a scale to measure “perception estimation skills, and academic self-perception
of self as reader.” In C. K. Kinzer & D. J. Leu in students of varying ability. Journal of Learn-
(Eds.), Literacy research, theory, and practice: ing Disabilities, 36, 437–448. doi:10.1177/
Views from many perspectives (pp. 111–117) . 00222194030360050501
Chicago, IL: National Reading Conference. Pajares, F. (1996). Self-efficacy beliefs in academi c
Henk, W. A., & Melnick, S. A. (1995). The Reader settings. Review of Educational Research, 66,
Self-Perception Scale (RSPS): A new tool fo r 543–578. doi:10.3102/00346543066004543