SELF-PERCEPTIONS OF “AT RISK” AND HIGH
ACHIEVING READERS: BEYOND READING RECOVERY
ACHIEVEMENT DATA
Sheila G. Cohen
SUNY Cortland
Gloria McDonell
Fairfax County Public Schools
Bess Osborn
‘Sam Houston State Universiy
This study looked at the impact of Reading Recovery on students’ beliefs about
their competence and capacity to direct their own learning activities. Causal atribu-
tions for success or failure (Weiner, 1972, 1979) and self-efficacy (Bandura, 1977), in
“other words, perceived competence to do various literacy activities, were selected as
measures of these variables. Bandura has maintained that self-efficacy can be consid
fered one's competence to accomplish a task. These self-appraisals affect a child's
choice of activities and their effort and persistence (Bandura, 1977). Weiner (1972,
1979) also suggested that an individua!’s achievement behavior will be a function of
the cognitive representation of related events, activated by a stimulus, that will produce
an expectation of success and failure. That is, if one is successful in performing a task,
fone is more likely to attribute future success to ability or effort. ‘Thus, successful
experiences ate likely to change children’s belief systems such that they believe they
are more competent, more in control, more likely to succeed with similar tasks and are
more likely to be motivated to engage in similar learning activities.
Reading Recovery isa tutorial intervention program for at-tisk first-grade students
as well as an inservice training program for teachers (Clay, 1985). Students in this
program work for 30 minutes each day with a teacher who has been trained to make
accurate instructional decisions. The program goals are to teach students to understand
the reading process, learn leter-sound relationships, and gain fluency through writing
tnd teading of progressively more difficult text. Students are dismissed from the
program as soon as they enter the middle reading group in their class, usually after an
average of 18 weeks of instruction,
More than simply designed to raise levels of functioning in reading, the goal of
Reading Recovery is to help students who are at-risk develop a self-improving system
that enables them to self-monitor and solve their own reading problems by helping
them discover strategies for themselves (Clay, 1972, 1985; Pinnell, 1985).Unlike other intervention or remedial reading programs where dependence on
others for learning is often a result, Reading Recovery seems to have an impact on
achievement motivation in that it appears to facilitate a sense of increased ability and
control over one’s reading and learning (Clay, 1972, 1985). In an earlier investigation
to document this claim within Social Cognitive Theory paradigms (Bandura, 1977;
Weiner, 1972, 1979), Cohen, Cohen, McDonell, and Osburn (1987) conducted a study
in which Chapter I Reading Recovery students were compared to other Chapter T
students who were in a highly successful remedial program. The results indicated that
both groups of first-grade children could make causal attributions and judge their own
competence to do school related work. However, the Reading Recovery students more
frequently attributed their success in various school tasks to their own ability and hard
work and also expressed greater feelings of competence to carry out various reading
and writing tasks in the classroom than did other at-risk students. In addition, anecdotal
data from classroom teachers indicated that Reading Recovery students indeed did use
independent strategies for solving their own reading problems.
In this study we attempted to examine the influence of Reading Recovery on
children’s self-perceptions by comparing Reading Recovery students not only with
other at-risk students but with high-achieving students as well. We addressed the
following specific questions: (a) Do Reading Recovery students attribute their success
to ability, effort, mood, task difficulty, or teacher help more than other at-risk students
and as often as high-achieving readers? (b) Do Reading Recovery students feel more
competent to do reading and writing activities than other at-risk students and as compe-
tent as high-achieving readers?
METHOD
Population
Participants were 138 first graders in a predominantly upper middle-class school
district outside of Washington, D.C. Ninety-cight of these students were at-risk and
well below the SES of the majority of students in this district. Fifty were in Reading.
Recovery and 48 were in other remedial programs. The additional 40 students were
high-achieving readers selected from the same schools as the at-risk students, The 98
at-risk students were identified for special services because they all scored below the
40th percentile on a standardized test at the end of their kindergarten year. Reading
Recovery students were selected from the lowest scores and enrolled in that interven-
tion program at different times during the 1987-88 school year, They were discon-
tinued from the program after an average of 18 weeks because they were able to work
Within the middle reading group in their class and were fet to have acquired sufficient
self-directed independent strategies to continue to improve in reading and writing on
their own,
‘Students in the remedial program were provided daily instruction for 45 minutes in
small groups of five or six. Reading specialists used an integrated language arts
approach and students in this program typically make NCE gains well above the
national average.Beyond Recovery Data 9
‘Materials and Procedure
Attribution was measured with a scale designed for our previous study (Cohen et
al, 1987) using as a model the ‘*What Kind of Student Am 1?” scale developed by
‘Cohen (1983). Four items were included, each of which specified a particular behavior
in the stem and then required students to respond to each of five causal statements thet
‘completed the sentence. The four behavior stems were the following: (a) Sandy usually
Jinishes worksheets on time because; (b) Sandy usually follows directions for school
work because; (c) Sandy usually figures out new words when reading because; (8)
Sandy understands what she reads because. The five causal attributions following each
of the above stems read as follows: (a) Sandy is smart most ofthe time; (b) Sandy works
hhard most ofthe time; (c) Sandy is ina good mood most of the time; (A) Sandy has easy
work to do most of the time; and (e) Sandy gets help from the teacher most of the time.
‘The five causal attribution statements were to be rated either not like me, a little like
‘me, ot just like me. To assess the internal consistency of the scale Cronbach’s alpha
‘was computed. The standard item alpha was .81, indicating that this scale was fairly
reliable.
‘The self-efficacy measure was developed using a method similar to that described
by Schunk (1985). On the scale we created, there were seven statements like J can read
books by myself, I can read books to someone else, and so forth (soe Table 3 for all
items); students were asked to respond not often, sometimes, or always in regard to
each. Although this instrument was administered as a 3-point scale, it was modified
when tabulating the responses, because we, as well as independent observers, agreed
there was litte or no difference between the first two categories (nor often and some-
times). Because the internal consistency of this messure, computed with Cronbach's
alpha, was .71, we collapsed the first two categories and treated this variable as
nominal rather than ordinal. The attribution and self-efficacy scales were individually
administered to all participants at the end of the 1987-88 school year by trained
RESULTS
‘To answer the first rescarch question, means and standard deviations were com-
puted for each of the attribution variables by group. These data are summarized in
Table 1. Correlations were also computed between these independent variables (Table
2), Because these data indicated there was a moderate to strong relationship between,
these variables, a MANOVA was computed to determine whether there were differ-
ences between the three groups on the attribution scale. The MANOVA indicated that
there were significant differences between the groups, F(2, 135)=4.05; p>001. Uni-
variate ANOVAs indicated that there were significant differences for effort, ability,
and mood but not for the other two attributions, task difficulty and teacher help,
To determine where the significant differences were, contrasts were computed.
Results of these contrasts indicated that there were significant differences between
Reading Recovery and other at-isk students om three of the causal attributions, ability
(PU, 135)= 14.56; p>.0001], effort (FUL, 135)=7.09; p=.009); and mood [F(1,Table 1
‘Means and Standard Deviations of Children's Attribution Scores by Group
‘Group
Other Reading
AcRiskt Recovery?
Ateibotion aM sD M sD
‘Ability 2.34 1982.64 sera 16) 1.09
Effort 2.45 Lg 2.67 Lay 28 129
Mood 2.45 2342.66 es 133
‘Teacher Help 215 266-2300 2.76 2a 292
‘Task Difhculty 197 298 2.06 211 168, 2.94
n= a8, "n=50, “nA,
Table 2
Interrcorrelations Among the Attributions :
Variable 1 2 3 4 5
1. Ability 1.00
2. Effort, sie 1.00
3. Mood eat are 1.00
4. Help 2st 24r 23" 1.00
5. Difficulty 10 alia ‘a1 04 1.00
Nowe 9137.
#01. **<.001,
Table 3
Proportion of "Always!" Responses to Self-Ejficacy Items by Group
Group
Other Reading High
AtRisk Recovery” __Achieving®
1. Tean figure out new words by myself 29 46 59
2. Lean write stories by myself 34 8 88
3. Lean spell words by myself 46 32 56
4. Tean write s0 someone can read it 56 ‘38 83
5. Lean read books by myself, 36 n 9s
6. Tcan read books to someone else 34 n 16
7.1 do good work on my workbook pages, 54 oe 75
Sn ab = 50, “nea,Beyond Recovery Data m1
135)= 10,39; p>.002]. Contrasts between Reading Recovery and high-achieving stu-
dents indicated there were no significant differences on any ofthese causal attributions.
‘To answer the second question, the researchers calculated percentages of students
in each group who responded always to the self-fficacy items. These percentages
(Table 3) indicate a clear tend, with percentages of always responses higher for
Reading Recovery students than for other at-risk students but lower than the percentage
of always responses for high-achieving students
DISCUSSION
Analysis of the attribution data indicated that Reading Recovery students mote
readily attributed their success in school to ability, effort, and mood than did the other
at-risk students, When comparing Reading Recovery students with high-achieving
first-grade readers, there were no significant differences found for any of the attibu-
tions, It would appear, therefore, that Reading Recovery students become much like
the group of high-achieving students; in other words, their attributional patterns are
similar, characterized by use of internal stable causes to account for achievement
Responses on the self-efficacy scale revealed a trend in the predicted direction
Reading Recovery students responded always a greater proportion of the time to each
of the self-efficacy items than the other at-risk students, but less often than the high-
achieving readers. This is an indication that Recovery students feel more competent
about their ability to do various reading and writing activities than the other at-risk
students. Theve were substantial differences between the Reading Recovery students
and the other at-risk students on five of the seven items. For those items where there
‘were minimal differences (items four and seven), the nature of the Recovery program
itself might explain this. Fewer always responses from Recovery students to item 4,
T-can write so someone can read, might be a reflection of the program emphasis on
strategies for reading; writing strategies are fostered for understanding sounds in
‘words, letter sequences, and so forth. Fewer always responses from the Recovery
students in reaction to item 7, [ do good work on my workbook pages, may also be
attributable to the program, Recovery students are not trained to do workbook pages.
TThe assumption is made that they will be able to do them and this is generally the case.
However, after doing a great deal of real reading in the Recovery program, workbook
pages may look tedious to these students.
‘What appears to be important about the findings, overall is that Reading Recovery
students, who wero judged to be the lowest functioning students in the kindergarten
population when they were selected for the program, not only could function in an
average reading group after treatment, but appeared more like high-achievers in their
autributions (ability, effort, mood, task difficulty, and teacher help). They also had
scores that indicated they felt mote competent on school-related tasks (self-efficacy)
than the other at-risk students. They were, however, not as high on these items as the
high-achieving group. With regard to self-efficacy, although it appears that students’
beliefs about their competence for selected learning tasks increase, they are not as high
as students who were initially high-achieving, which is not surprising. That ReadingRecovery students have self-learning strategies and continue to improve is an indice
tion that the statistical differences are also real differences.
‘The study has a number of limitations. Selection of students is @ problem i
‘esearch such as this in that it poses a threat to internal validity, yet a great deal i
learned when conducting research in real-life settings. Another limitation is that ther.
is no way of knowing if the students’ perceptions translate into self-directed learnin;
behaviors in the classroom. Follow-up longitudinal studies using observational tech
niques and interviews nced to be conducted to determine this.
‘This study suggests that the Recovery students perceive themselves to be compe
{ent and in contro! of their own learning, thus supporting the claims of Clay (1987)
Based on studies reporting gains in achievement due to Reading Recovery (Clay, 1987
land the findings of this and a previous study (Cohen et al., 1987) suggesting Rea
Recovery increases ability and effort attributions as well as self-efficacy, variables tha
hhave been demonstrated to mediate self-regulated classroom behavior and achievemen
‘motivation, it is suggested that educators consider adoption of this type of program fo
at-risk students,
REFERENCES
Bandura, A. (1970), Social learning theory. Englewood Cis, N: Prentice Hall
Gay, M: (1985). The early detect of reading eifculies. Poxsmenth, NH: Heinemann,
(Clay, M. (1987), Implementing reading recovery: Systematic adptations wo an edvatonlinovation, New
“Zealand Journal of Educational Stuer, 22, 35-58.
Coen, D. W. (1983). Children's temperament clusters, sei atbution of academie success and percevec
self conpeterce. Unpablised doctoral dissertation, University of Kentucky, Lexington,
Coen, S. G., Coben, D. W., MeDonnll,G., & Osburn, 8, (1987, Decembe). The effets of Reading
‘Recovery on achievement motivation. Paper presented atthe meeting ofthe National Reading Conor
‘nce, St Petersburg, FL.
Pinnll . S. (1985). Helping teachers elp children at risk: Insights ftom the reading recovery program,
Peabody Journal of Education, 62, 70-85,
Schunk, D. H, (1965) Selfffieay and classroom learning. Psychology in the Schools, 22, 208-223,
Winer, B. (1972). Awibition theory, achievement motivation, andthe edestional process. Review of
itucational Research, #2, 203-215,
Weiner, B. (197). A theory of motivation for some classroom experenées Journal f Educational Psychol
py. 72, 3-28