Professional Documents
Culture Documents
1
ANXIETY IN ACADEM IC ACHIEVEMENT SITUATIONS
RICHARD ALPERT RAL P H NORM AN HABER
AKD
Harvard University Yale University
http://slidepdf.com/reader/full/alpert-haber-1960-anxiety-in-academic-achievement-situations 1/9
7/30/2019 Alpert & Haber 1960 Anxiety in Academic Achievement Situations.
divergent positions regarding this matter. One experiences in a testing situation" (Mandler &
position is exemplified by Taylor (1953) with Sarason, 1952, p. 166). Thus, the TASprovides
her Manifest Anxiety Scale and the other by a score indicating the recalled intensity of
Handler and Sarason (1952) with their Test certain experiences and behaviors immediately
Anxiety Scale. antecedent to or concomitant with the taking
Taylor's general anxiety scale is made up of of various types of examinations in the past.
items drawn from the MMPI and is concerned The implication of this type of scale is that
with a wide variety of situations other than the increased situational specificity of its item
test taking. Underlying the construction of the content will allow for a more sensitive meas-
M AS is a theoretical assumption that there urement of anxiety and its effect in the aca-
is a relatively constant "level of internal anx- demic achievement situation. It follows that
iety or emotionality," and also "that the for other situations other specific scales will be
intensity of this anxiety could be ascertained needed; and, only if these specific scales, each
by a paper-and-pencil test consisting of items shown to be valid in its own situation, turn
describing what have been called overt or out to be highly intercorrelated may one posit
manifest symptoms of this state" (Taylor, a general anxiety state and thus justify a
1953, p. 285). If Taylor is correct in positing single measure of anxiety for all situations.
such a general anxiety state, then it should The relative merits of situational specificity
follow that a single measure of a set of manifest versus generality of item content can be
anxiety responses gathered from many situa- evaluated by two statistical techniques. The
tions would be an adequate predictor of the first is a comparison of the intercorrelations
presence and effects of anxiety responses in among a variety of general and specific scales
any situation, whether it be eyelid conditioning administered to the same subjects. The Taylor
(Spence & Taylor, 1951; Taylor, 1951), thera- M AS , the Welsh AI, and the Freeman AS are
peutic sessions (Peck, 1951), serial learning used as measures of general or manifest anx-
(Taylor &At
formance. Spence,
present,1952), or academic
however, per-
there is some iety;
scalesthe
of Mandler-Sarason
the AAT are used TASas and the two
measures of
evidence (Child, 1954; Mandler, 1954) of specific anxiety. High intercorrelations among
w/ra-individual differences in anxiety both in all six scales would support the theoretical
content and intensity from one situation to assumption of a single underlying state, and
another. This evidence warrants further exam- thus the appropriateness of a single general
ination of the basic general-anxiety-state measure. On the other hand, low intercorrela-
assumption underlying the construction of the tions would support the use of scales which
Taylor MAS-type scale before applying such a are specific to the situations in which they are
scale as extensively as has been done in the to predict. Table 1 reports the intercorrelations
past. Too often, conclusions have been drawn of the measures of anxiety for an N of 40
on the basis of correlations with the MAS (the only sample that received all six scales).
regarding the presence or absence of anxiety Because in all instances inspection of the
or the effects attributable to anxiety without scatter diagrams revealed no departure from
du e consideration of the possible limitations linearity, only product-moment correlation
of a general scale of this type as a sensitive coefficients were computed.
indicator of anxiety in any limited, recurring As Table 1 indicates, the correlations among
type of situation. the general anxiety scales range from .32 to .39.
Mandler and Sarason, who represent the The correlations between the general and
alternative current position, maintain that the specific scales range from .24 to .38. The cor-
items composing the measuring instrument relations among the specific anxiety scales
should be concerned with the specific situations range from .40 to .64, all of which appear to
in which it is to be used. These investigators, be higher than any of the correlations involving
studying stress in academic achievement the general scales. This observation is sup-
situations, devised the Test Anxiety Scale ported by significance tests of the differences
(TAS), "... a questionnaire which was spe- between correlations, which indicate that in
cifically concerned with the Ss attitudes and most instances the intercorrelations among
http://slidepdf.com/reader/full/alpert-haber-1960-anxiety-in-academic-achievement-situations 2/9
7/30/2019 Alpert & Haber 1960 Anxiety in Academic Achievement Situations.
TABLE 1 TABLE 2
INTESCORRELATION A M O N G Six MEASURES OF ANXIETY CORRELATIONS BETWEEN Six ANXIETY SCALES AND
(N =• 40) FOUR MEASURES OE ACADEMIC PERFORMANCE
(N = 40 to 379)
Test REL AI AS TAS AAT- AAT+
Final Midterm
Course Exam Exam
Grade- Grade Grade Grade
Test Point
General Anxiety Average
Scales:
Psych 1 Psych 1 Psych 1
MA S .89 .39*" .32* .32* .38* -.33*
AI .84 .34* .2 8 .37* -.25
AS .73 .38* .30 -.24 General Anxiety
Scales:
Specific Anxiety MA S .01 -.08 -.02 -.19
Scales: AI -.04 -.05 -.03 -.22*
TA S .82 .64** -.40* AS -.06 .14 .15
AAT- .87 -.48** —
AAT+ .83 Specific Anxiety
Scales:
Note.—Reliabilities of the scales appear in Column 1. TAS -.24* -.21 -.16 -.32*
a
Based on the nonoverlapping items. When overlapping items AAT- -.35* -.26* -.28* -.25*
the specific scales are significantly higher than correlations, based on combined samples with
the correlations involving the general scales total Ns ranging from 40 to 379. Thus, these
either with each other or with the specific relationships are based on all samples for
scales. This seems to throw some doubt on which both the appropriate anxiety and per-
the comparability and, therefore, substituta- formance scores were available. Before averag-
bility of a general anxiety scale for a specific ing was attempted, a check was made to be
anxiety scale.
A factor analysis w as also performed (Alpert, sure that neither
deviations means and
the samples
of all the thestandard
northe correla-
1957) on this matrix of correlations, but al- tions to be averaged were significantlydif-
though the above conclusions were supported ferent. No differences were found in any of
by this analysis, such support is little more these comparisons.
than suggestive because of the hazards of The data summarized in Table 2 indicate
drawing conclusions from a factor analysis that the specific anxiety scales are more often
based on only 40 cases. significantly correlated with academic per-
A second technique for determining the formance measures than are the general anx-
substitutability of the general for the situa- iety scales. Accepting the 5% level of signifi-
tionally specific anxiety scales, far more cance (two-tailed) as a criterion, only one of
rigorous than merely intercorrelating the the 11 available correlations for the general
scales, is to compare the two types of scales scales reached significance, while 9 of the
as predictors of various measures of academic 12 correlations involving the specific scales
performance. If the general scales predict as reached significance.
well as the specific scales, there is justification The implications of the findings are reason-
for the use of a single general measure for the ably clear. Specific anxiety scales and general
detection of anxiety in all situations. If, on anxiety scales measure, to a significant extent,
the other hand, the specific scales are able to something different. Furthermore, it appears
account for significantly more of the variance that the variable which the specific scales
in the performance measures than are the measure, and which the general scales do not,
general scales, then there is justification for is involved in academic performance to such
the use of scales which are specific to the situa- an extent that the specific scales are better
tion in which they are used. predictors of academic performance than are
Table 2 presents the correlations between the general anxiety scales.
the six anxiety scales and four measures of One may tentatively conclude, therefore,
academic performance. These are average that further use of a general anxiety scale as
http://slidepdf.com/reader/full/alpert-haber-1960-anxiety-in-academic-achievement-situations 3/9
7/30/2019 Alpert & Haber 1960 Anxiety in Academic Achievement Situations.
http://slidepdf.com/reader/full/alpert-haber-1960-anxiety-in-academic-achievement-situations 4/9
7/30/2019 Alpert & Haber 1960 Anxiety in Academic Achievement Situations.
measure
as is, because
its format, simplyofasome characteristic
measure of aptitudesuch
it- AAT- 93 .49 -.48 .30 .001
self, then one would expect that the aptitude AAT- 92 .42 -.02 .42 ns
AAT- 98 .35 -.34 .16 ns
and anxiety measures would account for the TA S 40 .40 -.35 .25 .05
same variance in academic performance . If this AAT+ 93 .36 .25 .30 .05
were, in fact, the case, then a multiple correla- AAT+ 92 .47 .2 3 .42 .05
AAT+ 98 .32 .28 .16 ns
tion based on the best weighted com bination of
aptitude and anxiety would not be expected to a
Final examination grade— Anxiety & Aptitude.
b
account for significantly more of the variance Final examination grade— Anxiety.
0
Final examination grade— Aptitude.
in academic performance than would aptitude
alone. Tables 4 and 5 present multiple correla-
tions of aptitude and anxiety, p redicting grade- findings refute the argument that the specific
point average and final examination grade. scales which are correlated with aptitude are
Also, they indicate the significance of the addi- simply measures of aptitude, because the
tion of the anxiety measure in the predictions. results indicate that specific anxiety scales are
The general anxiety scales are not included in able to account for variance in academic
the tables because they are not correlated w ith performance other than that accounted for by
achievement performance and hence, could a measure of aptitude. This does not mean to
make no contribution to the prediction of sa y that these anxiety instruments are not,
such perform ance. in part, measures of intellectual ability; it
The debilitating scale of the AAT adds does say, however, that they are something
significantly, over and above aptitude, to the more than that— and that therefore they are
prediction of the grade-point average (GPA) able to account for added variance in academic
in two out of three instances, and in one out performance, and, as such, are valuable supple-
of three for the final examination grade. Th e ments to a measure of aptitude.
facilitating scale of the AAT adds signifi- The second explanation holds that the
cantly, over and above aptitude, to the predic- significant correlation of specific anxiety scales
tion of GPA in the three instances reported, with ap titude is really the result of a m ediating
and in two out of the three instances for the process of rationalization and justification for
final examination grade. The Mandler-Sarason poor performance. In order to evaluate this
TA S adds significantly to prediction of final argument, the entering freshman class was
examination grade, but not for GPA, in the given a specific anxiety scale. Freshmen enter-
http://slidepdf.com/reader/full/alpert-haber-1960-anxiety-in-academic-achievement-situations 5/9
7/30/2019 Alpert & Haber 1960 Anxiety in Academic Achievement Situations.
for whom exactly comparable data were on for the third explanation of the relationship
hand, no significant differences were found. between aptitude an d anxiety. This experi-
This is a weak refutation of the second ment would be similar to that by Pomeroy
explanation, at least at the college level. But, (1950), who demonstrated increases in finger-
of course, college students have had anxiety maze performance after only 10 min. of
experiences regarding high school achievement, cathartic-type therapy. Until such time as ap-
and thus, it is still possible that the freshmen propriate data are available, however, one
could be rationalizing via the AAT for their must hold in abeyance any final conclusions
high school grades. But the two scales were regarding the validity of this explanation.
correlated w ith high school grade-point average In summ ary, regarding these three explana-
.29 and — .2 1, respectively. These correlations tions, the first is refuted, doubt cast on the
are both significant at the 5 % level, and they second, and the third neither refuted nor
are not significantly different from .42 and supported. Present evidence allows one to say
— .29. Thus, it seems the da ta can not provide a only that either the third or the fourth explana-
definite answer to the issue involved— nor, tion (that is, that the obtained correlation is a
probably, can any correlational study of this measure of the relationship between the two
nature. constructs— anxiety an d aptitude) may be
Third, if a timed aptitud e test arouses stress correct, or that both of them may be true.
cues, which elicit anxiety responses which Had the first and second explanations been
affect the aptitude test performance, then a supported, they would support a preference
correlation between an anxiety measure and for the use of a general anxiety scale because
an aptitude measure could be interpreted to it is free of the contaminating influence of
mean that the anxiety scale measures the aptitude. The last two explanations support a
anxiety which is a component in the aptitude preference for a specific anxiety scale because
test situation and which is reflected in the of its greater sensitivity. Since the evidence
http://slidepdf.com/reader/full/alpert-haber-1960-anxiety-in-academic-achievement-situations 6/9
7/30/2019 Alpert & Haber 1960 Anxiety in Academic Achievement Situations.
A N X I E T Y IN A C A D E M I C A C H IE V E M E N T S IT U A T I O N S 213
in the Mandler-Sarason TAS. All the items of tation. This assumption may be unnecessary.
the TAS are unidimensional, i.e., anxiety In fact, these two constructs of debilitating
responses are either debilitating or not (e.g., and facilitating anxiety may be uncorrelated.
"While taking a course examination, to what Thus, an individual may possess a large
extent do you worry? Worry a lot... [to]... amount of both anxieties, or of one but not
Worry not at all." [Item No. 35]). From the the other, or of none of either. The nature of
absence of negative responses for an S, that this correlation can be determined empirically
is, a low score on the test, they infer that following the construction of two such inde-
when anxiety-provoking cues are present in pendent measures of anxiety, a scale measuring
the environment, "these stimulus elements the facilitating effects of anxiety on achieve-
will raise [his] general drive level and result in ment performance, and a separate scale meas-
improved performance for that S." (Sarason, uring the debilitating effects of anxiety on
Handler, & Craighill, 1952, p. 561). This achievement performance.
results in their confounding the two alter- The Achievement Anxiety Test (AAT) was
natives of a facilitating effect of anxiety and constructed to make this determination. It
no effect of anxiety on academic performance consists of two independent scales: a facili-
and leads to their failure to allow for a third tating scale of nine items based on a prototype
alternative, the possible existence of an indi- of the item—"Anxiety helps me to do better
vidual whose anxiety responses in an anxiety- during examinations and tests"; and a debili-
provoking situation do not affect test per- tating scale of 10 items based on a prototype
formance either by improving or by de- of the item—"Anxiety interferes with my
pressing the score. Such an individual would performance during examinations and tests."
have to be described in their formulation as Both scales have gone through numerous
one whose negative or debilitating response revisions based upon item analyses, correla-
effect is just counteracted by the increased tions with various criteria, and theoretical
task-performance drive.
There is another approach which does not reformulations.
a 10-week interval aretest-retest
The reliabilities
.83 and .87, for
respectively.
require this unwarranted inference and accord- The test-retest reliability over an 8-month
ing to which an anxiety scale would measure period is .75 for the facilitating scale and .76
the presence and intensity of both kinds of for the debilitating scale. The two scales are
anxiety responses, those which facilitate per- administered in one questionnaire, the items
formance and those which interfere with it. randomly mixed (as presented below). The
If an individual scored low on both of these 5s answer each item on a five-point scale,
scale components, he would be considered indicating the degree to which the item applies
to be unaffected (insofar as test performance to them. The AAT includes neutral buffer
was concerned) by anxiety-provoking cues. In items in addition to the items listed. The
other words, the facilitating effects of anxiety numbers in the parentheses to the right are the
would then be measured independently and actual item numbers on the AAT.
not inferred from the absence of negative
Facilitating Anxiety Scale
responses. This independent measure would
1. I work most effectively under pressure, as when
allow for the possibility of the absence of the task is very important. Always—Never. (2)
both types of response as well as for the 2. While I may (or may not) be nervous before tak-
presence of either one. ing an exam, once I start, I seem to forget to be nervous.
I always forget— I am always nervous during an exam.
Theoretically, Handler and Sarason would
(9)
predict, on the basis of their ideas, that two 3. Nervousness while taking a test helps me do
such independent measures of anxiety would better. It never helps—It often helps. (11)
4. When I start a test, nothing is able to distract me.
be highly negatively
individual correlated.
had a great deal ofThat is, if an
facilitating
This is always true of me—This is not true of me. (12)
5. In courses in which the total grade is based
anxiety, he would have little debilitating mainly on one exam, I seem to do better than other
anxiety and vice versa. This is the assumption people. Never—Almost always. (14)
underlying their scale because it allows them 6. I look forward to exams. Never—Always. (16)
7. Although "cramming" under pre-examination
to use only the debilitating index and to infer tension is not effective for most people, I find that if
facilitating anxiety from the absence of debili- the need arises, I can learn material immediately before
http://slidepdf.com/reader/full/alpert-haber-1960-anxiety-in-academic-achievement-situations 7/9
7/30/2019 Alpert & Haber 1960 Anxiety in Academic Achievement Situations.
an exam, even under considerable pressure, and success- intercorrelation of the scales without affecting
fully retain it to use on the exam. I am always able to their validity coefficients. The final correla-
use the "crammed" material successfully— I am never
able to use the "crammed" material successfully. (19) tions between the facilitating and debilitating
8. I enjoy taking a difficult exam more than an scales were — . 37 , — .34, — .43, and — .4 8,
easy one. Always— Never. (21)
9. The more important the exam or test, the better drawn from four different samples. Fo r these
four samples, th e average means an d standard
I seem to do. This is true of me— This is not true of
deviations for the facilitating an d debilitating
me. (24)
scales of the AAT were as follows: for the
Debilitating Anxiety Scale A A T + , A f = 27.28, S Z > = 4.27; for the AAT-,
M=26.33, SD=5.33. The average correlation
1. Nervousness while taking an exam or test hinders
me from doing well. Always— Never. (1) for the combined N (379) was — . 37 . All these
2. In a course where I have been doing poorly, my correlations are significant beyond the 1%
fear of a bad grade cuts down my efficiency. Never— level. Hence, in spite of efforts to separate
Always. (3) the two scales empirica lly, a low but significant
3. When I am poorly prepared for an exam or test,
correlation remains. The point seems made,
I get upset,should
knowledge and doallow.
less well
This than
nevereven my restricted
happens to me— however, that the correlation is neither perfec t
This practically always happens to me. (5) nor high as Handler an d Sarason imply an d
4. The mo re imp ortant the e xam ination, the less require as an assumption for the use of their
well I seem to do. Always— Never. (6) scale.
5. During exams or tests, I block on questions to
which I know the answers, even though I might re- In general, the two scales of the AAT cor-
member them as soon as the exam is over. This always related about equally, though opposite in
happens to m e— I never block on questions to which I sign, with the other variables discussed in this
know the answers. (10) paper (see Tables 1, 2, and 3). The important
6. I find that my mind goes blank at the beginning
o f an exam, and it takes me a few minutes before I can
question is whether, when predicting an aca-
function. I almost always blank out at first— I never demic performance score such as college grade-
blank
7. Iout
amatsofirst.
tired(IS)
from worrying about an exam, that point average,
both scales therethan
rather is any advantage
using merelyina using
con-
I find I almost don't care howwell I do by the time I ventional debilitating anxiety scale such as
start th e test. I never feel this w ay— I almost always
feel this way. (17)
the AAT— alone or the Handler an d Sarason
8. Time pressure on an exam causes me to do worse TAS. These tw o measures of debilitating
than the rest of the group under similar conditions. anxiety (the AAT— and the TAS) correlate
Time pressure always seems to make me do worse on .64 with an N of 40, indicating considerable
an exam than others— Time pressure never seems to
make me do worse on an exam than others. (18) equivalence. Table 6 answers the question of
9. I find myself reading exam questions without whether a scale which measures the facilitating
understanding them, and I must go back over them so effect of anxiety justifies itself. In the last
that they will make sense. Never— Almost always. (23) two columns to Table 6 it can be seen that
10 . When
beginning I don't
of an exam,doitwell a difficult
on to
tends upset meitem at theI
so that for three separate samples, the multiple cor-
block on even easy questions later on. This never relations, using both the plus an d minus scales
happens to me— This almost always happens to me. (26) to predict grade-point average, are signifi-
TABLE 6
Part of the refinement of the scales was MULTIPLE C O R R E L A T I O N S P R E D I C T I N G G R A D E - P O I N T
accomplished through means of empirical A V E R A G E (GPA) F R O M THE AAT
techniques. After a large number of items ha d
N r* c 1
been constructed fo r each scale, the scales -R" r r< » #" t
were used to predict several performance
93 .50 .36 -.45 -.37 .001 .0 5
criteria, such as grade-point averages an d 92 .32 .32 -.08 -.34 ns .01
final examination grades. These data were item- 96 .54 .50 -.40 -.43 .05 .001
analyzed to give the correlation of each item
* GPA • AAT+ & AAT-.
with the criterion. Those items were retained b
GPA • AAT+.
which were highly correlated with the criteria 0
GPA • AAT-.
d
AAT- AAT+.
but which were not correlated w ith each other. 8
•
http://slidepdf.com/reader/full/alpert-haber-1960-anxiety-in-academic-achievement-situations 8/9
7/30/2019 Alpert & Haber 1960 Anxiety in Academic Achievement Situations.
http://slidepdf.com/reader/full/alpert-haber-1960-anxiety-in-academic-achievement-situations 9/9