Professional Documents
Culture Documents
Student feedback sensitivity and the efficacy of feedback interventions in public speaking performance improvement (1)
Student feedback sensitivity and the efficacy of feedback interventions in public speaking performance improvement (1)
To cite this article: Camille D. Smith & Paul E. King (2004) Student feedback sensitivity
and the efficacy of feedback interventions in public speaking performance improvement,
Communication Education, 53:3, 203-216, DOI: 10.1080/0363452042000265152
Although feedback from teachers to students occupies a central role in learning and
instruction, the manner in which feedback functions in performance improvement
remains unclear. The current study examined how students’ sensitivity to feedback and
variations in the wording of feedback messages impacted public speaking performance.
Results indicated that students high in feedback sensitivity exhibited preferred speaking
behaviors when messages were worded in a manner that was less likely to be taken as
a direct, personal criticism (low intensity). Students low in feedback sensitivity appeared
to be less adversely impacted by high intensity messages. Overall, markedly different
relations between student feedback sensitivity and speaking behaviors appeared for
students exposed to high intensity feedback messages compared with students exposed to
low feedback intensity.
Camille D. Smith (MS, Texas Christian University, 2000) is a marketing representative for the Fort Worth
Visitor’s Bureau. Paul E. King (PhD, University of North Texas, 1985) is Professor of Communication Studies
at Texas Christian University, Fort Worth, TX. This manuscript was adapted from Ms. Smith’s thesis, directed
by Dr. King. A previous version of this manuscript was presented at the meeting of the Southern States
Communication Association, April 3–7, 2002, Winston-Salem, NC. Paul King can be contacted at
p.king@tcu.edu
Feedback Sensitivity
Developed in the area of interpersonal communication and social perception, the
concept of feedback sensitivity (Edwards & Pledger, 1990) is concerned with
individual differences in the processing of socially relevant information. While
development of the feedback sensitivity concept has been somewhat limited, the
generalized notion that individuals vary in their sensitivity toward the recognition,
processing, and internalization of messages is well established. For example, the
concepts of conversational sensitivity (Daly, Vangelisti, & Daughton, 1987; Stacks &
Murphy, 1993), sensitivity to emotional states (Booth-Butterfield & Booth-But-
terfield, 1990), social sensitivity (Garrard & Kyriacou, 1985), and sensitivity to
supportive communication (MacGeorge, 2001) are all somewhat related to these
features.
Student Feedback Sensitivity 207
H1: Overall, students who receive feedback of any intensity level will outperform
students who have received no feedback.
H2: A significant positive correlation will hold between sensitivity to feedback and the
production of desired behavior in the low-intensity condition.
H3: A significant negative correlation will hold between sensitivity to feedback and the
production of desired behavior in the high-intensity condition.
Method
Participants
Ninety-one students enrolled in basic communication courses served as participants
in the study. While participation was voluntary, students received either extra credit
or partial fulfillment of their course’s research requirement for participation.
Student Feedback Sensitivity 209
Instruments
Feedback sensitivity was operationalized using the Sensitivity to Feedback scale
developed by Edwards and Pledger (1990). While two of the factors (anticipation of
response and sensitivity to attention) are directed toward the give and take of
interpersonal dialogue, the remaining two factors were deemed appropriate for use
in classroom settings and were utilized in the present study. For example, personal
attributions are normally associated with classroom feedback (Booth-Butterfield,
1989). Students are concerned whether instructors are reacting positively or nega-
tively, with happiness or disappointment, to their work. The remaining two factors
in the feedback sensitivity instrument relate to these issues.
Sensitivity to socially undesirable feedback was composed of four Likert-type items
such as, “I can tell when I have upset someone.” This sensitivity to socially desirable
feedback factor included four items such as “I can tell when my conversational
partner is reacting positively to me.” Adequate reliability (Cronbach’s alpha) was
found for both factors: sensitivity to socially undesirable feedback (.87), and sensi-
tivity to socially desirable feedback (.75).
Procedure
Conceived as a post-test only experimental study, participants were asked to prepare
an informative speech of at least a five minute duration and were randomly assigned
to one of three groups: no feedback, low-intensity feedback, and high-intensity
feedback. After completing the feedback sensitivity scale, participants delivered their
speech in the setting of a pedagogical rehearsal with a researcher, but (for control)
no other audience present. Aware of the danger of experimenter bias, the researcher
made deliberate efforts to avoid any behaviors that could have been interpreted as
forms of nonverbal feedback (e.g., head nods, smiles, frowns, etc.) and endeavored
to provide similar instructional demeanor for all participants.
Following the presentation, the participants were subjected to experimental treat-
ment. Participants in the no feedback condition were given no information. Partic-
ipants in the low-intensity condition received feedback stated in tactful and
nonconfrontational language, while participants in the high-intensity condition
received feedback in direct and frank language. Feedback was delivered both orally
and in writing. In all cases, the feedback recommended increasing eye contact,
overtly giving references as evidence for claims made and increasing the length of the
introduction. These three areas were chosen in order to provide a range of behaviors
that included both mindful, deliberative aspects of speech construction and behavior
that must be executed automatically to some degree.
As an example, in the low-intensity condition, participants were told, “Your eye
contact needs improvement. You don’t appear to be looking out toward the
audience as frequently as you should or maintaining the eye contact when you do
look up. In your next speech, try to increase your eye contact.” Participants in the
high-intensity condition were told, “Your eye contact was bad in this speech. You
210 C. D. Smith & P. E. King
rarely look up and when you do glance toward the audience, it’s only for a moment.
Your next speech requires significant improvement in eye contact.” Similar instruc-
tions were given for introduction length and overtly referencing information sources.
Following the sensitization to feedback, participants were given one week to work
on their speeches. At a second presentation, participants were videotaped and
debriefed. Tapes were reviewed and dependent variables calculated as follows: eye
contact was operationalized as the number of times that the speaker appeared to
look up at the camera, number of references given in the text was counted, and
length of introduction was timed in seconds. Introduction end-point was deter-
mined when the speaker, following instructions, stated “body.”
Two researchers independently reviewed the tapes to count and time behaviors.
Inter-rater reliability was assessed using Scott’s , which contrasts observed and
expected levels of agreement. Satisfactory agreement was found for eye contact
( ⫽ .96) and for number of references utilized ( ⫽ .99). Speakers were explicitly
asked to state “introduction, body, and conclusion” in their speeches, ostensibly to
assist them in organizing the speech but primarily in order to provide a reference for
determining introduction length. Since timing of introduction length did not involve
judgments by coders (beginning and end-points clearly demarked by the state-
ments), any small differences in timing were assumed to be the result of issues
related to the physical activities of working a stopwatch.
Validation
It was of some interest to determine whether the factor structure originally reported
by Edwards and Pledger (1990) held up in the present study. All 18 scored items that
compromise the Sensitivity to Feedback Scale were utilized in a principal compo-
nents analysis. Four factors emerged (Eigenvalues ⬎ 1.0) and were subjected to a
varimax (orthogonal) rotation to purify the factor structure. Rotated sums of
squared loadings (expressed as percentage of variance accounted for) ranged from
17.41% for factor 1 (anticipation of response) to 12.55% for factor 4 (sensitivity to
attention) with a total of 60.81% of variance explained.
Importantly, the items which loaded on the two factors, sensitivity to socially
undesirable feedback (factor 2) and sensitivity to socially desirable feedback (factor
3) mirrored the results reported by Edwards and Pledger (1990). While one of the
items loading on factor 3 also had a secondary loading above .40 (.430) on factor 1,
the item was retained on the final instrument for the sake of consistency with the
original analysis and because factor 1 was not employed in the present study.
In order to check the intended effect of the experimental manipulation of
feedback intensity, a series of four items relevant to feedback were constructed and
arrayed on 7-point semantic differential scales. The first pair of items asked whether
“The feedback indicated my performance was …” (good/bad or negative/positive).
The second pair dealt with frankness in presenting feedback (direct/indirect and
ambiguous/frank and to the point). Both scales were highly reliable (alpha re-
liability ⫽ .96 and .98, respectively). As evidence of the efficacy of the feedback
Student Feedback Sensitivity 211
* p ⬍ .05.
Results
The first hypothesis predicted an overall main effect of experimental treatment on
public speaking behaviors. Power for these tests was .84, anticipating a moderate
effect size (Cohen, 1988). Significant variation between feedback groups was ob-
served for eye contact, F(2,171) ⫽ 6.46, p ⬍ .05, 2 ⫽ .07. Post-hoc Scheffe tests (.05
alpha) revealed that the no feedback group (M ⫽ 52.0, SD ⫽ 22.5) differed
significantly from the low-intensity (M ⫽ 64.17, SD ⫽ 18.1) and high-intensity
(M ⫽ 65.58, SD ⫽ 19.4) groups on the eye contact variable.
A significant feedback effect was also observed for a number of references,
F(2,171) ⫽ 6.56, p ⬍ .05, 2 ⫽ .071. Again, the no feedback group mean (M ⫽ 1.79,
SD ⫽ 1.47) was significantly lower than either the low-intensity (M ⫽ 3.07, SD ⫽ 2.5)
or high-intensity (M ⫽ 2.91, SD ⫽ 2.2) group means.
No effect for feedback was found for introduction length, F(2,1.9) ⫽ 1.03. Re-
sponding to the second hypothesis, a correlation matrix was computed for sensitivity
to feedback factors and production of preferred public speaking behaviors in the
low-feedback-intensity condition. Power for these tests was .75, with a projected
alpha level of .05 and an anticipated moderate effect size of .30 (Cohen, 1988). The
correlations appear in Table 1. A significant positive correlation between feedback
212 C. D. Smith & P. E. King
sensitivity (for both socially desirable and socially undesirable feedback) and pro-
duction of desired public speaking behaviors following low-intensity feedback was
hypothesized and largely supported in the analysis. For the sensitivity to socially
undesirable feedback factor, significant correlations were found between sensitivity
and introduction length (r(58) ⫽ .416, r2 ⫽ .173, p ⬍ .001), and between sensitivity
and eye contact, (r(58) ⫽ .534, r2 ⫽ .285, p ⬍ .001), but not between sensitivity and
number of references used (r(58) ⫽ .203). Likewise, for the sensitivity to socially
desirable feedback factor, significant correlations emerged for the same public
speaking variables, between sensitivity and introduction length (r(58) ⫽ .639,
r2 ⫽ .41, p ⬍ .001), and between sensitivity and eye contact (r ⫽ .534, r2 ⫽ .285,
p ⬍ .001), but not between sensitivity and number of references (r(58) ⫽ .039).
The third hypothesis projected significant negative correlations for feedback
sensitivity and speaking behaviors under the high-intensity feedback condition. This
hypothesis received less support. Power for these analyses was .72, anticipating a
moderate effect size. Results are likewise shown in Table 1. Most correlations were
nonsignificant. The exception was for the number of references used, which was
inversely correlated to the sensitivity to socially desirable feedback factor (r(51) ⫽ –
.301, r2 ⫽ .09, p ⬍ .05) following high-intensity feedback (Cohen, 1988).
Post-Hoc Analyses
While hypotheses two and three were phrased for convenience and ready apprehen-
sion, it can be argued that the absolute magnitude of the correlations between
feedback sensitivity and public speaking behaviors is less important than whether the
high-intensity and low-intensity feedback conditions differed in terms of that
association between feedback sensitivity and speaking performance. One approach to
testing this difference is to examine whether there were significant differences
between the sensitivity/behavior correlations under the high- versus the low-inten-
sity conditions.
Fisher’s zr transformation is an appropriate and conservative procedure for
examining the difference between correlations of independent samples (Ferguson,
1981). For the correlation between the student’s sensitivity to socially undesirable
feedback and the length of the student’s introduction, no significant difference
between high- and low-intensity feedback was detected (zr ⫽ 1.91). However, for
both the sensitivity to socially undesirable feedback/eye contact (zr ⫽ 2.08, p ⬍ .05)
correlations and the sensitivity to socially undesirable feedback/references (zr ⫽ 2.06,
p ⬍ .05) correlations, significant differences between feedback-intensity levels were
detected.
For correlations between sensitivity to socially desirable feedback levels and
performance variables, significant differences were found between the group exposed
to high-intensity feedback and the group exposed to low-intensity feedback in all
dependent variables—for the correlations between feedback and introduction length
(zr ⫽ 4.07, p ⬍ .01), between feedback and eye contact (zr ⫽ 2.59, p ⬍ .01), and
between feedback and number of references (zr ⫽ 2.74, p ⬍ .01).
Student Feedback Sensitivity 213
Discussion
The present study sheds important light on the dynamics of instructional feedback.
The first hypothesis demonstrated that the use of feedback led, overall, to observed
increases in the performance targets of that feedback (introduction length, eye
contact, and number of references cited). However, if the analysis ended here, large
gaps would remain in explicating the efficacy of instructional feedback. The remain-
ing hypotheses were aimed at investigating the interaction between student individ-
ual differences (feedback sensitivity level) and the kind of feedback (intensity level)
associated with those observed differences in targeted behavior.
The second hypothesis thus posited a positive linkage between feedback sensitivity
and production of desired public speaking behaviors in the low-intensity condition.
This was observed for two of the three public speaking outcomes, eye contact and
introduction length. This correlation indicates that speakers high in feedback
sensitivity were conforming to desired behaviors at a higher rate than speakers low
in feedback sensitivity. Though no significant correlation was found for feedback
sensitivity and number of references used, this variable may have been more difficult
to adjust, involving a return trip to the library, or may have been seen as less relevant
to personal communication competence.
A lone significant inverse correlation was found between feedback sensitivity and
number of references only partially supporting the third hypothesis. However, the
trends emerging in the present study between feedback sensitivity and production of
targeted behaviors are certainly supported by studies associating feedback sensitivity
with the processing and internalization of messages (Edwards, 1990; Kinney &
Segrin, 1998). Moreover, post hoc analyses in the present study demonstrated that
the magnitude of the relations between feedback sensitivity and targeted perform-
ance variables differed between high- and low-intensity experimental conditions in
all cases save one.
Thus, it would appear that individuals with a high sensitivity to feedback most
readily adapt their behavior when the feedback is of a lower intensity. Presumably,
highly negative feedback causes the feedback sensitive learner to make negative
attributions, to focus on meta-task issues, such as seeing the feedback as punishment,
and to fail to modify behavior.
Edwards and Pledger’s (1990) feedback sensitivity scale was adapted for use in this
study in an instructional context. Two of the subscales in the original instrument—
sensitivity to attention and anticipation of response—did not appear to be particu-
larly relevant to the kind of summative instructional feedback given to students in
this study, and therefore those two factors were not utilized. Both of those non-used
factors assume ongoing interactions to which response and attention are highly
variable. Some, but not all, instructional feedback shares those same features (i.e.,
informal formative feedback and teacher–student consultation are usually more
interactive than summative feedback associated with graded assignments). That kind
of informal, ongoing feedback might be the object of subsequent study and might
very well display a different pattern than what emerged here.
214 C. D. Smith & P. E. King
In this study, we predicted that public speaking outcomes would be associated
with both socially desirable and socially undesirable feedback sensitivity factors when
feedback was of a low intensity. Further, we predicted that targeted public speaking
behaviors would not increase, however, when feedback was of a high intensity. A
question might be raised concerning the operational nature of the two types of
feedback sensitivity. Specifically, why should individuals who are highly sensitive to
socially desirable feedback respond in cases, such as the current study, where
feedback is entirely critical? While the nature of these factors cannot be discerned by
the analyses in this study, it should be noted that the factors did not measure
whether the messages used as stimulus were socially desirable or undesirable, but
rather, measured the reactivity of participants. It is possible that, in the present
study, individuals high in sensitivity to socially desirable feedback were especially put
off by the egregious lack of such feedback, particularly in the high-intensity message.
In that sense, the study may have been revealing what was lacking in the feedback
messages (i.e., any positive feedback at all), as well as what was present.
Future research should attempt to further explicate the circumstances under
which feedback works, or fails. It is certainly possible that two levels (at a minimum)
of sensitivity are involved. The present study supports the existence of an “I take it
personally” dimension, but there should also be variability in an individual’s ability
to decode or sense a feedback standard gap. In fact, a feedback efficacy model has
recently been proposed which accounts for (1) decoding sensitivity, (2) sensitivity to
punishment that can activate internal defense mechanisms (psychological sensi-
tivity), and (3) modification of behavior (King & Young, 2002).
What emerges from this analysis is an interesting difference in the instructional
requirements of students. A more traditional view is that greater feedback intensity
(more feedback, stronger language, blunt criticism) leads to greater improvement.
This is most likely true only for students low in feedback sensitivity, or reactivity.
Students high in feedback sensitivity may quickly respond to moderate levels of
criticism yet may rapidly overreact to feedback of significant intensity. Ultimately,
feedback must walk the fine line of being obvious enough to get the student’s
attention yet not too harsh to provoke negative attributions and defensive behaviors.
Similarly, for the learner (and unlike interpersonal sensitivity), it would be best to
maintain a golden mean of feedback sensitivity, being neither too high nor too low.
It is likely that personal knowledge of the student involved is absolutely essential in
understanding the efficacy of instructional feedback.
Potential application for feedback sensitivity ranges beyond the instructional
context. In a recent study of feedback in business organizations, Geddes and
Linnehan (1996) observe that, “ … some recipients of negative feedback regarded
this information as appropriate and helpful, and adjusted their performance accord-
ingly; however, others considered it inappropriate, even hurtful, and, in a surprising
number of cases, retaliated with hostile/aggressive acts such as work slowdowns,
machinery sabotage, and project abandonment (p. 328). Differences in feedback
sensitivity and in the intensity of feedback messages may ultimately prove beneficial
in accounting for and avoiding such reactions, in the classroom and beyond.
Student Feedback Sensitivity 215
References
Balzer, W. K., Doherty, M. E., & O’Connor, R., Jr. (1989). Effects of cognitive feedback on
performance. Psychological Bulletin, 106, 410–433.
Bangert-Downs, R. L., Kulik, C. C., Kulik, J. A., & Morgan, M. T. (1991). The instructional effect
of feedback in test-like events. Review of Educational Research, 61(2), 213–238.
Baxter, L. A. (1984). An investigation of compliance-gaining as politeness. Human Communication
Research, 10, 427–456.
Booth-Butterfield, M. (1989). The interpretation of classroom performance feedback: An attribu-
tional approach. Communication Education, 38, 119–131.
Booth-Butterfield, M., & Booth-Butterfield, S. (1990). Conceptualizing affect as information in
communication production. Human Communication Research, 16, 451–476
Brandl, K. K. (1995). Strong and weak students’ preferences for error feedback options and
responses. The Modern Language Journal, 79, 194–211.
Cadinu, M. R., Arcuri, L., & Kodilja, R. (1993). Self-serving biases: The role of perspective-taking.
European Journal of Social Psychology, 23, 97–102.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Hillsdale, NJ: Erlbaum.
Daly, J. A., Vangelisti, A. L., & Daughton, S. M. (1987). The nature and correlates of conversa-
tional sensitivity. Human Communication Research. 14, 167–202.
Edwards, R. (1990). Sensitivity to feedback and the development of self. Communication Quarterly,
38, 101–111.
Edwards, R., & Pledger, L. (1990). Development and validation of the sensitivity to feedback scale.
Communication Research Reports, 7, 83–89.
Ferguson, G. A. (1981). Statistical analysis in psychology and education. New York: McGraw-Hill.
Frandsen, K. D., & Millis, M. A. (1993). On conceptual, theoretical and empirical treatments of
feedback in human communication: Fifteen years later. Communication Reports, 6, 79–91.
Garrard, G. G., & Kyriacou, C. (1985). Social sensitivity among young children. Journal of Social
and Personal Relationships, 2, 123–135.
Geddes, D., & Linnehan, F. (1996). Exploring the dimensionality of positive and negative
performance feedback. Communication Quarterly, 44, 326–344.
Green, T. D., Bailey, R. C., Zinser, O., & Williams, D. E. (1994). Causal attribution and affective
response as mediated by task performance and self-acceptance. Psychological Reports, 75,
1555–1562.
Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.
Jones, E. E. (1990). Interpersonal perception. New York: W. H. Freeman.
Jussim, L., Soffin, S., Brown, R., Ley, J., & Kohlhepp, K. (1992). Understanding reactions to
feedback by integrating ideas from symbolic interactionism and cognitive evaluation theory.
Journal of Personality and Social Psychology, 62, 402–421.
Kelley, H. H. (1971). Attribution in social interaction. In E. E. Jones et al. (Eds.), Attribution:
Perceiving the causes of behavior (pp. 1–26). Morristown, NJ: General Learning Press.
King, P. E. (2001). Automatic responses, target resistance, and the adaptation of compliance-seek-
ing requests. Communication Monographs, 68, 386–399.
King, P. E., & Behnke, R. R. (1999). Technology-based instructional feedback intervention.
Educational Technology, 39, 43–49.
King, P. E., & Young, M. J. (2002). An information processing perspective on the efficacy of
instructional feedback. American Communication Journal, 5. Retrieved October 24, 2002,
from http://www.acjournal.org/holdings/vol5/iss2/
King, P. E., Young, M. J., & Behnke, R. R. (2000). Public speaking performance improvement as
a function of information processing in immediate and delayed feedback interventions.
Communication Education, 49, 365–374.
Kinney, T., & Segrin, C. (1998). Cognitive moderators of negative reactions to verbal aggression.
Communication Studies, 49, 49–73.
216 C. D. Smith & P. E. King
Kluger, A. N., & DeNisi, A. (1996). The effects of feedback interventions on performance: A
historical review, a meta-analysis, and a preliminary feedback intervention theory. Psycho-
logical Bulletin, 119, 254–284.
MacGeorge, E. L. (2001). Support providers’ interaction goals: The influence of attributions and
emotions. Communication Monographs, 68, 72–97.
Snodgrass, S. E., Hecht, M. A., & Ploutz-Snyder, R. (1998). Interpersonal sensitivity: Expressivity
or perceptivity? Journal of Personality and Social Psychology, 74, 238–249.
Stacks, D. W., & Murphy, M. A. (1993). Conversational sensitivity: Further validation and
extension. Communication Reports, 6, 18–24.
Watt, S. E., & Martin, P. R. (1994). Effect of general self-efficacy expectancies on performance
attributions. Psychological Reports, 75, 951–961.