.—

 

.3 I
"'22. I:

. '9. £4 .3 '5}, ’33:;
\ . . . ‘9 3-0 "v . ‘5' "
.9: ”3.9.9.; 3 : .93 3 39.9.2: 3939.79.93"-

3..'9¢..,.w3‘ av.
133:“.

   
     
 
    
   
  

L-ln

\
.vf... ‘69
=999 3.9.; 99%.?”-
9?“ guy “:1. ‘ 4.9:
3‘":'wﬁ

“(1.:
g;

z'a—I'i‘L-‘b‘ﬁMTXJ ‘

Kb! Stu.
s-l‘ .999: ‘ﬁ 5!
£193. 9 .99.... . . - - r... 9999‘ ‘9‘
. . :1. I t .. ‘ 3 3:3; '4. m}: $¢%;q..hl .16?“
.3 . , Saguaxwfﬁﬁ .\
’\
N . «Ma». "59;.9

 

hi3"

 

 

 

“ 359. ..
~ “51%;!" :ngi: *

’v'y'
_ ‘3. at
£3'. ;3.£:9\9

J-
Y 3.‘
.“ 4;:“1‘tfﬁgn.
“. ' £1w1‘-.-’..;..'.. 9‘ 1‘“ ‘ “- " - I ‘"
-J"'iJ'v.". '9 “W“. “:xu .._.. 2"ng
9 , '- 'v . u '99
.

   

‘L .- I 9.,329-5253}.

u» v( .
w , “~39“ haun-

 

\‘r1.9-nir¢'.1":‘_~l. ' 13‘.W
~43 1.1:. ‘ $9.339
3 .

f'.‘ 1.339333“
.95: .

 

C»
‘1‘ «Fiﬁ?

}A L )3:
ﬁ'ﬁuﬁ. um...‘ "

3.2. fluff—i...
“I.”

 

61-" 9‘9

‘ if

.
.n.
”.991
5

. . ’ . ‘ '9' ‘1 '3 l q. ”2r"
. . . . 9 .. .u ,.
. - . . 9 3 . , . .3 31
.3 Hwy...»
. .
[“3“

lullllIllllllwlugltlllllulll p

3129

This is to certify that the
thesis entitled

THE EFFECTS OF INFORMATION CONSISTENCY, TIMING

OF CATEGORY SALIENCE, AND MOTIVATION ON
THE RECALL OF RATEE BEHAVIOR

presented by

Miguel Angel Quinones

has been accepted towards fulﬁllment
of the requirements for

M.A. degree in Psychology

   

 

Major prof%

Date 11/26/90

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

~—

LIBRARY
“lent“. State
University

“—‘I

—_

 

ﬂ

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

_—____—__————-—————_—_————_——————————T
DATE DUE DATE DUE DATE DUE

1 I
+_l|

 

 

 

 

 

 

 

 

 

 

 

 

J

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘ ' J

’ 4_J
MSU Io An Affirmative ActiorVEqual Opportunity Institution
cszNmea-M

 

THE EFFECTS OF INFORMATION CONSISTENCY, TIMING
OF CATEGORY SALIENCE, AND MOTIVATION ON
THE RECALL OF RATEE BEHAVIOR

By

Miguel Angel Quiﬁones

A THESIS

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

MASTER OF ARTS

Department of Psychology

1990

ABSTRACT
THE EFFECTS OF INFORMATION CONSISTENCY, TIMING
OF CATEGORY SALIENCE, AND MOTIVATION ON
THE RECALL OF RA'I'EE BEHAVIOR
By

Miguel Angel Quiﬁones

This study explored the effects of a performance category (i.e. good performer) on recall
in a performance appraisal situation. More speciﬁcally, the study investigated the effects
of providing a category prior to and after the observation of a ratee on the recall of ratee
behavior. The category used was that of good secretary. In addition, the effects of
motivation (accountability) as a way of eliminating the biasing effects of providing a
category were also investigated. Undergraduate psychology students (n=132) participating
in groups of 2-7 (mode=6) were presented with a bogus performance appraisal form
depicting a university secretary as being a good performer prior to or after viewing a
videotape of this secretary containing instances of good and poor performance. Motivation
was manipulated by telling half of the participants that they would have to justify their
responses on a behavioral checklist to the rest of the group members. The hypotheses
relating to accuracy of category consistent and inconsistent behaviors were not supported.
Subjects were more accurate in recalling instances of poor behaviors. The failure to ﬁnd

the predicted effects and directions for future research are discussed.

ACKNOWLEDGEMENTS

I guess the ﬁrst person I need to thank is myself for doing all the work and the typing
required for ﬁnishing this project. However, I would not be myself if it wasn’t for my
beautiful wife Sherry. She has helped me become the person I am today. Without her
love and constant support I would probably never have ﬁnished my thesis. My Chairperson,
Dan Ilgen, has been instrumental in making this a quality project. Thanks a lot Dan for
not letting me get away with mediocre work. My other committee members, Steve
Kozlowski and J. Kevin Ford, have also provided me with invaluable guidance and support
during my tenure here at Michigan State. I would also like to thank my colleagues,
Stephen Gilliland and Jeffrey Schneider, for showing me that this was not an impossible
task. They have provided excellent role models for me. Dawn Rosen deserves to be
mentioned for helping me run the experiments. Finally I would like to thank my parents.
The have always believed in me and my ability to succeed. This unconditional support has

helped me overcome every obstacle I have encountered in my life, including this thesis.

iii

TABLE OF CONTENTS

Page

LIST OF TABLES ........................................... vii
LIST OF FIGURES .......................................... viii
INTRODUCITON ........................................... 1
Filter Model ............................................. 5
Organizer Model .......................................... 8
Strength of Organizer Effects .............................. 13
Contradictions ......................................... 13
Conclusion ........................................... 15
Motivation .............................................. 15
Framework for Current Research .............................. 20
Criteria .............................................. 20
Accuracy .......................................... 20
Recognition vs Recall ................................. 22

Nature of Categories .................................... 23

Time of Activation of Categories ............................ 24
Motivation ........................................... 26

iv

Stimulus Materials .........................................
Procedure ..............................................
Dependent Measures .......................................

Measures Derived From Behavioral Checklist ...................

Other Measures ........................................

Descriptive Data ..........................................
Manipulation Checks .......................................
Analysis of Hypotheses .....................................
OverallTests. ............ ..............
Overall Accuracy .......................................
False Alarm Rates ......................................
DISCUSSION ..............................................
Summary of Results .......................................
Overall Tests ..........................................
Overall Accuracy .......................................
False Alarms ..........................................
Limitations of the Study .....................................
Future Research ..........................................

LIST OF REFERENCES ......................................

29

29

29

29

34

35

35

36

38

38

38

38

38

39

46

49

50

50

56

58

61

62

APPENDD( A: BARS Scales Used To Rate Behavioral Incidents ..........

APPENDDi B: Consent Form ...........
APPENDIX C: Instructions .............
APPENDD( D: Rating Form ............
APPENDD( E: Bogus Rating Form .......

APPENDD( F: Distractor Task ..........

7O

74

75

76

77

78

83

LIST OF TABLES

Page
TABLE 1: Experts Ratings of Behavioral Incidents Performed in the Videotape 32
TABLE 2: Description of Behavioral Incidents on Videotape .............. 33
TABLE 3: Means, Standard Deviations, and Intercorrelations of variables ..... 40
TABLE 4: Manipulation Check for Accountability ..................... 41
TABLE 5: ANOVA Summary Table of Consistency by Timing by Motivation . . 42
TABLE 6: Means of Consistency by Timing ......................... 43

TABLE 7: ANOVA Summary Table of Total Accuracy by Timing by Motivation 45

TABLE 8: ANOVA Summary Table of False Alarms by Timing by Motivation . 47
TABLE 9: Means of False Alarms by Timing ........................ 48
TABLE 10: ANOVA Summary Table of Consistency by Impression ......... 54

vii

LIST OF FIGURES

Page

FIGURE 1: Expected Relationship Between Consistency, Category Timing,
and Motivation ...................................... 28
FIGURE 2: BARS Scale for Dimension 1: Job Knowledge and Skill ........ 31
FIGURE 3: Interaction of Consistency by Timing ...................... 44
FIGURE 4: Interaction of Consistency by Impression ................... 55
FIGURE 5: False Alarms by Category ............................. 60

INTRODUCTION

Early research in performance appraisal focused on the rating instrument as a
possible source of bias (i.e. halo, leniency, central tendency). This approach, as well as
attempts to train raters to be more accurate, did little to reduce common rating errors
(Landy & Farr, 1980). Researchers have since begun to look at performance appraisal ’
as a complex and ongoing process in which the instrument and the rater are only a
subset of the factors that can bias ratings (DeNisi, Cafferty, & Meglino, 1984). This
more recent approach to performance appraisal has been labeled the process approach
(Ilgen & Feldman, 1983).

A process perspective views the ﬁnal rating as the result of an interaction
between the organizational context within which the appraisal takes place, the rater’s
information processing system, and the behaviors of the ratee (Ilgen & Feldman, 1983).
Inherent in this perspective is the recognition that a rater acquires information about a
ratee over a period of time from different sources (DeNisi et al., 1984; Ilgen &
Feldman, 1983). A primary goal of appraisal process research is to ﬁnd key factors
throughout the rating process that may bias the ﬁnal rating. One such factor that has
stimulated a number of research projects is the Tater’s information processing system.

It is suggested that raters process information about others who are to be

evaluated by (1) attending to behavioral information about the ratee, (2) categorizing

2
this information in memory, (3) recalling it at the appropriate time, and (4) integrating
the information for purposes of coming up with a ﬁnal rating (DeNisi et al., 1984;
Feldman, 1981). However, people are not perfect, unbiased perceivers of such
information. Instead they rely, in part, on simpliﬁcation heuristies that make
categorization of incoming information easier (Cantor & Mischel, 1979; Fiske & Taylor,
1984). To the extent that incoming information is similar to a readily accessible
category in memory, the information is more likely to be stored under that category
(Rosch, Mervis, Gray, Johnson, & Boyes-Braem, 1976). Given the fact that people rely
on categories to process information that they encounter, researchers have focused on
studying the nature and the effects of categories on the information processing system
(e.g. Cohen, 1981; Nathan & Lord, 1983).

Feldman (1981) noted that there are many situational factors that inﬂuence the
salience of a particular category at any given time. Furthermore, research has shown
that the use of categories inﬂuences the type of information to which people attend and
recall (DeNisi et al., 1984; Srull & Wyer, 1989). Thus, if the performance appraisal
process is ongoing and categories are made salient by a number of situational features, it
is reasonable to conclude that these categories could be activated at any point in the
appraisal process. However, most process models of performance appraisal view
categories as inﬂuencing the type of information that is sought out and recalled after
the category has been made salient (DeNisi et al., 1984). This condition will be referred
. to as the Filter Model.

In some rating situations, raters learn something about a ratee after he or she
has observed that ratee for a period of time. That is to say, the rater is not presented

with category information prior to the observation of performance. Rather, the

3
information about categories either evolves from the observations or is presented at
some point after observations begin. Research has shown that under these conditions
raters tend either to distort information previously encoded or to recall events that did
not even occur (Srull & Wyer, 1989, Srull & Wyer, 1980). Other researchers have also
noted differences in the kind of information that is recalled under conditions where
category information follows exposure to relevant stimuli (Wyer, Bodenhausen, & Srull,
1984; Wyer & Srull, 1989). Information processing models describe the process involved
when category information is received after observation and storage as one in which the
activation of a category does not act as a ﬁlter of incoming information but rather as an
organizer of information already available in memory (Wyer & Gordon, 1984; Wyer &
Srull, 1989). This condition will be referred to as the Organizer Model.

In addition to the rater’s information processing system, it has also been
suggested that the rater’s motivation and ability to provide accurate ratings will inﬂuence
the performance appraisal process (DeCotiis & Petit, 1978). Ability and motivation are
believed to interact to affect rating accuracy. Raters who are highly motivated to give
accurate ratings may not be able to do so if they do not have the ability to do so and
vice versa. Ability to give accurate ratings could be inﬂuenced by the fact that raters
may have, in memory, biased information about ratees as a result of selective attention,
encoding, or recall of information due to the presence of a category prior to or
following observation. Thus, similar amounts of motivation could lead to differential
accuracy in ratings due to the fact that biased information is available to raters as a
result of categories being made accessible at different points in the performance

appraisal process thus inﬂuencing the ability of the raters to give accurate ratings.

4

For example, a rater who believes that a particular ratee is usually a good
performer may only pay attention to information that conﬁrms this expectation (i.e. ﬁlter
model). When the time comes to rate this particular individual, no amount of
motivation will enable the rater to recall information that was never encoded in the ﬁrst
place such as instances of poor performance. However, if this rater learned about a
ratee’s past performance after he or she has had a chance to observe the ratee for a
period of time (i.e. organizer model), the rater might bias the information that is
recalled even though they have attended to and encoded all behavioral instances equally.
In this case, high motivation at least has the possibility of overcoming the biasing effect
of this performance category on the recall of the information since it is assumed that
the information was encoded in an unbiased fashion.

This study will look at the effects of making a category salient prior to the
presentation of behavioral information on the recall of this information and compare
that effect to that observed when an identical category is made salient immediately after
the presentation of behavioral information (i.e. will compare the effects of the ﬁlter
model versus those of the organizer model). Furthermore, the inﬂuence of the type of
information recalled on the ability of the rater to give accurate ratings will be explored
under varying levels of rater motivation to accurately recall the observed behavioral
information. It is expected that under similar motivational conditions, the rater’s ability
to give accurate ratings will be inﬂuenced by the time (before or after observation) at
which a category was made salient.

A review of the literature concerning the effects of making a category accessible
prior to and following observation will ﬁrst be presented. Second, the results presented

will be interpreted with respect to the speciﬁc problem of performance appraisals.

5
Third, a discussion of the effects of motivation and ability to give accurate ratings will
be presented. Finally a framework for the current research will be presented along with

a discussion of the speciﬁc hypotheses.

Filter Model

Many researchers have been interested in studying the effects of providing a
category prior to observing a target person on information processing. The primary way
that this effect has been studied has been to determine whether behaviors that are
consistent or inconsistent with the category are better remembered (see Srull & Wyer,
1989). This situation has sometimes been referred to as impression testing because
individuals use the behavioral information that is being presented about a target person
to test the category or impression that they already possess about the person.

As has already been stated, impression testing occurs when a judgement about a
person has already been made or an impression has already been formed and new
information is presented about that person. Preexisting judgments or impressions of
people are believed to serve as anchors or standards for evaluating new information
about a focal person by categorizing them as belonging to a particular group of people
(i.e. poor performer); these anchors are sometimes called stereotypes or prototypes
(Cantor & Mischel, 1979; Fiske & Taylor, 1984). Research has shown that such
categories (stereotypes/ prototypes) affect subsequent processing of information about
the focal person.

In a study by O’Sullivan and Durso (1984) people were given category labels

about a person along with two pieces of information highly congruent with the category.

6

The participants in the research were then presented with biographical information that
varied in its congruency with the initial stereotype label. The results showed that
participants had higher recall for category consistent items. Other researchers have
concluded that the effect of providing a category prior to the observation of a person is
to increase recall for category consistent behaviors (Higgins & Bargh, 1987).

Other researchers have studied the ﬁlter effects but found different results.
Hastie (1980) presented participants with a category about a focal person. The
participants were then asked to read a series of descriptive statements about that person
while forming an impression of him or her. The statements contained information that
was both consistent and inconsistent with the category that was originally presented.
The results indicated that participants had higher accuracy of recall for category
inconsistent behaviors. The authors concluded that this effect was due to participants
forming more associative bonds between the inconsistent items because they could not
be associated with the general category.

To explain the inconsistencies between Hastie’s (1980) data and the data of
those who found consistent information better recalled, Higgins and Bargh (1987)
pointed out that studies such as Hastie’s (1980) ask the participants to form their own
impression regardless of the category given to them prior to observation. This
manipulation puts the subjects in a mode of forming an impression along with that of
testing one. Support for the position that impression formation is a separate process
comes from studies which ask subjects to either form an impression or remember as
much as they can about a target person (Lichtenstien & Srull, 1987; Srull, 1981; Wyer &
Gordon, 1982). These studies ﬁnd that when subjects are asked to remember as much

as they can, the impression formation effect disappears. Higgins and Bargh (1987)

7

concluded that participants in impression formation experiments are not relying on the
original category presented to them prior to observation. Contrastingly different
processes than these are at work in an impression testing paradigm which is the reason
why they obtain different results.

Returning to the impression testing situation, there are several mechanisms that
have been proposed to account for the ﬁlter effect of impression testing in which a
recall advantage for consistent information is found. One proposed mechanism is that
of discounting (Higgins & Bargh, 1987). Discounting occurs when people have a
category and interpret inconsistent information in a way that is consistent with the
category (Bodenhausen, 1988). For example individuals may attribute the inconsistent
behaviors as being caused by factors outside of the person and thus not view these
behaviors as being inconsistent with the category assigned to the person (Crocker,
Hannah, & Weber, 1983). Thus the category acts as a ﬁlter which blocks out
inconsistent behaviors and does not allow them to be processed into memory. Another
possible mechanism which causes the ﬁlter effect is the way an individual "chunks” the
behavioral information being presented to them into discrete behavioral events.
Individuals may group information in a way that is consistent with their overall
impression or category. This conclusion is evidenced by the lack of congruence in
results found for studies using different modes of presenting the behavioral information.

All of the research that has been reviewed presents a rater with an extended
amount of information about the person who is to be rated, evaluated, or simply
described. There is some evidence that the mode through which the detailed
information about the stimulus person is presented has an effect on the results of recall

experiments. In particular, when the richer sources of information such as videotapes

8
are used and compared to less rich sources (such as paper and pencil descriptions of the
same information) consistent information is remembered better (Cohen, 1981; Rothbart,
Evans, & Fulero, 1979). Cohen (1981) argued that richer sources produce these effects
because participants must reduce the cognitive load created by the variety of
information. According to Cohen, they do this by chunking the behavioral episodes in a
manner that favors consistent information.

In conclusion, it appears that providing a category prior to observation serves to
affect recall by increasing accuracy for category consistent items. Several mechanisms
for this effect have been proposed all of which are presumed to work by affecting the
initial encoding of the information in the form of a ﬁlter. The effect has been most
acute when rich formats such as a videotapes are used which allow the participants to

reduce cognitive load by using the existing category.

Organizer Model

Up to now, the effects of a category have been discussed in terms of its effect
on information gathering. However, categories can be made salient ﬁg the
observation of information. For example, after working with an individual for a period
of time, one might ﬁnd out that a colleague was ﬁred from his or her previous job for
stealing. One might begin to recall instances in which the colleague acted in a
suspicious manner. Thus, activating a category after observation affects the way an
individual recalls previome stored information.

The effects of providing a categorization schema after the observation of

behavior have been less conclusive than those based on categories provided prior to

9

observation. However, similar to the ﬁlter effects, the organizer effects have been
studied by assessing their impact on recall of category consistent and inconsistent
behaviors. Several studies have found consistent information better recalled, others
inconsistent information, while still others have found no difference in recall advantage
for either form of information. All studies have investigated these hypotheses by
presenting participants with information about the ratee before the introduction of a
categorization schema (Cohen, 1981; Pyszczynski, LaPrelle, & Greenberg, 1987; Snyder
& Uranowitz, 1978). Some researchers have not explicitly stated that they were testing
the organizer effects but also introduced the category after observation (Phillips & Lord,
1982). The latter research has produced results similar to the explicit tests (Phillips &
Lord, 1982). There are, however, other studies which have failed to ﬁnd the predicted
effect for the organizer model (Belleza & Bower, 1981; Bodenhausen, 1988; Clark &
Woll, 1981; Rothbart et al., 1979; Wyer, Srull, Gordon, & Hartwick, 1982). Before
attempting to resolve the inconsistencies in results of the organizer effect studies, a brief
summary of the most important ones will be presented. It should also be added that
most studies use the organizer model as a control condition in studying the ﬁlter model.
This is done in order to rule out the possibility that the category in the ﬁlter model
serves as a recall cue. Thus if, a recall advantage for either consistent or inconsistent
information is found in the ﬁlter model and not in the organizer model then the
hypothesis that category effects are due to recall cues can be ruled out (e.g. Rothbart et
al., 1979).

In a study by Rothbart et al. (1979), subjects were presented with 50 behaviors
attributed to a ﬁctitious person. The behaviors were coded as friendly (n=17),

intelligent (n=17), unfriendly (n=3), nonintelligent (n=3), unrelated (n=10). Their

10
design was a 2 x 2 x 2 x 2 factorial design with 2 between-subjects and 2 within-subjects
factors. The between-subjects factors were the type of behaviors the subjects were led
to expect of the ﬁctitious person (friendly vs. intelligent) and the time at which they
received the categories of friendly or intelligent (before vs. after observing the
behaviors). The within-subjects factors were type of behavior (friendly behaviors vs.
intelligent behaviors) and type of instance (expectancy conﬁrming vs. expectancy
disconﬁrming). The behaviors were presented in the form of written descriptions on
randomly ordered slides. The authors expected to ﬁnd a difference in recall of
conﬁrming vs disconﬁrming behaviors as a result of providing the category after
observation. Results showed no advantage in recall for expectancy inconsistent
behaviors as was expected in the organizer model condition; the organizer model was
not supported.

In another study, Bodenhausen (1988) examined the effects of a category label
on recall and guilt judgments of a mock jury when presented with information about a
defendant during a mock trial. In this study, the author was also testing the hypothesis
of biased processing observed in the ﬁlter model and incorporated the organizer model
in order to rule out the possibility of the organizer effect acting only as a recall cue.
Results showed no difference in recall for category consistent or inconsistent behaviors
in the organizer model condition.

Cohen (1981) presented subjects with behavioral information on videotape. The
subjects in the study were led to believe that a woman depicted in the tape was either a
librarian or a waitress. This information was provided to half of the participants prior to
viewing the tape and to the other half after they had viewed the tape. Results showed

a signiﬁcantly higher accuracy in recall for consistent information in both conditions.

11

The author concluded that both processing as well as weak recall effect were responsible
for the differences in accuracy observed in the results. Since no difference was found
when the category label of h'brarian or waitress was presented before or after
observation, participants were not only relying on this category to selectively process
incoming information but also to aid them in recalling observed behaviors.

Similarly, in a study by Pyszczynski, LaPrelle, and Greenberg (1987), subjects
were presented with behavioral information about a person either before or after they
were allowed to read a general characterization (category) of that person. The
behavioral incidents were presented in the form of a recorded audio message. The
results showed a higher recall for category-congruent items in the condition where the
category was introduced after the behavioral incidents were presented. There was no
effect for the ﬁlter model. In their discussion, the authors hypothesized that the
reason for an organizer effect was that the audio messages made it difﬁcult for
participants to form associations among behavioral incidents. Since the incidents were
encoded independently, the category served as a retrieval cue for category-consistent
items. This is in contrast to studies which use a written format. The written format
allows subjects to memorize the items as a list. It was suggested that when participants
are asked to recall the behaviors, they were merely recalling a memorized list and did
not need to use of the category (Pyszczynski et al., 1987). This is a possible explanation
as to why the failure of Rothbart et a1. (1979) to ﬁnd an organizer effect.

In a study by Phillips and Lord (1982), subjects were presented with behavioral
information of a four-person problem-solving group on a videotape. After watching the
tape, they were presented with a bogus performance category about the group leader

(i.e. good leader) and were then asked to ﬁll out a questionnaire containing effective

12
and ineffective leadership behaviors. The subjects were presented with a list of
behaviors and asked to mark off those behaviors they recalled observing the leader of
the group perform. A higher recall for consistent items was found. That is, subjects
who were told that the group leader had performed well recalled more behaviors
associated with good leadership than poor. The results also indicated a higher false
alarm rate for consistent behaviors. This means that subjects reported having observed
more consistent than inconsistent behaviors that were not present in the videotape.

The Phillips and Lord (1982) study suggests that the organizer effect occurs
because participants use the category provided to them after observation and use it to
interpret the information previously observed. In recalling this information, they ”ﬁll in”
behavioral instances that are consistent with this category but were never observed.
This is the reason why Phillips and Lord (1982) found a large false alarm rate for
category consistent behaviors. This implies that one must take into account false alarm
rates when developing an accuracy estimate due to people’s tendency to distort their
recall toward consistent items. When the distortion was taken into account in this study
by correcting for false positives, inconsistent behaviors were more accurately recalled
than consistent. Since other studies which found consistent information more accurately
recalled than inconsistent did not correct for false alarms, their results (after correction)
may be consistent with Phillips & Lord (1982). This possibility will be examined in the
section entitled ”Contradictions”.

In conclusion, there are studies that have failed to ﬁnd an organizer effect (eg.
Bodenhausen, 1988) while other studies have found an organizer effect (e.g. Phillips &
Lord, 1982). Of these, some have found higher accuracy for items consistent with the

category given after observation (cg. Pyszczynski et al., 1987). Others have found higher

13
accuracy for category inconsistent behaviors (eg. Phillips & Lord, 1982). Thus there
appears to be evidence for an organizer effect but whether it acts by increasing accuracy
for category consistent or inconsistent behaviors is not clear. A possible explanation for
this is the correction for false alarms. There are two issues that need to be resolved.
One issue is why studies found differences in strength of organizer effect. A second
issue is why some found higher accuracy in recall for consistent information whereas
others found exactly the opposite effect. Each of these issues will be dealt with

separately.

Strength of Organizer Effects

The studies that failed to ﬁnd a difference in recall accuracy when a category is
presented after observation were those that used a written format rather than a video or
audio tape to present the behavioral information. This ﬁnding is consistent with
Cohen’s (1981) idea that a richer context is necessary for people to begin using
simplifying mechanisms such as processing biases and retrieval cues to deal with
incoming information. Thus it appears that there is an effect associated with providing
subjects with a category after the presentation of behavioral information but this effect
is present only when a richer more realistic format is used to present the behavioral

information.

Contradictions
The second issue deals with the seemingly contradictory results which have been
obtained. Some studies found higher accuracy in recall for category congruent items

while others found higher accuracy in recall for category incongruent items. A closer

14
look at the dependent measures used may be helpful in resolving this inconsistency. In
the Cohen (1981) study which found higher recall for category-consistent information,
accuracy was measured using a two-alternative forced-choice instrument which contained
only items that were present in the videotape. Even though the authors accounted for
errors that were made by the participants, these errors consisted only of participants
choosing items that did not go with the appropriate occupational category. The
responses coded as errors did not include items which were stereotypical but which did
not occur on the videotape. This is critical because Phillips and Lord (1982) have
suggested that people are more likely to choose category-consistent behaviors and thus
Cohen’s results may be a reﬂection of guessing on the part of the subjects. Correcting
Cohen’s data for false positives may have cleared up the discrepancy between the
studies.

In the Pyszczynski et a1. (1987) study, accuracy was measured by counting the
number of items that the subject wrote down which actually occurred in the audio tape.
However, there is no mention in the results of taking into account the number of items
written down which were not presented in the audio tape. Given the fact that accuracy
is measured in different ways, it is difﬁcult to directly compare the results of these
studies as they relate to accuracy in recall. However, as some studies suggest, false
alarm rates are an important issue in the study of processing biases. Thus the results of
the Phillips and Lord (1982) appear to be the most appropriate because their measure
accuracy takes into account the biases which can occur in the context of categorization
effects. Therefore it seems that in the situation in which a category is made salient

after the presentation of behavioral information, it is inconsistent information that is

15
more likely to be accurately recalled when the measure of accuracy corrects for false

alarms.

In conclusion, two issues need to be resolved in order to test the organizer
model. First, the stimulus material should be as realistic as possible. This would include
video or audio taped behavioral incidents. Second, the results should be compared with
a similar condition in which the categorical information is presented before observation,
and both measures must take false alarm rates into account. This would result in a
direct comparison of the pre and post situation which controls for the effects of
categorization (namely false alarms).

The information processing effects presented above assume that people have
equal motivation to recall the information presented to them. However, it is reasonable
to assume that not everyone is equally motivated to recall information and that these
differences can inﬂuence the way people process information. This is an issue that will

be dealt with in the following section.

Motivation
It has been recognized that the rater’s motivation to give accurate ratings will
affect the ﬁnal ratings (DeCotiis & Petit, 1978). It has also been recognized that a
rater’s ability will limit the accuracy of ratings regardless of the level of motivation. As

Taft (1971) has pointed out:

16

if the judge is motivated to make accurate judgments

about his subject and if he feels himself free to be

objective, then he has a good chance of achieving his aim,

provided of course that he has the requisite ability and can

use the appropriate judgmental norms (p. 177).
As has been noted earlier, a person’s ability to accurately recall information and hence,
give accurate ratings, is affected by the processes used to encode the information. If
selective processing took place (i.e. the ﬁlter model) then it is predicted that biased
information will be present in memory. In this case, high motivation might only serve to
increase the number of category consistent items recalled but should have no effect on
inconsistent information because such information should not be properly encoded.
However, in the situation where categories serve as retrieval cues (i.e. the organizer
model), motivation should affect recall of both category consistent and inconsistent
behaviors due to the fact that all behaviors should have been encoded equally during
observation.

Thus it appears that motivation will have a differential impact on the recall of
consistent and inconsistent information under the organizer condition but not on the
ﬁlter condition. In the latter case, there will be no differential effect although recall
should be elevated for both forms of information when low motivation is compared to
high.

Motivation can be manipulated in a number of ways. Some studies have used
monetary incentives. However, this manipulation assumes that all subjects will value a
given amount of money similarly. In studies dealing with decision making, motivation

has been successfully manipulated by varying accountability for the decisions (T etlock,

1985). In the present study, deciding whether or not a behavior was observed requires

17
the participants to make a choice among several alternatives. Therefore, it was decided
to manipulate motivation to be accurate by holding participants accountable to their
decisions. As justiﬁcation for the use of such a motivation manipulation, a review of the
effects of using accountability to manipulate motivation in decision tasks follows.

The effects of accountability on decision making behavior has been studied by a
number of researchers (see Tetlock, 1985). Accountability typically refers to the extent
to which the decision maker’s performance is open for review by others. Motivationally,
it is suggested that the presence of accountability increases the probability of evaluation
apprehension (Marcy, 1990). Accountability typically reduces the level of tolerance for
mistakes and thus increases the demands for a correct decision (Beach & Mitchell,
1978).

Most studies have explored the effects of accountability on decision making
strategy. More speciﬁcally, studies have focused on the type of information used by
decision makers to make their decisions under conditions varying in accountability.
Rozelle and Baxter (1981) manipulated accountability by telling subjects that they would
be required to attend a later session to discuss their ratings with other students. The
task was to rate students applying to graduate school by watching a videotape of the
candidate during an interview. The results showed that high accountability subjects’
descriptions of the candidate more closely matched the behaviors observed on the
videotape than low accountability subjects. This increase in accuracy reﬂects an increase
in motivation towards giving a "correct" answer.

In a study by Tetlock and Kim (1987), accountability was manipulated before and
after participants were allowed to read responses by a test taker to 16 items from

Jackson’s Personality Research Form. The task was to predict the responses that the

18
same test taker would make to another set of questions. Accountability was
manipulated by telling participants that they would be taking part in an interview
describing the information used to base their predictions. Results showed that in the
pre-exposure accountability condition, subjects reported more integratively complex
impressions of test takers, made more accurate behavioral predictions, and reported
more appropriate levels of conﬁdence in their ratings than no accountability and
post-exposure accountability subjects. The authors concluded that the accountability
manipulation inﬂuenced the way in which subjects encoded and processed information
resulting in a more complex integration of the information. The authors believed that
this was because when participants are accountable, and know this prior to information
presentation, they actively prepare for counter arguments to their expressed opinions. It
can be argued that this active preparation for counter arguments reﬂects an increase in
motivation toward producing correct responses.

McAllister, Mitchell & Beach (1979), studied the effects of accountability on
decision strategy choice and anticipated time for performing a task. The task included
rating the marketability of two groups of products and deciding which product had the
best market potential. The results indicated that accountable subjects chose more
cognitively complex strategies, anticipated needing more time to complete the task, and
took more time in completing the actual task; all these were considered to be a
reﬂection of higher motivation.

Researchers have concluded that when subjects expect to justify their decisions
to others, they use more cognitively complex strategies which rely more in the
information available to them than on information inferred through categorization

(T etlock, 1987). Furthermore, accountability seems to increase cognitive effort as

19
evidenced by the types of decision strategies employed by subjects in an accountability
condition (Tetlock, 1983).

While accountability does not seem to have an effect on processing biases when
introduced after the presentation of information, it does seem to increase the use of
information that is available to the decision maker. Thus, in situations where
participants are asked to recall information that they have observed at a previous time
and are also informed that they will have to justify their responses to a group, one can
expect that they will give a more accurate representation of the information that is
available to them in memory. Furthermore, if a category is introduced after the
observation of a target person, accountability should work to reduce the subjects’
reliance on this category for making decisions and thus give a more accurate picture of
the information that they posses in memory. However, if accountability is not present,
subjects should be more likely to rely on an available category which requires less
cognitive effort on their part.

Although motivation to perform decision tasks well has been assumed to increase
with accountability, a more direct measure of motivation is required if one is to
conclude that accountability has the effect of making participants think more deeply
about the information that they posses in memory. One possible measure of motivation
is the time that it takes for subjects to complete the task for which they will be held
accountable. It is hypothesized that accountable subjects will take more time in
completing the task due to the fact that they will have to come up with counter
arguments to any criticism leveled against their decisions (McAllister eta], 1979). It is
also possible that accountable subjects will want to review the information that was

presented to them more often than non-accountable subjects. Thus, if subjects are

20
asked whether they would like to review the information presented to them, accountable
subjects should be more likely to say yes than non-accountable subjects. Accountable
subjects may also be expected to rate themselves higher than non-accountable subjects in

measures of effort and motivation toward the task.

Framework for the Current Research

The current research is based on the process view of performance which suggests
that the rater’s information processing system will have an impact on the accuracy of the
ﬁnal ratings (Feldman, 1981; Ilgen & Feldman, 1983). More speciﬁcally, the current
research takes the approach of Feldman (1981) that categories and the time of their
activation will affect the way information is encoded and retrieved. This study does not
look at rating accuracy in terms of a rating scale but looks at accurate recall of
behavioral events. This is based on the assumption that a necessary condition of being
able to rate accurately is being able to accurately recall the events that lead to that

rating.

993%

Accurag. Since the present study deals with the effects of categories on
processing of information, the literature reviewed suggested that a measure of accuracy
that takes into account the systematic distortions produced by the presence of categories
must be used. Lord (1985) argued that, when dealing with recall accuracy in these
situations, one must take into account hit rates (correctly identiﬁed behaviors which

were observed) as well as false alarm rates (behaviors which are reported to have been

21
observed but which were actually not exhibited). This can be done by using a
questionnaire that contains an equal number of behavioral descriptions that are both
exhibited and not exhibited in the stimulus materials on the ratee. Raters can then be
asked to indicate whether each behavior on the list was or was not presented in the
observed segment of behaviors, and, from these responses, accuracy indices can be
constructed. First, the proportion of correctly and incorrectly identiﬁed behaviors can
be calculated for both category consistent and category inconsistent behaviors using the

following formulas:

Proportion of consistent ncc
behaviors correctly = Pcc= --
Identiﬁed tCc
Proportion of inconsistent n":
behaviors correctly = Pr c= --
Identiﬁed t

where n c is the total number of consistent behaviors correctly identiﬁed, nlc is the total

c
number of inconsistent behaviors correctly identiﬁed, tIc and th are the total number of
consistent and inconsistent behaviors actually exhibited by the ratee. In the present
study tic will be set equal to tee in order avoid confounding consistency with set-size

(Higgins & Bargh, 1987). Incorrectly identiﬁed behaviors can be calculated with the

following formulas:

Proportion of consistent nci
behaviors incorrectly = Pei = --
identified tCi
Proportion of inconsistent nli
behaviors incorrectly = P r i = --

identiﬁed tl i

22
where nci is the total number of consistent behaviors incorrectly identiﬁed, nxi is the
total number of inconsistent behaviors incorrectly identiﬁed, tcr and t1i are the total
number of consistent and inconsistent behaviors on the questionnaire which were not
exhibited by the ratee. Similarly to the correct behaviors, tci will be set equal to t,i in
order to eliminate the set-size confound.

Finally, accuracy can be calculated for total, consistent, and inconsistent behaviors
by subtracting the prOportion of correctly identiﬁed behaviors from the proportion of

incorrectly identiﬁed behaviors for each of these in the following manner:

Pee-PC i

Prc'Pri

Ac”:

Accuracy for Consistent Behaviors

ll
.3’ 3':

Accuracy for Inconsistent Behaviors

Accuracy for Total behaviors

II
.P’
u

Recognition vs. Recall. Memory is most frequently indexed by two forms of
measures, recognition and recall. The present study uses a recognition measure of
memory. The participants were presented with a list of items and required to mark the
ones they remember as occurring in the stimulus material. It is called recognition
because participants only have to recognize the items and not remember them. Recall,
on the other hand, refers to measures that provide no cues as to what occurred; the
participant is asked to write down every item which they remember occurring in the
stimulus material.

Many studies have been carried out as an attempt to understand the similarities

and differences between responses to recognition and recall measures (see Gillund &

23
Shiffrin, 1984 for a review). Researchers agree that both forms of memory tasks require
the individual to utilize the information that they possess in memory. However, there is
some disagreement as to whether or not a recognition task requires a complete search
or merely the use of more familiar and easily accessed memory (Gillund & Shiffrin,
1984). This is important because if subjects rely on the familiarity of the object to be
recognized, then their judgments will be biased toward those that are consistent with
their impression (Wyer, Bodenhausen, & Srull, 1984). However, studies of these
differences have been inconclusive and researchers have concluded that in general, both
types of memory tasks require some search of the contents of memory (Gillund &
Shiffrin, 1984).

This study uses recognition because a signal detection theory form of accuracy
was selected to be consistent with the advice of Lord (1985). When one presents
information to individuals and then ask them to make memory judgments, the number of
correct items that they can remember is known by the researcher. However, when false
alarms are assessed, it is impossible to know how many behaviors an individual is
considering unless the total set of stimuli to be considered is controlled. With the
recognition format, the total number of incidents, both consistent and inconsistent, is

ﬁxed by the list presented to the participants therefore ratios can be constructed.

W

The current research deals with performance appraisal situations and the effects
of performance categories on recall of ratee behaviors. Performance categories are
prototypes or schemas about a person’s general level of performance on their job.

Speciﬁcally, in this study raters were informed that the person they watched (or were to

24
watch) was either a good or a poor performer. DeNisi et a1. (1984) have noted that job
performance categories can be constructed to be salient to raters by studying past
performance appraisals and adopting information from the appraisal forms. Thus,
performance category in the present research was operationalized as the rating (i.e. good

or bad) supposedly given to the target ratee (secretary) by another group of students.

Time of Activation of Categories

The time at which a category is made salient has been shown to inﬂuence the
type of information that is recalled (e.g. Pyszczynski et al., 1987). More speciﬁcally, it
has been found that in the ﬁlter model category consistent items are more accurately
recalled while in the organizer model it is category inconsistent behaviors that are more
accurately recalled. This variable was operationalized as the time at which the subjects
got the see the bogus performance appraisal ﬁlled out by another group of students
portraying the ratee (secretary) as a good worker. From the review of the literature
presented above the following hypotheses were made regarding the time at which a

category is made salient:

Hypothesis 1a: Subjects for whom a category has been made salient prior
to the observation of performance will more accurately recognize
category-consistent behaviors than category-inconsistent behaviors (i.e.

Ac>A,).

Hypothesis 1b: Subjects for whom a category has been made salient after
the observation of performance will more accurately recognize
category-inconsistent behaviors than category-consistent behaviors (i.e.

AI >Ac).

25
Another way to look at the combination of Hypotheses 1a and 1b is to say that an
interaction is predicted between consistency and timing of activation of the performance
category.

Cohen (1981) found that having a category present while observing behaviors
helped subjects remember more total behaviors than not having one present. This result
is consistent with the general conceptualization of categories as serving to organize
incoming information and reduce cognitive chaos (Cohen, 1981). This is also consistent
with the concept of advanced organizers studied in learning and training (Ausubel, 1968;
Howell & Cooke, 1989). This literature has found that providing individuals with an
organizing schema at a higher level of abstraction than that of the material that is to be
learned aids in the recall of the material (Mayer & Bromage, 1980). Thus, the following

hypothesis were made regarding total number of behaviors recognized:

Hypothesis 2: Subjects for whom a category has been made salient prior

to observation will more accurately recognize a greater number of

behaviors than those for whom a category has been made salient after

observation when the accuracy measure removes categorically consistent

and inconsistent behaviors that are falsely reported as having been

observed (i.e. Pre A1 > Post A1).
Phillips and Lord (1982) found that introducing a category after observation resulted in
a higher false alarm rate for category-consistent behaviors (i.e. organizer effect). The
subjects were more likely to report remembering category-consistent behaviors that did

not take place. Therefore, the following hypothesis were made regarding false alarm

rates in recognition:

26

Hypothesis 3: Subjects for whom a category has been made salient after
observation will have a higher false alarm rate for category-consistent
behaviors than for category-inconsistent behaviors (i.e. PCi>Pli)'

Motivation

In the present research, motivation was operationalized as the absence or
presence of accountability which has been shown to inﬂuence motivation. This was
manipulated by having one group of subjects believe that they would have to justify
their responses in the checklist to the rest of the group members (accountability). The
other group was not given this instruction. In order to test the motivational effect of
the accountability manipulation, the time that each individual took in ﬁlling out the
checklist was recorded. Thus if an individual took more time, it can be inferred that
they were more motivated and put more effort into the task. The following general

hypothesis was made regarding the effect of motivation on recognition:

Hypothesis 4: Subjects in the motivation condition will more accurately

recognize more total behaviors than those in the no motivation condition

(i.e. Mot. A1 > No Mot. A1).
As discussed earlier, subjects presented with a category prior to observation (ﬁlter
model) will engage in selective processing and thus are not able to recall as many
category-inconsistent behaviors due to failure to attend to these behaviors. In contrast,
subjects for which a category is made salient after observation (organizer model) should

have most of the information encoded in a similar fashion therefore a motivation

manipulation should have a greater impact. Therefore the following hypotheses were

27
made regarding motivation, timing of the introduction of a category, and

consistency/inconsistency of behaviors observed:

Hypothesis 5: There will be a three way interaction between consistency,
timing of category, and motivation. More speciﬁcally, of the subjects who
receive the category information after observation, those who are not
motivated will show a greater recall advantage over category inconsistent
behaviors than those who are motivated (i.e. for motivated subjects
A,=Ac, for not motivated subjects A,>Ac).

Of the subjects who receive the category information prior to observation,
motivation should have no effect on the recall advantage of category consistent
behaviors(i.e. Ac>A, for both motivation groups). This interaction is depicted in
Figure 1.

28

Figure 1
Expected Relationship Between Consistency, Category Timing,

and Motivation

NO MOTIVATION

 

 

 

 

 

MOTIVATION

 

 

 

 

 

METHOD

Subjects

Subjects were 132 undergraduate psychology students who received course credit
for their participation in the study. A power analysis revealed that this number of
subjects provided enough statistical power (.80) to detect a "medium” effect size with the
design used (Cohen, 1988). Subjects participated in groups ranging from 2-7 individuals
(mode=6). Each group was randomly assigned to one of four experimental group

conditions.

Design

The current study was a 2 X 2 X 2 mixed design with 2 between-subjects factors
and 1 within-subjects factor. The between-subjects factors were timing of category
information presentation (before vs. after observation), and motivation/accountability
(absent vs. present). The within subject factor was information consistency (category

consistent vs. category inconsistent).

Stimulus Materials
The stimulus materials consisted of videotaped incidents of good and poor

behavior exhibited by a secretary in a university department. The videotapes were part

29

30
of a larger set of videotapes and incidents developed for a study by Padgett and Ilgen
(1989). The videotape contained 10 behavioral incidents out of which 5 were instances
of good behavior and 5 were instances of poor behaviors. The behavioral incidents fell
into one or more of four categories. These categories were (1) Job Knowledge and
Skill, (2) Dealing with professors, (3) Working with Others, and (4) Organization of
Work. The level of performance for each of the behavioral incidents was determined by
having expert judges (secretaries) rate each of the incidents. Prior to making their
ratings, the raters were asked to indicate which of the four dimensions were represented
in the incident. They then rated the level of performance on the selected dimension(s)
using seven point BARS deveIOped speciﬁcally for clerical workers by Padgett and Ilgen
(1989). The BARS scale anchors for dimension 1 (Job Knowledge and Skill) can be
found in Figure 2. Appendix A contains the BARS scale anchors for the four
dimensions. In order to determine the overall rating for the incident, the rating(s) for
all the dimension(s) were averaged. The mean ratings for each incident on each
dimension as well as the overall rating can be found in Table 1. Table 2 contains a
description of each of these behavioral incidents. Only a subset of the behaviors was
chosen for this study in order to have equal number of good and poor performance
incidents. A behavioral incident was regarded as representing good performance if its

overall rating was above 4.00 and poor behavior if its overall rating was below 4.00.

31

Figure 2

BARS for Dimension 1: Job Knowledge and Skill

Job Knowledge and Skill: The extent to which the secretary has acquired information
relevant to doing his/her job (e.g. about university policies and procedures) as well as
knowledge about the different technical aspects of his/her job and the extent to which
he/she is able to execute them skillfully.

 

 

 

Anchor Description

7 Secretary could be expected to correctly answer questions
requiring specialized knowledge about university policies and
procedures.

Above Average
Performance

6 Secretary could be expected to have mastered the ﬁner points of
operating complicated ofﬁce equipment.

5 Secretary could be expected to correctly ﬁx spelling and
grammatical errors in letters and manuscripts he/she is asked to
type.

4 Secretary could be expected to follow correct style requirements

Average typing letters and manuscripts.
Performance

3 Secretary could be expected to turn in manuscripts or letters that
contain typographical errors.

2

Below Average
Performance
1

Secretary could be expected to be unable to properly operate
equipment used on the job (such as telephone, typewriter, or
copier).

32

Table 1

Experts’ Ratings of Behavioral Incidents Performed on the Videotape

 

Behavioral Job Dealing Working Organization Overall for
Incident Knowledge W/Profs W/Others of Work Incident
1 -- -- 1.5“ 1.5 1.50

2 -- 1.5 -- 1.3 1.40

3 5.4 5.9 -- -- 5.64,

4 2.1 2.9 -- -- 2.50

5 6.2 -- 5.9 -- 6.05

6 -- 3.4 -- 2.0 2.70

7 -- -- 1.4 -- 1.40

8 -- -- 6.4 -- 6.40

9 6.2 6.1 -- -- 6.16

10 -- 5.0 4.9 -- 4.93

 

' See Appendix A for a description of the seven point BARS scales used to rate each
of the four dimensions.

33

Table 2

Description of Behavioral Incidents

 

On Videotape

Behavioral

Incident Number Description

1. Gets angry at another secretary for not returning a borrowed stapler.

2. Does not get a professor’s presentation notes and overheads typed on time.

3. Repeats a phone message from the wife of one of the professors back to her to
be sure that it is correct.

4. Does not know that a particular professor is in her department and is unable to
tell a student where the professor’s ofﬁce is and what his ofﬁce hours are.

5. Demonstrates detailed knowledge about the various idiosyncrasies of the
professors in the department (e.g. about correcting grammatical errors in their
work or giving out their phone number).

6. Does low priority work, such as ﬁling, before typing exams, the latter of which
should be given top priority.

7. Refuses to change her plans to take a day off to accommodate the other
secretaries.

8. Brings up boxes from the mail room for another secretary.

9. Allows a professor to use the copier even though she needs to use it. She
instead goes down to the business library to use the copier.

10. Does not let another secretary see conﬁdential evaluation forms that she has

been trusted to type up.

34

minus

Upon arrival to the experimental site, subjects were given a consent form to read
and complete (see Appendix B). They were then asked to read a set of instructions
concerning the present study (see Appendix C). These instructions stated that they
were to watch a videotape which depicts a secretary working at a university. They were
also told that their task was to observe the secretary for the purpose of rating her
performance on a rating form as accurately as possible. A blank copy of the rating form
was shown to all subjects prior to their watching of the videotapes (see Appendix D).

The category was manipulated by ﬁlling out one of the rating forms depicting the
secretary was being "good" (i.e. circling fours and ﬁves on the rating scales) (see
Appendix E). Subjects were told that these ratings represented the mean ratings given
to the secretary by another group of students. Category timing was manipulated by
presenting subjects in the prior category with this rating form prior to viewing the
videotape. The post category subjects were presented with this form after viewing the
videotape and performing a ﬁller task.

Accountability was manipulated by telling subjects that they would have to justify
their responses to the other students in their group who were completing the task at
that time. These groups ranged in size from 2-7 with a mode of 6.

After viewing of the videotape, the subjects were given a 15 minute distractor
task (see Appendix F). This ﬁller task was a baseball and a general knowledge quiz.

All subjects were given ﬁfteen minutes to work on this ﬁller task and were told not to
talk or work on anything else if they ﬁnished early. The purpose of this ﬁller task was

to allow enough time to pass between observation and recognition in order to place a

35
greater cognitive demand on the participants. After the ﬁller task, subjects were given a
checklist containing the dependent measure. After completing this checklist the
participants were asked to rate the secretary on the four dimensions as well as an
overall rating. The rating form used is the same one containing the category
manipulation. The participants were then debriefed and told not to discuss the

experiment with other potential subjects.

Demndent Measures

Measures Derived From Behavioral Checklist. The main dependent measure in
this study is the checklist describing 20 behavioral incidents (see Appendix G). Half of
them were actually represented on the videotape and the other half were not. Of each
of these, 5 described instances of good behavior (consistent) and 5 described instances
of poor behavior (inconsistent). The incidents which occurred on the videotape are
those presented in Table 2. The behaviors which did not occur in the videotape were
derived from Padgett and Ilgen (1989). Subjects were told to check off all incidents
that they recalled observing in the videotape. The checklist contained some blanks
where the subjects could write themselves notes that might help them in recognizing the
behaviors. The checklist was used to compute proportions of consistent and inconsistent
information that was correctly and incorrectly recognized. The following dependent
measures were calculated from the responses to this checklist:

Proportion of consistent n

behaviors correctly = P = --
Identiﬁed 5

Proportion of inconsistent nlc

behaviors correctly = P":= --

Identiﬁed 5

Proportion of consistent “ct

behaviors incorrectly = Pci = --

identiﬁed 5

Proportion of inconsistent nH

behaviors incorrectly = P1i = --

identiﬁed 5

Accuracy for Consistent Behaviors = Ac: Pec'PCi
Accuracy for Inconsistent Behaviors = A,= Pm-Pn

Accuracy for Total behaviors

II
P
II

Act“:

Other Measures. Several other measures were obtained. These were the

following:

TIME: Time in minutes spent working on the checklist. This measure was
obtained by having the experimenter write down the time at which the
participants in the group began working on the checklist. The
participants in the group began working on the checklist at the same
time. The participants were told to turn over the checklist as soon as
they were ﬁnished. The experimenter then wrote down the time at which
each of the participants ﬁnished the checklist. By subtracting the starting
time from the ending time, a measure of time spent working on the

checklist was obtained.

37

RATINGS: The participants were asked to rate the secretary on the four
dimensions using the checklist in Appendix E. The checklist also

contained a scale for an overall rating.

RESULTS

Descriptive Data

Table 3 contains the means, standard deviations, and intercorrelations for the

variables in the study.

Manipulation Checks

In order to see if the accountability manipulation had the effect of increasing
motivation in ﬁlling out the checklist, a t-test of the time spent working on this form
was conducted. The results of this test can be found in Table 4. As the results
indicate, the no accountability group spent an average of 2.7 (SD=1.3) minutes working
on the checklist versus 3.8 (SD=2.2) minutes for the accountability group. These mean
differences were statistically signiﬁcant (t=-3.31, df (130), p<.01), and the manipulation

accounted for 7.8% of the variance in time spent (eta2=.077).

Analjgis of Hmtheses

Overall Tests. A 2 X 2 X 2 repeated measures analysis of variance was
conducted to assess the impact of timing of category and motivation on the recognition
of category wnsistent and inconsistent behaviors (Hypotheses 1a, 1b, 5). The dependent

variables were accuracy for consistent (Ar) and inconsistent (A!) behaviors which made

38

39
up the within subjects effect of Consistency. The between subjects effects were Timing
(pre vs post) and Motivation (absent or present).

The results of this analysis are presented in Table 5. The results indicated a
nonsigniﬁcant interaction between consistency, timing, and motivation as was predicted
by Hypothesis 5 (F<1). Hypotheses 1a and 1b predicted an interaction between
consistency and timing. Results from Table 5 indicate a marginal effect (F=3.11,
df(1,128), p<.10). Table 6 presents the means associated with this interaction. An
inspection of the means indicate that this effect is in the opposite direction than that
predicted. The means as well as the signiﬁcant main effect for consistency indicate that
overall, subjects in both timing conditions recognized more category inconsistent
behaviors. The interaction was due to the fact that in the pre condition, subjects
recognized a much larger proportion of category inconsistent than consistent behaviors
whereas those in the post condition recognized almost equal proportion of consistent
and inconsistent behaviors (see Figure 3).

Overall Accuracy. The effects of category and motivation on overall accuracy
were also tested. A 2 X 2 ANOVA with overall accuracy (A1) as the dependent
variable was conducted. Table 7 presents the results of this analysis. Hypothesis 2
predicted that subjects presented with the category prior to observation will more
accurately recognize a greater number of behaviors than those receiving the category
after observation. The results indicated a non-signiﬁcant effect for timing (F<1).
Hypothesis 4 predicted that subjects in the motivation condition will more accurately
recognize a greater number of behaviors than those in the no motivation condition.
The results indicate a marginally signiﬁcant effect for motivation (F=3.63, df (1,130),

p<.10).

40

.§:§§«8§8§ .moaanqopluognnﬁn
8.3ﬂnaouﬁqoﬂuo8oﬁonﬂma

.5

Saggguogﬂﬂouag .moaaumgoﬁumuom.

.~+B«.£§o§noonn

 

 

4+87£oohuo§8on~

A on o 3 g 3538 .

«an: moi—L.

.mm. 3. 8. .S. 8... .56 8. .8. «a. B. odn “436.3
8. 8. a. 8... .26 .8... .2... .3. B. on." Ed

8... .8... 8. no. mo... 3. 3. no.” 9...... a.»

.2. .mm. .mm.- .8. .8... .2.. an. 34 ”£4.

.8. .8... .3. .36 .mm. 2. 8. ~66

.9... .«n. .8... .8. an. 2.. note

8.- .3. 8.: 3. 8. .:n .e

B... .8. 3. mm. .2n .n

O O O ’U .

5. B. .an A

S m m n m m e n n A B 25! mg

41

Table 4

Manipulation Check for Accountability

Accountabilig Time (min) S_D_ t 9
Absent 2.72 1.33

Present 3.78 2.23 3.31 .001

42

Table 5

AN OVA Summary Table of Consistency by Timing by Motivation

With Accuracy (Ac and A!) as the Dependent Variable

 

SOURCE gr M_S 13 eta2
Timing (T) 1 .05 .97 .00
Motivation (M) 1 .18 3.63" .02
T x M 1 .00 .09 .00
Within (ser) 128 .05

Consistency (C)° 1 .36 13.59“ .03
T x C 1 .08 3.11" .01
M x C 1 .01 .54 .00
T x M x R 1 02 69 .00
Within (CS_/TM) 128 .03

' p<.10

" p<.05

“Consistency is 8 within subjects factor of accuracy for consistent (Ac) and inconsistent
(Ar) behaviors

43

Table 6

Means of Consistency by Timing

Consistent Inconsistent
Pre Ac=.752° Al =.861
Post Ac=.814 AI =.852

 

' The numbers indicate the proportion of behaviors correctly recalled after subtracting
the proportion of behaviors incorrectly recalled (i.e. corrected for false alarms).

Figure 3

Interaction of Consistency by Timing

 

 

0.975
0.95
0.9 25
0.9
0.875
0.85
0.825
0.8
0.775
0.75
0.725
0.7
0.875
0.55

 

 

 

 

—-— re' °°°°° “X"?b'lt

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Conslstent

Inconslstent

 

 

 

45

Table 7

ANOVA Summary Table of Total Accuracy (A1) by Timing by Motivation

.___SOURCE Eli M_5 E Q3.
Timing (T) 1 .09 .97 .01
Motivation (M) 1 .35 3.63' .03
T x M 1 .01 .09 .00
Within (srrM) 128 .10

 

* p<.10

46
An analysis of the cell means indicates that subjects in motivation condition actually had
lower overall accuracy (m=1.59) than those in the no motivation condition (m=1.69).

Thus the trend was in the opposite direction than predicted by Hypothesis 4.

False Alarm Rates. Hypothesis 3 stated that subjects who received category
information prior to observation would have a greater false alarm rate for category-
consistent behaviors than for category-inconsistent behaviors. To explore this effect an
analysis of variance with the proportions of incorrectly identiﬁed behaviors (consistent
and inconsistent) as the dependent variables was conducted. The independent variables
were timing and motivation. The hypothesis predicted a signiﬁcant interaction between
consistency and timing. The results of this analysis are presented in Table 8. The
results indicate a signiﬁcant interaction between consistency and timing (F=4.34, df
(1,130), p<.05). Table 9 presents the means associated with this interaction. The
means indicate that subjects receiving the category prior to observation had nearly equal
number of false alarms for consistent and inconsistent behaviors. Thus Hypothesis 3 was
not supported. The interaction was due to the fact that subjects receiving the category
after observation had a greater false alarm rate for inconsistent (m=.10) than consistent

(m=.03) behaviors.

47

Table 8

AN OVA Summary Table of False Alarms (Pu, P”) by Timing by Motivation

____SOURCE .df Mi 1:" Sit—33
Timing (T) 1 .02 48 .00
Motivation (M) 1 .01 .57 .00
T x M 1 - .00 .05 .00
Within (S/I'M) 128 .02

 

Consistency (C) 1 .13 9.47* .03
T x C 1 .06 4.24“ .01
M x C 1 .02 1.64 .00
T x M x C 1 .01 .94

Within (CS_/I'M) 128 .01

' p<.01

" p<.05

Consistent (Pei)

Inconsistent (PI i)

Table 9

Means of False Alarms by Timing

Pre Post
.07 03
.08 10

DISCUSSION

The purpose of this study was to test the effects of a category label on
information recognition. More speciﬁcally, the purpose was to test the effects of
providing a category label either prior to, or after, observing a ratee. The introduction
of a category at these two points in time was hypothesized to differentially affect
recognition of information about the ratee by either ﬁltering information going into
memory or organizing information previously stored in memory. These effects were
studied by providing information about the ratee which was either consistent or
inconsistent with the category label. Then, the effects of providing the category label at
the two points in time was assessed by its impact on the recognition of category
consistent and inconsistent behaviors.

The discussion of this study is organized in three sections. First a summary of
the study results is presented as well as a summary and discussion of unexpected
ﬁndings. Second, limitations of the study are presented and discussed in terms of their
possible effects on the ﬁndings in this study. Finally, a discussion of future research is

presented.

49

50

Summary of Results

Overall Tests. The main thesis presented in this study was that a category
presented prior to observation affects the encoding of information by ﬁltering out
category inconsistent information (Hypothesis 1a). A category presented after
observation was hypothesized to bias recognition by organizing the information stored in
memory prior to receiving the category. This organization effect was hypothesized to
result in a greater false alarm rate for category-consistent behaviors hence resulting in
greater accuracy for category-inconsistent than consistent behaviors (Hypothesis 1b).
These hypotheses predicted an interaction between consistency and timing. The results
indicated a marginal interaction (F=3.11, df (1,128), p<.10). However, an examination
of the means indicated that subjects receiving the category label prior to observation
recognized a greater proportion of category-inconsistent behaviors whereas for those
receiving the category label after observation this difference was not as great. This
marginal interaction suggests that the category manipulation may have not been strong
enough to have the hypothesized effect. This possibility will be discussed in more detail
later.

Hypothesis 5 stated that under conditions of heightened motivation to recall
accurately, those who received the category label after they have been presented with
information about the ratee and presumably stored it in memory should be more able to
counteract the bias created by the category as opposed to those receiving the category
label prior to observation. The latter group should have biased information stored in
memory due to the ﬁltering effect of providing a category label prior to observation and

motivation should have not effect in overcoming this bias.

51

This hypothesis predicted a three-way interaction between consistency, category
timing, and motivation, with recognition accuracy for consistent and inconsistent
information as the dependent variable. This interaction was not observed. The ability
to detect this interaction depended on ﬁnding the interaction predicted by Hypotheses
1a and 1b. However, since this interaction was only marginally signiﬁcant, it is not
surprising that the three way interaction was not signiﬁcant. There are several possible
explanation for the failure to ﬁnd an effect for the category manipulation.

The participants in the study were presented with a category label depicting the
secretary as being a good performer. Furthermore, the videotape contained equal
number of good and poor performance instances. However, the results indicate that all
subjects had greater accuracy in recognizing category-inconsistent behaviors. In short,
subjects were more accurate in recognizing instances of poor performance. The design
of this study deﬁned consistency and inconsistency in relation to the category presented
to the participants. This category depicted the secretary as being a good worker.
Therefore, consistent behaviors were those which were good behaviors and inconsistent
were poor behaviors. The fact that this design confounds consistency with performance
level is important in the failure to ﬁnd some of the hypothesized results.

Deﬁning consistency in relation to the ”good secretary” category presented to the
subjects makes the results of this study depend on the subjects accepting that category
as their impression of the secretary at the point in time at which it is introduced. This
category should then affect information acquisition in the pre group and information
recall in the post group. If for some reason the category does not become the subjects’
own impression, then it is very unlikely that the hypothesized effects for timing would be

observed. There was some evidence to suggest that the category label of "good

52
secretary” did not have the effect of making the subjects observe or recognize the
secretary with this category in mind.

If the category label was not accepted by the subjects, then it is possible that
subjects formed their own impression and observed and recognized information
according to that impression. One way to examine this possibility was to reanalize the
data using the subjects’ own impression of the secretary as the category. A surragate
measure of the subjects’ impression of the secretary was the overall rating from the
rating form which the subjects completed at the end of the experiment. The mean
rating given to the secretary was 2.19 (SD=.69) on a 5 point scale. This represents an
average rating below the midpoint. Furthermore, 68% of the subjects rated the
secretary 2 or below and 98% rated her 3 or below on this 5 point scale. These ratings
were not affected by the time at which the category was introduced. This is evidenced
by the results of an ANOVA of ratings by timing which was nonsigniﬁcant (F<1).

A further test of the hypothesis that the subjects formed their own impression
involves analyzing some of the study hypotheses using the subjects’ overall rating as a
measure of their impression. Given the fact that there was no difference in ratings by
timing, subjects were collapsed across the two levels of timing. Subjects who rated the
secretary a 3 or above on their overall rating were put into a “good impression" group
(n=41). Those which rated her 2 or below were put into a ”poor impression" group
(n=88). An analysis of variance was conducted with accuracy for good (Ac) and poor
(Ar) behaviors as the dependent variables. The independent variable was the impression
formed (good or poor). The results of this analysis can be found in Table 10. A main
effect for Consistency was found (F= 18.95, df (1,127), p<.001). This indicates that

overall, the subjects had greater accuracy for poor behaviors. The results also showed

53
an interaction between Consistency and Impression (F=4.80, df (1,127) p<.05). Figure
4 shows the means associated with this interaction. These results indicate that subjects
with a "good impression” had greater accuracy for poor behaviors. Subjects with a "poor
impression” also had greater accuracy for poor behaviors than good ones but this
difference was not nearly as large. This interaction indicates a trend toward greater
accuracy for behaviors inconsistent with the subjects’ impression. This result is
consistent with Hastie’s (1980) study who found that when subjects were asked to form
an impression of a target person, inconsistent behaviors were better recalled. Thus, it
appears that even though a category was presented to the subjects, they still formed
their own impression which was different from the manipulation. A possible explanation
for this ﬁnding can be found in the intructions given to the subjects. Even though they
were presented with the category, the subjects were told to watch the videotape for the
purpose of rating the secretary as accurately as possible. So, although the subjects were
presented with a category label which represented another group of students’ impression
of the secretary, they might have felt that they needed to form their own impression in
order to be as accurate as possible in their ratings. It is possible that a stronger
manipulation of category might have overcome this ambiguity in the instructions.
Another way to manipulate the category could be to let the subjects view a videotape of
the secretary performing all good behaviors either prior to or after showing them the

tape with good and poor performance instances.

54

Table 10

AN OVA Summary Table of Consistency by Impression

__SOURCE Qf M_S E 91.8.2.
Impression (1) 1 .10 16 .01
Within (_S_/I) 127 .05

Consistency (C) 1 .50 18.95' .05
Ix C 1 .13 4.80” .01
Within (CS_/I) 127 .03

 

' p<.001
" p<.05

55

Figure 4

Interaction of Consistency by Impression

 

 

 

 

 

 

 

 

 

0.970 “‘ CGD’CTIU"""‘°X"’N5P'INI§

 

0.95

 

0.925

 

09 /

 

0.875 /
0.85 /

 

0.825 / ,x

 

’
’
I
’
I
I
0.8

 

 

0.775 (

 

0.75

 

0.725

 

0.7

 

0.575

 

 

 

0.85
Good Iehsvlore Poor Iohevlore

 

 

56

Overall Accuracy. Overall accuracy (A1) was deﬁned as the sum of the
pr0portions of category-consistent and inconsistent behaviors after subtracting the
proportions of category consistent and inconsistent behaviors incorrectly identiﬁed.
Hypothesis 2 predicted that subjects receiving the category label prior to observation
would have greater overall accuracy than those receiving the category label after
observation. Contrary to Hypothesis 2, the results indicated a nonsigniﬁcant effect for
timing of category. As stated earlier, a possible reason for the nonsigniﬁcant interaction
was the failure of the category manipulation to make salient a "good secretary”
impression to the subjects at two different points in time. If this were the case, then
receiving the category label prior to observation may have formed their own impression
in a manner similar to those receiving the category after observation. The data are
consistent with this interpretation due to the absence of an effect for category timing on
overall accuracy.

Hypothesis 4 predicted that subjects in the motivation condition would have
greater overall accuracy than those in the no motivation condition. Although a marginal
effect was found, it was in the opposite direction. Subjects in the no motivation
condition tended to have greater accuracy than those in the motivation condition.

There are at least two possible explanations for this. Given that accuracy is dependent
on the proportion of behaviors correctly as well as incorrectly identiﬁed, accountability
could have had an impact on either of these two measures. As a result, heightened
motivation could have decreased accuracy by (1) increasing the number of false alarms
or (2) decreasing the number of behaviors correctly identiﬁed. Each of these possible

explanations was explored separately.

57

In order to test the possibility that motivation increased the number of false
alarms, the effects of motivation on false alarms were investigated using an analysis of
variance. The dependent variable was false alarms for consistent (Pei) and inconsistent
(Pr i) behaviors as the within subject factor of Consistency. The independent variable
was motivation. No effect for motivation on false alarms was found.

The second possible explanation was that motivation caused subjects to record
fewer correct behaviors. Such a response would have occured if the increased
motivation led the subjects to be more cautious in an attempt to reduce their mistakes,
and it would decrease overall accuracy if the number of false alarms stayed constant.

An analysis of variance was conducted with the proportion of consistent (Pee) and
inconsistent (P, c) behaviors correctly identiﬁed. A signiﬁcant main effect for motivation
was found (F=3.93, p<.05, eta2=.03). The pattern of the means indicated that the high
motivation group recorded fewer correct behaviors (m=.87) than the no motivation
group (m=.91). This is consistent with the hypothesis that the motivation group had
lower overall accuracy because they recorded fewer items. A possible explanation for
this ﬁnding is that manipulating motivation with accountability might have had the effect
of causing accountable subjects to mark only those behaviors of which they were
absolutely sure because they knew they would have to justify their answers to the rest of
the group members. It should be noted that these explanations are purely exploratory.
Furthermore, the original interaction between accuracy and motivation was only
marginally signiﬁcant. A stronger test of this hypothesis would be to have subjects rate
their certainty of having observed a particular behavior. One would expect a higher

degree of certainty for the no accountability group on more total behaviors than the

58
accountability group. The latter group might show a high degree of certainty on a fewer

number of behaviors.

False Alarms. Hypothesis 3 predicted that subjects receiving the category
information after observation would have a greater false alarm rate for category-
consistent than category-inconsistent behaviors. The results indicated a signiﬁcant main
effect for Consistency and 8 Consistency X Timing interaction when false alarms are
used as the within subjects factor of Consistency. The main effect indicated that all
subjects had a greater false alarm rate for poor behaviors (inconsistent). The interaction
indicated that subjects receiving the category after observation showed a greater
difference between the number of consistent and inconsistent behaviors incorrectly
identiﬁed with inconsistent behaviors having the larger false alarm rate (see Figure 5).
Given the fact that consistency was deﬁned in terms of the ”good secretary" category
presented, the hypothesis was not supported because poor performance was inconsistent
with this category. However, since the data seemed to suggest that subjects formed a
negative impression of the secretary on their own, then poor behaviors were consistent
with this impression. Therefore, the results would not have been the opposite predicted
if one assumes that the category held by the subjects was poor rather than good
performance.

Further evidence for the conclusion that the results must be interpreted using
the subjects’ own impression is the correlation between overall ratings and proportion of
consistent and inconsistent behaviors incorrectly identiﬁed. The correlation between
subjects’ overall ratings and the proportion of good performance instances incorrectly

identiﬁed was .20 (n=129, p<.05), and it was -.21 (n=129, p<.05) with poor

59
performance instances incorrectly identiﬁed. That is, the more positive the subjects’
impression were, the higher the false alarm rate for good behaviors. Conversely, more
positive impressions were related to lower false alarm rates for poor behaviors. These
results are consistent with those of Phillips and Lord (1982) who found that subjects had
greater false alarms for behaviors consistent with their overall impression.

An issue which needs to be resolved is the fact that category timing had an
effect on false alarms as well as accuracy as evidenced by the Consistency x Timing
interaction for Hypotheses 1a, 1b, and 3. As has already been stated, the failure to
make the ”good secretary" category the one used by subjects to process and recognize
information led to the conclusion that the subjects were forming their own impression
and the category manipultation was not strong enough to counter this impression.
Evidence for this was found in the fact that when the data were analyzed using the
subjects’ impression as the category, they were consistent with past literature. However,
the interaction between false alarms and timing seems to suggest that the category might
have had some effect. But the differences in false alarms, which amounted to less than
one half of a behavior difference, may not have been enough to affect accuracy. This
ﬁnding points to the possibility of restriction of range limiting the ﬁndings in the present

study. This will be discussed in the following section.

Figure 5

False Alarms by Timing

 

 

0.1 3
0.12
0.1 1

0.1
0.09
0.08
0.07
0.08
0.05
0.04
0.03
0.02

0.0 1

 

 

 

4‘9“ ------- 'X-ﬁut

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Consletent Inconslstem

 

 

 

61

Limitations of the Study

There are several limitations to this study that could pose a potential threat to its
internal validity. The first has to do with the category manipulation. In general, the
subjects did not accept the manipulation that described the secretary’s performance as
good. If, rather than using "good secretary” as the category for evaluating the results of
the study, the subjects’ reported impression is used, the results are more similar to those
predicted. Thus, it appears that subjects might have paid attention to this manipulation
but then felt free to form their own impression. Furthermore, since subjects were told
that they would be rating the secretary, this could have had the effect of asking subjects
to form their own impression.

The second limitation has to do with restriction of range on the dependent
variables. A ﬁller task was used to take the subjects’ mind off the tape before ﬁlling
out the checklist, but it appears that the amount of time taken up by this task may have
not been enough. Most subjects recognized an average of 4 behaviors (out of 5) for
each of the two performance types (good and poor). Furthermore, most subjects falsely
identiﬁed less than 1 behavior of each type. It appears that either a longer delay
between observation and recognition and/or the inclusion of category irrelevant
information might have helped to make the subjects to rely on a category which is given
to them in order to reduce the cognitive load. This restriction of range is important
because most of the hypotheses in the present study predicted diﬂ'erences across
conditions due to penalties imposed by subtracting false alarms. However, when false
alarms account for less that one half of a behavior difference across groups, it is not

surprising that no difference in accuracy was found.

62
A ﬁnal limitation which has already been discussed has to do with the motivation
manipulation. It appears that instead of motivating the subjects to recognize as much as
possible, the manipulation served to make subjects more cautious when answering the
checklist. This effect severely limited the possibility of ﬁnding an effect for the main
hypothesis presented in this study because it relied on this manipulation increasing

accuracy and not decreasing it.

Future Research

There are several ways of improving the present study which future research can
address. The ﬁrst issue has to do with the cognitive demands placed on the subjects. A
longer delay between observation and recognition should make the raters rely more on a
performance category. Also the inclusion of category-irrelevant behaviors should aid in
making the task more cognitively demanding.

Another issue is that of motivating individuals to accurately recognize behaviors.
As has already been discussed, accountability in this case appeared to serve to make the
raters mark only those behaviors which they thought they could defend. To explore this
more, it would have helped to have subjects give conﬁdence ratings along with their
recall responses. One would expect that accountability subjects would show a high
degree of conﬁdence on the behaviors that they identify as having been seen when
compared to the conﬁdence ratings of the no accountability raters. If this were the
case, then it would suggest that making managers accountable for justifying their ratings
in terms of the behaviors observed could lead to ratings based on only a few behavioral
instances. This is only a problem if the behaviors that are recalled are not

representative of the total behaviors exhibited by the ratee. So for example, if a

63
subordinate exhibits 40 instances of good performance and 10 instances of poor
performance and the manager is only certain of 4 instances of good behavior and 1
instance of poor behavior, then the ratings of the subordinate would probably not be
biased. However, if the manager is sure of 10 instances of poor behavior and only 5 of
good behavior bias should be more likely.

Future studies should address the question of what happens when the category
which is introduced is opposite to the impression that participants form on their own. If
the participants in the present study formed a poor impression, then those receiving the
category after observation were receiving a category which was opposite to the one
which they formed on their own. It is possible that in this situation, characteristics of
the source, such as source credibility, become important. One may be more apt to
change his or her impression if the source is highly credible. This effect has been found
in research dealing with attitude change (e.g. see review by Chaiken & Stangor, 1987).
However, in the present study, characteristics of the source of the category were kept
constant. It is interesting to think of what would be the effects of substituting a
subject’s impression with another from a more or less credible source. Would the
effects on recall occur in relation to their original impression or with the newly acquired
impression? The issues raised above should help in eliminating some of the weaknesses
in the present study and gaining further understanding of the effects of a category on

the recall of behavior.

LIST OF REFERENCES

LIST OF REFERENCES

Ausubel, DP. (1968). Educational psychology: A cognitive view. New York: Holt,
Rinehart & Winston.

Bargh, J .A., & Thein, RD. (1985). Individual construct accessibility, person memory,
and the recall-judgement link: The case of information overload. Journal of

Personality and Social Psychology, 42, 1129-1146.
Beach, LR, & Mitchell, T.R. (1978). A contingency model for the selection of decision

strategies. Academy of Management Review, 3, 439-449.
Belleza, F.S., & Bower, GB. (1981). Person stereotypes and memory for people.

Journal of Personalig and Social nghology, 4_1, 856-865.

Bodenhausen, G.V. (1988). Stereotypic biases in social decision making and memory:
Testing process models of stereotype use. Journal of Personalig and Social
Psychology, 55, 726-737.

Cantor, N. & Mischel, W. (1979). Prototypes in person perception. In L. Berkowitz
(Ed), Advances in experimental social psychology (Vol 12). New York: Academic
Press.

Chaiken, S. & Stagnor, C. (1987). Attitudes and attitude change. Annual Review of

Psychology, .3_8, 575-630.

65

Clark, LR., & Woll, SB. (1981). Stereotype biases: A reconstructive analysis of the
role in reconstructive memory. Journal of Personalig and Social Psychology, 4_1,
1064-1072.

Cohen, CE. (1981). Person categories and social perception: Testing some boundaries
of the processing effects of prior knowledge. Journal of Personali‘g and Social
Psychology, ﬂ, 441-452.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences 2nd ed.

Hillsdale, N.J.: Lawrence Erlbaum Associates.

Crocker, J., Hannah, D.B., & Weber, R. (1983). Person memory and causal attributions.
Journal of Personality and Social Psychology, ﬂ, 55-66.

DeCotiis, T., & Petit, A. (1978). The performance appraisal process: A model and some
testable hypothesis. Academy of Management Review, 21, 635-646.

DeNisi, A.S., Cafferty, T.P., & Meglino, BM. (1984). A cognitive view of the
performance appraisal process: A model and research propositions.

Organizational Behavior and Human Performance, 3, 360-396.

Feldman, J .M. (1981). Beyond Attribution Theory: Cognitive processes in performance
appraisal. Journal of Applied Psychology, Q, 127—148.

Fiske, S.T., & Taylor, SE. (1984). Social Cognition. Reading, MA: Addison-Wesley.

Gillund, G. & Shiffrin, RM. (1984). A retrieval model for both recognition and recall.
chhological Review, 9_1_, 1-67.

Hastie, R. (1980). Memory for behavioral information that conﬁrms or contradicts a
personality impression. In R. Hastie, T. Ostrom, E. Ebbensen, R. Wyer, D. '
Hamilton, & D. Carlston (Eds), Person memory: Cognitive basis of social

perception (pp. 155-178), Hillsdale, N.J.: Erlbaum.

66

Higgins, ET., & Bargh, IA. (1987). Social cognition and social perception. M
Review of chhology, 58, 369-425.

Howell, W.C. & Cooke, NJ. (1989). Training the human information processor: A
review of cognitive models. In Goldstein, I.L. & Associates (Eds), Training and
Development in Organizations (pp. 121-182), San Francisco, CA: Jossey-Bass.

Ilgen, D.R. & Feldman, J .M. (1983). Performance Appraisal: A process focus.

Research in Organizational Behayjor, 5, 141-197.

Landy, F.J. & Farr, IL. (1980). Performance Rating. Psychological Bulletin, 8_7, 72-107.

Lichtenstein, M., & Srull, T.K. (1987). Processing objectives as a determinant of the
relationship between recall and judgement. Jougpal of E_xmrimental Social
Psychology, 25, 93-118.

Lord, RG. (1985). Accuracy in behavioral measurement: An alternative deﬁnition based
on rater’s cognitive schema and signal detection theory. Journal of Applied
Psychology, 19, 66-71.

Marcy, LM. (1990). Accountability of the decision maker: A closer examination of it’s
role in decision making. Unpublished Master’s Thesis. Michigan State
University, East Lansing, MI.

Mayer, R.E. & Bromage, B.K. (1980). Different recall protocols for technical texts due
to advance organizers. Journal of Eucational chhology, 2, 209-225.

McAllister, D.W., Mitchell, T.R., & Beach, LR. (1979). The contingency model for the
selection of decision strategies: An empirical test of the effects of signiﬁcance,

accountability, and reversibility. Organizational Behavior and Human
Eepfomance, 24, 228-244.

67
Nathan, B.R. & Lord, RC. (1983). Cognitive categorization and dimensional schemata:
A process approach to the study of halo in performance ratings. Journal of

Applied Psychology, §§, 102-114.
O’Sullivan, C.S., & Durso, ET. (1984). Effect of schema-incongruent information on

memory for stereotypical attributes. Journal of Personalig and Social Psychology,
ﬂ, 55-70.

Padgett, M.Y. & Ilgen, DR. (1989). The impact of ratee performance characteristics on
rater cognitive processes and alternative measures of rater accuracy.

Organizational Behavior and Human Decision Processes, 14, 232-260.
Phillips, J.S., & Lord, RC. (1982). Schematic information processing and perceptions of

leadership in problem-solving groups. Journal of Applied Psychology, 51,
486-492.
Pyszczynski, T., LaPrelle, J., & Greenberg, J. (1987). Encoding and retrieval effects of

general person characterizations on memory for incongruent and congruent

information. Personaligj and Social Psychology Bulletin, 1_3, 556-567.
Rosch, E., Mervis, C.B., Gray, W.D., Johnson, D.M., & Boyes-Braem, P. Basic objects

in natural categories. Cogpitive Psychology, 8, 382-439.

Rothbart, M., Evans, M., & Fulero, S. (1979). Recall for conﬁrming events: Memory
processes and the maintenance of social stereotypes. Journal of ngrimental
Social Psychology, g, 343-355.

Rozelle, R.M. & Baxter, J .C. (1981). Inﬂuence of role pressures on the perceiver:

Judgments of videotaped interviews varying judge accountability and

responsibility. Journal of Applied Psychology, Q, 437-441.

68

Snyder, M., & Uranowitz, S.W. (1978). Reconstructing the past: Some cognitive
consequences of person perception. Journal of Personalig and Social
chhology, 55, 941-950.

Srull, T.K. (1981). Person memory: Some tests of associative storage and retrieval
model. Journal of momenta] chhologyg. Human Learning and Memog, '_7_,
440-463.

Srull, T.K., & Wyer, RS. (1980). Category accessibility and social perception: Some
implications for the study of person memory and interpersonal judgments.
Journal of Personality and Social Psychology, 88, 841-856.

Srull, T.K., & Wyer, RS. (1989). Person memory and judgement. Psychological

Review ﬂ, 58-83.

 

Taft, R. (1971). The ability to judge people. In W.W. Ronan & E.P. Prien (Eds),
Perspectives on the Measurement of Human Performance. New York:
Appleton-Century-Crofts.

Tetlock, RE. (1983). Accountability and the complexity of thought. Journal of
Personality and Social Psychology, 55, 74-83.

Tetlock, RE. (1985). Accountability: The neglected social context of judgement and

choice. In B.M. Staw & L. Cummings (Eds), Research in Organizational

Behavior Vol. 7. Greenwich, CI‘: JAI Press.

 

Tetlock, P.E., & Kim, M. (1987). Accountability and judgement processes in a

personality prediction task. Journal of Personalig and Social Psychology, 5;,
700-709.

Wyer, R.S., & Gordon, S.E. (1982). The recall of information about person and groups.

Journal of gmrimental Social Psychology, E, 128-164.

69

Wyer, R.S., & Gordon, SE (1984). The cognitive representation of social information.
In R.S. Wyer & T.K. Srull (Eds), Handbook of social cognition (Vol. 2, pp.
73-150). Hillsdale, N.J.: Erlbaum.

Wyer, R.S., Bodenhausen, G.V., & Srull, T.K. (1984). The cognitive representation of
persons and groups and its effect on recall and recognition memory. Journal of
Epsrimental Social chhology, E. 445-469.

Wyer, R.S., Srull, T.K., Gordon, S.E., & Hartwick, J. (1982). Effects of processing
objectives on the recall of prose material. Journal of Personality and Social

Psychology, 13, 674-688.

APPENDICES

APPENDIX A

70

BARS for Dimension 1: Job Knowledge and Skill

Job Knowledge and Skill: The extent to which the secretary has acquired information
relevant to doing his/her job (e.g. about university policies and procedures) as well as
knowledge about the different technical aspects of his/her job and the extent to which
he/she is able to execute them skillfully.

Anchor Description

 

7 Secretary could be expected to correctly answer questions
requiring specialized knowledge about university policies and
procedures.

Above Average
Performance

6 Secretary could be expected to have mastered the ﬁner points of

operating complicated ofﬁce equipment.

 

5 Secretary could be expected to correctly ﬁx spelling and
grammatical errors in letters and manuscripts he/she is asked to

type-
4 Secretary could be expected to follow correct style requirements
Average typing letters and manuscripts.
Performance
3 Secretary could be expected to turn in manuscripts or letters that

contain typographical errors.

 

2

Below Average
Performance
1 Secretary could be expected to be unable to properly operate
equipment used on the job (such as telephone, typewriter, or
copier).

71

BARS for Dimension 2: Dealing With Faculty, Students, etc.

Dealing With Faculg, StudentsI etc.: The extent to which the secretary interacts in a
positive manner with professors, students, callers, or other people that he/she comes into
contact with on the job (e.g.positive attitude, willing to help out)

Anchor Description

 

 

7 Secretary could be expected to volunteer to take work home with
him/her in order to get an emergency assignment done for a
professor on time.
Above Average
Performance
6 Secretary could be expected to offer help a professor with some
of their special work that isn’t part of his/her regular job duties.

 

5 Secretary could be expected to be pleasant and helpful to people
who call the ofﬁce on the phone or who come in to ask
questions.

 

4
Average
Performance

3 Secretary could be expected to get upset and defensive when a
professor asks him/her to correct his/her work.

2 Secretary could be expected to gossip with other people in the
ofﬁce about administrators, professors or students that he/she
doesn’t like.

Below Average
Performance

72

BARS for Dimension 3: Working Cooperatively With Other Secretaries

Working Coopgratively With Other Secretaries: The extent to which the secretary works
as a part of a team with the other secretaries (e.g. by sharing the workload with them,
sharing important information with them, checking with them before taking time off,

 

 

 

etc.).
Anchor Description

7 Secretary could be expected to offer to help other secretaries with
their work if he/she was ﬁnished with his/her work.

Above Average
Performance

6 Secretary could be expected to share with the other secretaries
any new or special information he/she learns that might be helpful
to them.

5 Secretary could be expected to be willing to teach an
inexperienced ofﬁce worker basic information about ofﬁce
functioning.

4

Average
Performance

3 Secretary could be expected to agree to help if another secretary
asked him/her to help proofread a manuscript the other secretary
had typed.

2 Secretary could be expected to insult or ridicule another secretary
who makes a mistake on the job.

Below Average
Performance

73

BARS for Dimension 4: Organization of Work

Organization of Work: The extent to which the secretary is able to organize his/her time
and work to get things done efﬁciently, is able to set priorities concerning what work is
most important, is able to handle multiple demands on his/her time and can meet

 

 

 

deadlines.
Anchor Description

7 Secretary could be expected to develop an innovative procedure
that is more efﬁcient and saves work time.

Above Average
Performance

6 Secretary could be expected to use time when there is no work to
be done in the ofﬁce to straighten up and reorganize some ﬁles
rather than reading the paper or talking.

5 Secretary could be expected to prioritize work assignments so that
the most important things get done ﬁrst.

4 Secretary could be expected to have his/her work Performance
station organized so that he/she can quickly and easily locate
supplies and information that he/she needs.

Average
Performance

3 Secretary could be expected to be sufﬁciently organized the he/she
is able to get work assignments completed on time.

2 Secretary could be expected to get upset and ﬂustered if he/she
was in the ofﬁce alone when several people needed to speak with
her and the phone was ringing frequently.

Below Average
Performance
1

Secretary could be expected to miss important work assignment
deadlines.

APPENDIX B

74

CONSENT FORM
Employee Rating Study

This study is designed to investigate how people rate employees. You will be asked to
watch a short videotape of a secretary at work. You will then be asked a series of
questions relating to the tape.

Your participation in this research is completely voluntary and you are free to decline to
answer any questions or to terminate the session at any time. Termination prior to
completion of the session will not affect your credit for participation. Your participation
in this study will be conﬁdential. Your data will be included in a summary report along
with those of others, but will not be broken down in any way in which the responses
can be identiﬁed individually as yours. If you have any questions regarding the study,
you may ask the experimenter.

Subject’s Statement
1 agree to participate in the Employee Rating Study. I understand that I will be asked

to ﬁll out a questionnaire. Furthermore, I understand that I will be asked to ﬁll out a
questionnaire. Furthermore, I understand that my participation is voluntary, that I may
discontinue at any time without any adverse consequences to me, that my answers will

be anonymous, and that my responses will not be analyzed in any way that allows me to
be identiﬁed.

Signature
Date

 

Printed Name

 

Experimenter: Miguel Quiﬁones
22 Baker
355-2171

 

APPENDD( C

75
Employee Rating Study

Instructions

Welcome to the Employee Rating Study. In this study you will be playing the role of a
manager that has just begun a new job. One of your ﬁrst tasks is to rate the
performance of your new employees. Your employees are university secretaries working
in the management department. Please take a moment to examine the EMPLOYEE
EVALUATION FORM which you received from the experimenter. In this form there
is a rating scale ranging from Unacceptable to Outstanding behavior. This scale is used
to rate the employee on the four dimensions shown. These dimensions are (1) Job
Knowledge and Skill, (2) Organizational Ability, (3) Dealing with Faculty/Students, and
(4) Working Cooperatively with Others. There is also a section for an overall
evaluation and some comments.

You will be watching a videotape of your employee during a normal work day. It is
your task to watch this videotape for the purpose of rating the employee as
ACCURATELY as possible. If you have any questions please ask the experimenter.

APPENDIX D

76

EMPLOYEE EVALUATION FORM

Employee Name Job Title

 

Review Date

 

 

Use the following scale to rate the employee's performance on the following
dimensions.

 

 

 

Unaccgptable Belzow Aveiage Abgve Outstfnding
Average Average
Job Knowledge and Skill 1 2 3 d 5
Organizational Ability 1 2 3 4 5
Dealing With Faulty/Students 1 2 3 4 5
Working Cooperatively With Others 1 2 3 l 5

 

 

Overall Rating 1 2 3 4 5

 

Comments

OOOOOOOOOOOOOOOOOOOOO

 

 

 

 

 

 

 

APPENDIX E

77

EMPLOYEE EVALUATION FORM

Employee Name 5"" “9"” job Title Secret"? 1

Review Date 1/15/89

 

Use the following scale to rate the employee's performance on the following
dimensions.

 

 

 

Unaccgptable Belzow Avezage Abgve Outstfnding
Average Average
Job Knowledge and Skill 1 2 3 4®
Organizational Ability 1 2 3 4 5
Dealing With Faculty/Students 1 2 3 4 6‘)
Working Cooperatively With Others 1 2 3 C4) 5

 

 

Overall Rating 1 2 3 4 (S)

 

omments
Sue is an excellent worker. She $5 very knowledgeable and gets along well

 

"'tiv'i'th"B't'li'é'r'ii'Im'Sli'e" is very efficient around the office. Sue is definitely

 

..An..ssss.t...tn..nrtr..denattnent.

 

 

 

 

 

 

 

APPENDD( F

78

Baseball Knowledge Quiz

Instructions: Read the following questions and circle the correct answer.

1. Who holds the record for the most home runs hit in a single season?

a. Roger Maris ' c. Reggie Jackson
b. Micky Mantle (1. Jose Canseco

2. How many feet are there between home plate and ﬁrst base?

a. 79.5 b. 100
c. 60 d. 90

3. The left ﬁeld wall in Boston’s Fenway Park is referred to as:

a. the ivy border c. the brick barrier
b. the wall of fame (I. the green monster

4. The "Cy Young Award" is an award for the best:

a. rookie c. pitcher
b. batter d. ﬁelder

5. A ”texas Leaguer" refers to:

a. a baseball player from Texas Minor Leagues
b. an error made by an outﬁelder

c. a hit made between an inﬁelder and outﬁelder
d. a rookie who is still "wet between the ears”

6. What is a squeeze play?

a. a runner on third runs home while the batter bunts the ball

b. two outﬁelders run into each other when chasing a ball

c. a shortstop and second baseman convert a double play

d. a pitcher pretends to throw the ball to home plate and the runner
at ﬁrst is tagged out by the ﬁrst baseman who has the ball

7. What is the name of the league that is immediately below the Major Leagues?

a. Texas League c. Independent League
b. AAA League d. Minor League

10.

11.

12.

13.

14.

15.

79
When a "balk" is called during a game what occurs?

a. the batter is given a strike

b. play is halted to allow the umpires to confer on a decision

c. runners are allowed to move to the next base

d. the pitcher is thrown out of the game for throwing a "spit" ball

The "inﬁeld ﬂy” rule is invoked only when:

a. a popup goes foul c. there are two outs already
b. there are no runners on base (1. none of the above

If the batter walks with the bases loaded:

a. the batter gets an RBI and is charged with a time at bat

b. the batter does not get an RBI 9; a time at bat

c. the batter gets the RBI but not the time at hat

(I. the batter does not get the RBI and is charged with a time at bat

How many innings does a starting pitcher have to work to be elegible for the win?

a. one c. ﬁve
b. nine d. seven

Who won last year’s (1989) World Series

a. Chicago Cubs c. Oakland Athletics
b. San Francisco Giants (1. Baltimore Orioles
A passed ball is:

a. the same as a wild pitch

b. charged as an error to the pitcher
c. charged as an error to the catcher
d. both a and b

Using standard scorekeeping notation, a 5-4-3 play starts with:

a. the shortstop c. the third baseman
b. the second baseman d. the ﬁrst baseman

Which of the following is NOT a current major league baseball team?

a. Seattle Mariners c. Washington Capitals
b. Texas Rangers (1. San Diego Padres

16.

17.

18.

19.

20.

21.

22.

23.

24.

80
How many umpires work during a Major League regular season game?

a. four c. three
b. six (1. two

What kind of wood are bats made of?

a. pine c. oak
b. ash d. hickory

A good fast ball pitcher like Dwight Gooden can throw the ball at approximately
what speed?

a. 50 mph c. 95 mph
b. 80 mph (1. 110 mph

A ”knuckle ball” is a pitch that is thrown by holding the ball:

a. on the ﬁngertips c. in the palm
b. with the index ﬁnger and d. with the knuckles
thumb

The rubber of the pitcher’s mound is how far from home plate?

a. 60’6" c. 90’8"
b. 46’10" d. 72’2"

How many teams are there in the Major Leagues?

a. 14 c. 26
b. 22 d. 32

The oldest baseball park in which Major League games are still played is:

a. Tiger Stadium c. Chicago Wrigley Field
b. Boston Fenway Park d. Chicago Cominski Park

If the home plate umpire hinders the catcher’s attempt to throw the ball, it is:

a. tough luck c. an error
b. umpire interference d. obstruction

Home plate is:

a. in foul territory c. ﬁve sided
b. pointed toward the pitcher d. none of the above

81
25. The base path is deﬁned as:

a. the imaginary line 3 feet on either side of the direct line between the
bases

b. only the direct line between the bases

c. the space between the foul line and the direct line

(1. the distance between ﬁrst base and third base

82

General Knowledge Quiz

Please answer the following open-ended questions as best as you possibly can. Write
your answers in the spaces provided.

1. Who invented the light bulb?

 

2. How many continents are there?

 

3. What is the freezing point of water (F)?

 

4. In what state is Mt. Rushmore?

 

5. What is the capitol of Alabama?

 

6. Name the largest ocean

 

7. What oceans border Australia?

 

8. Who is the president of France?

 

9. How many great lakes border Michigan?

 

10. What year did the US enter WWII?

 

11. What does PLO stand for?

 

12. How many senators are there in congress

 

13. Who ran for president against Lyndon Johnson in 1964

 

APPENDIX G

83

Behavioral Incident Checklist
No Accountability Condition

SUBID

Employee Rating Study

Using the following checklist, place a mark by the items which you recall having seen
the secretary perform. It may help you to try to think about the speciﬁc context in
which the situation described took place (i.e. who was there, where did it take place,
etc).

Afteryouﬁnishﬁllingoutthisformplemeturnitoverandplaceitinfrontofyou.

Makes sure she takes accurate phone messages

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

................................................
................................................
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

84

Provides a student with detailed information about university hiring
procedures in order to answer their question

Tells a student that she is too busy to answer their question and that they
should ask one of the other secretaries

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

Does not have knowledge of all of the professors and the TA’s in the
department

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

Demonstrates detailed knowledge about the various idiosyncrasies of the
professors in the department

................................................
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

85

Refuses to do work for a professor that she feels was given to her on short
notice

She is friendly and courteous to professors when they ask her if they can use
the ofﬁce equipment

Does not let another secretary see conﬁdential information that she has been
trusted with

Volunteers to do some work for a professor that is not part of her regular
job description

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

Develops a method to keep track of all the assignments that she is working
on and when they are to be ﬁnished

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

86

Does very sloppy and inefﬁcient ﬁling

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

PLEASE TURN THE QUESTIONNAIRE OVER WHEN
. YOU ARE FINISHED

87

Behavioral Incident Checklist
Accountability Condition

SUBID

Employee Rating Study

Using the following checklist, place a mark by the items which you recall having seen
the secretary perform. It may help you to try to think about the speciﬁc context in
which the situation described took place (i.e. who was there, where did it take place,
etc).

After you ﬁnish ﬁlling out this form please turn it over and place it in front of you.
After everyone in the group is ﬁnished you will be asked to state your answers to the
rest of the group members and justify that you indeed saw the behaviors that you
marked.

Makes sure she takes accurate phone messages

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
................................................
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

................................................

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

88

Provides a student with detailed information about university hiring
procedures in order to answer their question

Tells a student that she is too busy to answer their question and that they
should ask one of the other secretaries

000000000000000000000000000000000000000000000000
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

Does not have knowledge of all of the professors and the TA’s in the
department

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

Demonstrates detailed knowledge about the various idiosyncrasies of the
professors in the department

Refuses to do work for a professor that she feels was given to her on short
notice

She is friendly and courteous to professors when they ask her if they can use
the ofﬁce equipment

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

Does not let another secretary see conﬁdential information that she has been
trusted with

Volunteers to do some work for a professor that is not part of her regular
job description

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

Develops a method to keep track of all the assignments that she is working
on and when they are to be ﬁnished

90

Does very sloppy and inefﬁcient ﬁling

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

PLEASE TURN THE QUESTIONNAIRE OVER WHEN
YOU ARE FINISHED