"INIWIWIWIWWWWHHIHHKHNIHWWI

   

MICHIGAN sure umvs

 

 

 

“r“ nIii/wiwflx/rgIg/MEiL/lliigii

Haas)

\

This is to certify that the

thesis entitled

CONSTRUCTIVE THINKING
AND
REACTIONS TO A DIFFICULT TASK

presented by

Janelle Cayo Ettema

has been accepted towards fulfillment
of the requirements for

.MastezLdegree in Psycholoqy

   
 

Major professor

Date jZ/ZO/ff

0-7539 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

LGIRARY
Mlchigan State
University

 

 

 

PLACE II RETURN BOX to roman thb chodtout from your "cord.
TO AVOID FINES Mum on or Moro dd. duo.

DATE DUE A DATE DUE DATE DUE

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU I. An Affirmative ActionEqunl Opportunity Intuition
mm:

c——_-—.____

CONSTRUCTIVE THINKING AND REACTIONS TO
A DIFFICULT TASK

By

Janelle Cayo Ettema

A THESIS

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

MASTER OF ARTS

Department of Psychology

1994

ABSTRACT
CONSTRUCTIVE THINKING AND REACTIONS To A DIFFICULT TASK
By

Janelle Cayo Ettema

To explore the behavioral correlates of the experiential system of Epstein's (1992)
Cognitive-Experiential Self Theory, 233 undergraduates completed measures of
Constructive Thinking (Epstein & Meier, 1989), emotional state, intelligence, and self-
esteem before a 40-minute task involving both easy and extremely difficult anagrams. Half
instructed to "do your best" on each anagram, the others received more relaxed
instructions. Emotional state was reassessed at posttask.

Constructive Thinking associated weakly and negatively (r = -.1 l, p < .10) with
nonproductive task persistence. Lower Constructive Thinking and self-esteem (median
splits) associated with longer (9 < .05) efforts to solve extremely difficult anagrams only
under the "do your best" instruction. Despite overlap of these Constructive Thinking and
self-esteem scores (1 = .60, p < .001), increments in posttask emotional discomfort
consistently associated more negatively with Constructive Thinking than with self-esteem.
The findings support the Constructive Thinking measure's utility as well as the validity of

Epstein's concept of experiential thinking.

TABLE OF CONTENTS

LIST OF TABLES .......................................................................................................... v
LIST OF FIGURES ........................................................................................................ vi
CHAPTER 1: INTRODUCTION ................................................................................... 1
Constructive Thinking .......................................................................................... 2
Persistence ........................................................................................................... 4
Self-esteem ..................................................................................................... 6
Intelligence ..................................................................................................... 8

Task difficulty cues ......................................................................................... 8

Need to please ................................................................................................ 9

Overview of the Study .......................................................................................... 9
Constructive thinking and persistence ........................................................... 10

The nature of constructive thinking: Exploration of the construct ................ 12
CHAPTER 2: METHOD ........................................................................... . .................. 15
Subjects ............................................................................................................. 15
Materials ............................................................................................................ 15
Procedure .......................................................................................................... 17
CHAPTER 3: RESULTS .......................................... . ................................................... 21
CHAPTER 4: DISCUSSION ........................................................................................ 30
Constructive Thinking and Persistence ............................................................... 30

The Nature of Constructive Thinking ................................................................. 33
APPENDIX A: Research Consent Form ....................................................................... 36

iii

APPENDIX B: Items of the Pre- and Posttask Questionnaires ...................................... 37

APPENDIX C: Anagrams Rated for Difliculty .............................................................. 38
APPENDIX D: Feedback Sheet .................................................................................... 39
LIST OF REFERENCES ............................................................................................... 40

iv

LIST OF TABLES

Table 1: Descriptive Statistics for 233 Students' Anagrams Task Performance
Measures ......................................................................................................... 19

Table 2: Descriptive Statistics for 233 Students' Scores on Principal Measures
and Related t-tests ........................................................................................... 22

Table 3: Correlations of Change in Emotion Ratings (Pretask Minus Posttask)
with the Global Constructive Thinking Scale and the Rosenberg
Self-Esteem Scale (N = 233) ........................................................................... 27

Table 4: Pearson Correlations Among Constructive Thinking, Self-esteem, and
Emotion Ratings (Pretask, Posttask, and Changes) for 233 Participants
(Decimals Omitted) ......................................................................................... 28

Table 5: Anagrams Rated for Difficulty ......................................................................... 38

LIST OF FIGURES

Figure 1: Interaction of Constructive Thinking scores with instruction
condition in relation to time spent on the first insoluble anagram. .................. 25

Figure 2: Interaction of Rosenberg self-esteem scores with instruction
condition in relation to time spent on the first insoluble anagram .................... 25

vi

CHAPTER 1

INTRODUCTION

Why do some intelligent people commonly behave so irrationally? If intelligence does
not predict success in life, as evidenced by previous research (F elsman & Vaillant, 1987;
Vaillant, 1977), what characteristics do? Questions like these led Epstein and Meier
(1989) to develop the Constructive Thinking Inventory (CTI), an instrument intended to
assess the degree to which the individual's habitual, preconscious thoughts are constructive
and facilitative of adaptive behavior.

The primary purpose of the present study was to explore the relationship between
constructive thinking and persistence. As speculated by Epstein ( 1992), one might expect
that constructive thinkers would be more likely to persist in the face of obstacles, whereas
those who are low in constructive thinking may give up more easily. A pilot study of this
hypothesis (Ettema, 1993) paradoxically found that college students who scored low on the
CTI's Constructive Thinking scale tended to persist longer on the practically insoluble items
of an anagrams task than did higher scorers. The present work addresses a post hoc
explanation of this result, that is, persistence in attempting to solve essentially insoluble
anagrams is irrational and nonproductive persistence. More constructive thinkers, it is
hypothesized, are better able to discriminate when persistence will or will not pay off, and
to invest their energies more appropriately. Less constructive thinkers, on the other hand,
tend to feel compelled to persist in situations in which success is unlikely, thereby foregoing
opportunities for more promising endeavors. A second purpose of this study was to more
broadly explore the nature of the construct of constructive thinking and its relationship to

measures of intelligence and self-esteem.

Constructive Thinking

In his undergraduate psychology class on emotions and self-concept, Epstein (Epstein
& Meier, 1989) was impressed by the degree to which some othenvise bright students led
their lives in an unintelligent and self-defeating manner. Rather than approaching their
problems analytically with a view toward behavioral experimentation, they often addressed
these problems in ways that tended to exacerbate, rather than alleviate, them. These
students often overreacted and overgeneralized from minor events, blaming themselves and
others, in spite of the fact that this habitual pattern typically made matters worse for them.
Epstein also found that this behavior, which appeared to be related to emotionally charged
past experiences, seemed extremely resistant to correction by rational argument. These
observations led him to believe that success in everyday living might be affected by the
individual's capacity to manage emotions.

The development and maintenance of nonproductive thought and behavior may be
analogous to animal studies of avoidance conditioning. In a study by Miller (1948), rats
learned to avoid moderate electric shock in the black side of a shuttle box by escaping to
the white side. After the electric current was turned off, their learned fear of the black side
of the box was highly resistant to extinction, apparently because the rats did not remain in
the now safe black section long enough to learn that it was safe. Perhaps in a similar
manner, humans also commonly appear to persist in their learned, but later dysfunctional,
behavior long after it has become inappropriate. What was once an adaptive response may
become so ingrained that one no longer considers whether this behavior is still helpfiil.
This would seem particularly likely when the original learning was traumatic or emotionally
charged. .

Epstein's (Epstein & Meier, 1989) cognitive-experiential self-theory (CEST) influenced
his conception of constructive thinking. CEST posits the existence of three semi-
independent systems, which together constitute the structure of the personality. In this

schema, the associationistic system operates at the unconscious level, analogous to Freud's

3

concept of primary process thinking. More pertinent to present purposes, however, are
Epstein's rational and experiential systems. The rational system operates on the basis of
logic and at the level of conscious awareness. Operating at the preconscious level, the
more holistic and loosely organized experiential system is developed on the basis of one's
past experience and has its own (non-rational) rules of inference. It is closely associated
with the emotions, and its inferences are experienced as self-evidently valid. It firnctions
more automatically than the rational system, and operates both more quickly and more
crudely.

From experiences with his students, Epstein deduced that the experiential system played
an important role in their problem-solving abilities. He developed the CTI to access this
system, tapping into the habitual thoughts that have resulted fiom the person's accumulated
past experiences, which may either help or hinder in dealing with life's problems and
stresses. Drawing on the work of cognitive theorists such as Beck (1976), Ellis (1962),
and Meichenbaum (1977), and on data regarding emotional incidents reported by his
students, Epstein and Meier (1989) developed items to describe the kinds of constructive
and counterproductive thoughts that people report having in everyday life. Taken together,
these are considered to be indicative of the individual's typical pattern of everyday thinking.

The CTI provides a composite global scale of Constructive Thinking and six factor-
based, semi-independent subscales. Supporting its construct validity (Cronbach & Meehl,
1955), the authors found that the scale correlated significantly with diverse criteria of
success in everyday living. These included success in the workplace, in social and intimate
relationships, and in mental and physical health. Furthermore, the CTI scales were found to
better predict such criteria than other measures of coping ability, apparently yielding a more
inclusive measure than these other scales provided (Epstein & Meier, 1989).

Epstein and Meier also attested that the concept of constructive thinking is
distinguishable from the concept of intelligence. This was supported by factor analytic

findings and also by the correlation of each construct with different criteria for success in

4

living. Among diverse criteria, the Constructive Thinking scale correlated with all except
academic achievement, whereas their IQ measure correlated m with academic
achievement. However, Epstein and Meier's (1989) study employed only the Shipley
(1946) Institute of Living Scale to estimate IQ scores. This measure was designed to
assess intellectual impairment, and may not be an adequate measure of the intelligence of
college students. The extent to which the CTI and standard measures of intelligence are
correlated has not yet been reported in the literature.

Subsequent research has confirmed anticipated links between counterproductive
thinking and emotional and behavioral symptoms (Katz & Epstein, 1991); between the
several CTI scales and rational beliefs as measured by Kassinove, Crisci, and Tiegerman’s
(1977) Idea Scale (Hurley, 1991); between low CTI scores and the tendency to
overgeneralize negative outcomes directed at the self (Epstein, 1992); and between
Constructive Thinking and the tendency to rate oneself above peers for acceptance of self
(Hurley, 1990).

Persistence

The importance of persistence and its societal value are evidenced in sayings and folk
wisdom: "If at first you don't succeed, try, try again." and "Winners never quit and quitters
never win. " Because of the importance of persistence and its applications in schools, at the
workplace, and in many other areas, this concept has interested psychologists from at least
the early years of this century (Feather, 1962). Persistence has been studied in relation to
many different variables, including self-esteem, intelligence, attributions for success and
failure, and attributional style.

Persistence has traditionally been viewed as a positive, adaptive phenomenon, and most
studies have taken this perspective in explorations of its personality correlates. Indeed,
persistence is often very desirable and valued in our society because it enables people to
accomplish things that they would never achieve if they gave up too readily. This view of

persistence is so ingrained in most of us that it is hard to conceive of any possible

5

disadvantages of persistence. Yet, despite the fact that we tend to think of persistence as
an unmitigated asset, clearly there are also situations in which it may be a liability.
Although giving up when persistence would have later paid ofl‘ may be maladaptive,
persistence at impossible tasks is also maladaptive, wasting time and personal resources.
JanofilBulman and Brickman (1984) argued that this second type of error may be even
more debilitating than the first, leaving pe0ple in destructive relationships, with
occupational bum-out, and with stress-related illness.

Brockner and Rubin (1985) explored nonproductive persistence in their book on
"entrapment," which they define as "a decision making process whereby individuals escalate
their commitment to a previously chosen, though failing, course of action in order to justify
or 'make good on' prior investments" (p. 5). According to these authors, such irrational
commitment to a course of action may have serious consequences in a wide variety of
situations at both the personal and the societal level. This process may result, for example,
in an inability or unwillingness to make productive career changes or to end unsatisfying
relationships. At the societal level, it may lead to the increasing unwillingness on the part
of both sides to negotiate a labor dispute, or even to the perpetuation of war when there is
little to be gained by a nation's continuing military involvement.

In addition to the conflict resulting from previous investments of time, effort, and/or
money described by Brockner and Rubin, the sheer momentum of behavior, as well as the
plethora of pro-persistence messages that we receive throughout our lives, often make
quitting very difficult (Janoff-Bulman & Brickman, 1982). Klinger (1975) has found that
some degree of depression--ranging from mild disappointment to severe depressive
episodes--is a normal part of disengagement from a task. A related problem is that people
commonly overestimate their ability to solve problems (F ischhoff & Slovic, 1980). So
discriminate persistence, or knowing when to quit and when to press on, involves skills that

may be more important than commonly recognized.

6

In evaluating the extent to which behavior is adaptive or maladaptive, it is neither
persistence, per se, nor the lack thereof, that is essential, but rather the ability to
discriminate those Situations in which persistence will pay off from those in which quitting
may be the better strategy (J anoff-Bulman & Brickman, 1982). A few studies illustrate the
importance of this skill in everyday living (Gurin, Gurin, Lao, & Beattie, 1969; Roistacher
1974). Gurin et al., for example, showed that black youths who distinguished between
areas of personal life that were under their control and elements in the general situation of
black Americans that were beyond their control were more likely than others to have an
active, realistic coping orientation both toward personal achievement and toward social
change.

Self-esteem.

Several studies have linked persistence with self-esteem. Orbach and Hadas ( 1982)
found that, after being subjected to a failure experience, persons who received self-esteem
enhancement in the form of positive feedback about their personalities were more persistent
on a subsequent task than those not receiving this treatment. Shrauger and Sorman (197 7)
also found that low self-esteem adversely affected persistence, but only if persons with low
self-esteem experienced consistent failure, and not if they were aware of some
improvement across trials. Similarly, Brockner and his colleagues found that highly self-
focused persons who were low in self-esteem performed particularly well on a task after
limited failure, but particularly poorly after extended failure, as compared to persons high in
self-esteem (Brockner et al., 1983).

In a study assessing persistence by the amount of time spent on each puzzle in an
anagrams task, Sandelands, Brockner, and Glynn (1984) found that persons higher in self-
esteem made better use of task information that suggested persistence would not pay ofi‘ in
some cases (the fact that some anagrams were insoluble) than did those low in self-esteem.

The former moved on from insoluble anagrams sooner when they had this information,

7

whereas those lower in self-esteem persisted equally long with and without this
information.

In contrast, McFarlin, Baumeister, and Blascovich (1984) found that persons m in
self-esteem made better use of failure feedback on a task involving Mednick's (1962)
Remote Associations Test (RAT) than did those higher in self-esteem, reducing their level
of persistence on each item, and actually performing better than those high in self-esteem,
who tended to spend an inordinate amount of time on earlier items. A related phase of
McFarlin et al.'s study found that persons high in self-esteem did not respond as well to
advice to go on to the next in a series of puzzles if they did not succeed in solving a puzzle
after two or three tries.

These contradictory conclusions as to the relationship of self-esteem to task
disengagement may be explained in terms of difi‘erences in self-presentational concerns
under conditions of receiving failure feedback or taking advice on the one hand, versus
receiving information as to the nature of the task on the other. It may be that under the
previous conditions, those high in self-esteem were more inclined to preserve their self-
esteem by continuing their efforts, since acknowledging failure or accepting advice to quit
might be viewed as reflecting adversely on them. Information about the solubility of the
problems would not reflect in the same way on their competence. Persons low in self-
esteem, on the other hand, might be more inclined to view failure feedback as a reflection
on their ability, making it pointless to persist very long, and to accept advice as a way of
protecting themselves against failure. In comparing their work with the McFarlin study,
Sandelands et a1. (1988) argued in a similar vein that while persons low in self-esteem may
be especially influenced by explicit external cues, those high in self-esteem may be more
likely to make use of relevant information suggesting the advantage of disengagement if
their active role in decision-making is maintained.

In a later work (McFarlin, 1985) more similar to that of Janofi‘-Bulman and Brickman

(1982), in which participants were given information rather than advice, McFarlin's results

8

were also similar: Persons high in self-esteem were more likely to use such information
appropriately than were those low in self-esteem. Nevertheless, the conclusions of
McFarlin, Baumeister and Blascovich (1984) may be valid under certain circumstances.
They concluded that while most researchers tend to equate high self-esteem with optimal
functioning, it may, in some situations, be the source of delusional over-confidence, causing
the person to ignore sometimes subtle cues as to the likelihood of success, or to disregard
legitimate advice not to persist too long.

Intelligence.

Exploring the transsituational consistency of behavior, Violato and Travis (1988) found
a statistically nonsignificant trend for lower IQ persons to be more stable in persistence
behavior than those with higher IQ scores, but suggested that the lack of significance may
be due to the small number of subjects in the IQ subgroups. Although inconclusive, this
trend may be of some interest in predicting the ability to persist or disengage as appropriate
to the situation.

Task difficulty cues.

While studying test anxiety and achievement motivation, Atkinson and Litwin (1960)
found that persons high in achievement motivation and low in test anxiety took more trials
in a ring-tossing task from the middle range of distance than did those higher in test anxiety
and lower in achievement motivation. This could be interpreted as a greater ability on the
part of high achievement motivation/ low test anxiety persons to discriminate where
persistence was most likely to pay off, since they concentrated their efforts at the middle
range, and took fewer shots from the very easy and the very difficult ranges where
persistence was either unnecessary for success or less productive. Similarly, Weiner,
Heckhausen, Meyer, and Cook (1972) found that task difficulty served as a cue to
determine the efficacy of extended effort, with the greatest expenditure of energy put forth
for tasks of intermediate difliculty. It may be that more constructive thinkers are more

attuned to task difficulty cues, and are more inclined to use even subtle cues to determine

9

the eficacy of effort expenditure. This ability is of particular interest in the present study,
since participants will be given no explicit information as to the practical insolubility of
some anagrams.

Need to please

It is also possible that eagerness to be a " good subject" and please the experimenter
may affect persistence behavior. This may be particularly applicable to persons lower in
constructive thinking, who have been found to be more concerned about evaluation when
working onstasks (Epstein, 1992) and more concerned about the impression they are
making on others (Katz & Epstein, 1991) than persons higher in constructive thinking. It
may be, then, that persons lower in constructive thinking persist longer at the insoluble
items, not because they cannot discriminate when persistence will and will not pay ofi‘, but
because they are more concerned about making a positive impression, and are therefore
more conscientious in following the directions to "do your best to solve each anagram."
This can be tested by varying the task instructions for one group of participants, de-
emphasizing the importance of each individual anagram and encouraging participants to go
on to the next item when experiencing difficulty with an anagram. If persons low in
constructive thinking are more conscientious than high constructive thinkers in following
directions, they should be less inclined to stick with the essentially insoluble anagrams
under this "go on" instruction condition than under the original instruction.
Overview of the Study

This study addresses two sets of questions, one involving the relationship of the global
Constructive Thinking scale to productive and nonproductive persistence, and the other
involving the nature of constructive thinking and its relationships to intelligence and self-

esteem.

10

Constructive thinking and persistence

The review of the literature concerning constructive thinking and persistence provides
the basis for a hypothesized relationship between these constructs. The post hoc
explanation suggested in my pilot study (Ettema, 1993 ), that high constructive thinkers are
less obsessively persistent, and better able to discriminate situations in which persistence
will pay off, seems reasonable in light of what we know about constructive thinking. If this
relationship holds, the concept of constructive thinking also may shed more light on
discriminate persistence and what constitutes the ability to know when to persist and when
to quit.

First, the ability to know when to quit and when to persist fits well with Epstein and
Meier's (1989) broad description of the characteristics of constructive thinkers, which is
based on the items in their scale. Constructive thinkers are characterized by a general
optimism, tempered by reality considerations. They are not prone to gross
overgeneralizations (e. g., "If I can't do this, I'm no good at anything"), nor to grandiose
self-enhancement which could lead to extreme overconfidence of their ability to perform
the task. They seem less likely to take things personally, and less sensitive to failure or
disapproval, giving them greater freedom to assess the profitability of disengaging fiom
tasks. Their greatest advantage may be flexibility, which allows them to take account of
situational cues and to change their expectations for success appropriately. This is
important since such expectations have been found to mediate persistence behavior
(Feather, 1982). High constructive thinkers generally feel good about themselves and
others, cope well with negative emotions, and behave effectively in daily life.

Secondly, differences between constructive and nonconstructive thinkers have been
confirmed by research, and seem consistent with differences in the ability to know when to
quit and when to persist. Persons low in constructive thinking tend to generalize negative
outcomes toward the self (Epstein, 1992; e.g., "If I fail at this, I will fail at other things

too. "). They are more concerned with evaluation when working on a task than are high

ll

constructive thinkers (Epstein, 1992), indicating that they may be less inclined to quit when
appropriate. Also, they are more likely to produce spontaneous negative thoughts during
tasks, to see stressors as more threatening, to judge themselves more harshly, and to be
more concerned about the impression they are making on others (Katz & Epstein, 1991).
All of these findings suggest that it may be especially dimcult for low constructive thinkers
to disengage from tasks, even when quitting is appropriate.

The theoretical advantage of people higher in constructive thinking in being able to
"solve problems in everyday living at a minimal cost in stress" (Epstein, 1992, p. 813) could
account for this greater ability to know when to persist and when to quit. Another possible
explanation to be tested is that people lower in constructive thinking may be more
conscientious in following directions in order to please the experimenter.

Individuals higher in constructive thinking may also be more sensitive to cues of task
difficulty and may have a keener intuitive sense as to whether or not an anagram is soluble.
In some problem-solving tasks, Zajonc (1980) found that intuitive or afl‘ective responses
may precede formal, cognitive judgments. Other studies have indicated the importance of
awareness of what one knows and the resulting awareness of where extended effort will or
will not pay off (Hart, 1965; Shaughnessy, 1979). Hart found that those who performed
best on a multiple-choice examination were also more accurate in their "feeling of
knowing," differentiating better between those items which they knew and those which they
did not know. High constructive thinkers may be better at making similar discriminations,
due to greater intuitive awareness, or possibly due to stronger verbal skills.

Other possible explanations are beyond the scope of this study, but may be worth
pursuing in the future. One particularly interesting possibility is that persons high and low
in constructive thinking may differ in the attributions they make for success and failure.
Such attributions may then affect persistence, with persons who make global, internal, and
stable attributions for failure (Abramson, Seligman, & Teasdale, 1978) being less inclined

to persist. Seligman, Nolen-Hoekstra, Thornton, and Thornton's (1990) study of the

12

efl‘ects of attributional style on swimming performance may be relevant here. Afier
receiving falsely poor reports of their times in different swimming events, the subsequent
performance of swimmers with a pessimistic style was negatively affected, whereas the
performance of those with an optimistic style was not. Perhaps in a similar manner, the
different attributions made by persons high or low in constructive thinking afi‘ect their level
of persistence.

The n_ature of constructive thinking: Exploration of the construct

Although Epstein (1989) demonstrated that the CTI and one IQ measure were
predictive of different types of success, it could still be that intelligence is a component of
the CTI, or that the two are somehow related, even if they do not measure exactly the same
thing. In light of the trend in the Violato and Travis study ( 1987) for persons with higher
IQ scores to vary more widely in their levels of persistence across situations, this
relationship bears further exploration.

Since self-esteem seems to be correlated with the ability to determine when to persist
and when to disengage from a task (Brockner, et al., 1983; Janoff-Bulman & Brickman,
1982; McFarlin, Baumeister, & Blascovich, 1984; Sandelands, Brockner, & Glynn, 1988),
and since it also seems plausible that self-esteem is intimately related to constructive
thinking, this relationship should also be explored further. It is interesting that while self-
esteem has failed to differentiate discriminate persisters fi'om nondiscriminate persisters in
the absence of explicit information about the efficacy of persistence in situations (e. g.,
McFarlin, 1985), data fi'om the pilot study (Ettema, 1993) suggest that CTI scores may do
so. This might indicate the ability of high constructive thinkers to pick up on more subtle
cues as to the likelihood that persistence will pay off.

Finally, the relationship of the Constructive Thinking scale to emotional reactions to a
frustrating task was also explored and compared to the relationship of self-esteem and
intelligence to such reactions. The anagrams task employed to measure persistence in this

study was anticipated to be rather fi'ustrating. It was further expected that constructive

l3

thinking would moderate negative reactions to the task, as high constructive thinkers would
be less inclined to overgeneralize this aversive experience, to blame themselves, and to be
sensitive to disapproval or failure than would those low in constructive thinking. Since
self-esteem and constructive thinking are likely related, a similar pattern was expected for
the relationship between self-esteem and emotional reactions. Intelligence was not
expected to correlate with emotional reactions to this task.
Based on the preceding considerations, these hypotheses were formulated:
Ifthe Constructive Thinking scale is measuring practical intelligence in relation to
persistence, Hypotheses I and H follow:
Hypothesis I : There is a positive relationship between constructive thinking and
productive persistence, as defined by the number of anagrams attempted and
solved, after controlling for intelligence.
Hypothesis II: Constructive thinking is negatively related to nonproductive
persistence, as defined by time spent on insoluble anagrams.
If low constructive thinking involves greater concern with making a positive impression,
Hypothesis III follows:
Hypothesis III: The proposed general negative relationship between constructive
thinking and nonproductive persistence will be attenuated if different instructions
suggest bypassing an anagram when experiencing difficulty, since low constructive
thinkers will follow this alternate instruction more conscientiously than high
constructive thinkers.
Ifconstructive thinking involves keener intuition as to which items are relatively easy and
which are practically impossible, Hypothesis IV follows:
Hypothesis IV: Persons higher in constructive thinking will be more accurate in their
intuitive ratings of the difficulty of anagrams than are those lower in constructive

thinking.

14

Further hypotheses regarding the nature of constructive thinking are:
Hypothesis V: Constructive thinking is positively related to both intelligence and self-
esteem.
Hypothesis VI: Constructive thinking is positively related to positive emotional

reactions to the task and negatively related to negative emotional reactions.

CHAPTER 2
METHOD

m

The participants were 233 undergraduates (152 women and 81 men) from the subject
pool of the Michigan State University (MSU) Psychology Department. Participation in
such studies was one option for the fulfillment of research requirements of students in
introductory psychology courses. Twenty-two subjects were eliminated from those
statistical analyses involving the Wonderlic Personnel Test (1989) due to a timing error
that invalidated their scores.
Materials

Each participant's informed consent was obtained using the form shown in Appendix A
(p. 36). Form A of the Wonderlic Personnel Test (Wonderlic, 1989) was administered to
determine each person's IQ score. The Wonderlic is a written 12-minute test of mental
abilities, originally developed for use in industrial settings. Its 50 items address diverse
areas of cognitive functioning, including general information, similarities and differences in
verbal and nonverbal material, vocabulary, visual-spatial tasks, reasoning, abstraction, and
calculations. Dodrill (1981) reported that Wonderlic total scores correlated 91-93 with
the Wechsler Adult Intelligence Scale (W AIS; Wechsler, 1955) and that Wonderlic IQ
scores were within 10 points of the WAIS Full Scale scores in 90% of the cases for a
sample of 120 normal adults. Reported correlations between the Wonderlic and the revised
version of the WAIS (WAIS-R; Wechsler, 1981) range from .75 to .96 (Dodrill, 1983;
Edinger, Shipley, Watkins, & Hammett, 1985; Frisch & Jossop, 1989; Hawkins, Faraone,
Pepple, & Seidman, 1990).

The 26-item Constructive Thinking scale from the Constructive Thinking Inventory
(Epstein & Meier, 1989) was used to assess this variable. Rosenberg's (1965) Self-Esteem
Scale (RSES) was used to measure self-esteem. This brief, convenient measure was

considered adequate for present purposes, although measures of global self-worth have
15

16

been found less successful than domain-specific measures in predicting behaviors (Stake,
1994). Rosenberg (1965) cited as evidence for construct validity that persons who score
higher and lower on this measure differed in expected directions on diverse other measures,
such as mood, neuroticism, group status, and so forth. Hagborg (1993) reported a
concurrent validity coefficient of .76 with the Global Self-worth Scale of the Self
Perception Profile for Adolescents (Harter, 1988). Silber and Tippett (1965) found a test-
retest correlation of .85 over two weeks. However, Byrne (1983) found a test-retest
correlation of only .62 over six months, suggesting that Silber and Tippett's higher
correlation was largely due to their much shorter time interval.

As is common in persistence research, an anagrams task was used to measure
persistence. Productive persistence was defined by the number of soluble anagrams
attempted and solved, partialing out intelligence to control for this probable moderator
variable. This persistence measure has face validity, in that high scorers are those whose
continued efforts were rewarded by successful solutions. Low scorers are presumably
those who either did not try very hard or who spent too much time and effort on insoluble
items.

Nonproductive persistence was defined by the amount of time spent on each of the
essentially insoluble anagrams. Continuing to work on an insoluble puzzle constitutes
evidence of nonproductive persistence, since the task is arranged so that persistence cannot
accomplish anything. This measure's internal consistency was assessed by comparing the
amount of time spent by each individual on each of the first two essentially insoluble
anagrams. This correlation was .73 ( r; = 231, p < .001, two-tailed test). Pre— and post-
task questionnaires (see Appendix B, p. 37) were used to gather information as to the
participants' gender, native language, attitude toward word puzzles, involvement in the
task, and mood. These items were created to assess likely mediator variables between
constructive thinking and persistence, and to determine the relationship between

constructive thinking and emotional reactions to the anagram-solving task.

17

The persistence task was generated using Micro Experimental Laboratory (MEL)
software, as was a brief task in which participants rated the perceived difficulty of ten 7-
letter anagrams. The CTI, the Rosenberg, and the pre- and posttask questionnaires were
also administered via MEL software. Instructions and stimuli for these phases of this study
were presented via computer, and are delineated in the following section. Participants
responded via keyboard. Data were recorded on discs.

Procedure

Students participated in groups of 5 to 23 in MSU computer laboratories. After
receiving each participant's informed consent, Form A of the Wonderlic Personnel Test was
administered first as a timed paper-and-pencil measure. The CTI, the Rosenberg, and the
pretask questionnaire were each administered in turn, followed by the anagram-solving
task. Finally, the anagram-rating task and posttask questionnaire were administered. All
measures were scored, cross-checked for accuracy, and the results recorded.

Participants were familiarized with the anagram-solving task through a brief
description of anagrams and the rules for solving them. They were given two sample
anagrams and their solutions, and then received the following instructions:

The anagrams you will be given vary in difficulty. It is VERY IMPORTANT that
you DO YOUR BEST TO SOLVE EACH ONE. If you cannot solve an anagram,
you can go on to the next one by pressing the F10 key.

You will NOT be able to return to previous anagrams once you go on to the next
one.

You may use pencil and paper to work out the solutions to the anagrams.

Please solve as many anagrams as you can until you are asked to stop.

Please do not communicate with other participants in this experiment.

If you have any questions, please raise your hand and the experimenter will assist
you. When you are sure that you understand the instructions, press the F1 key to

continue.

18

Some anagrams were so obscure as to be practically impossible to solve. The first two of
these were:

A C M O R R S U U MACRUROUS, third in the series;

 

A A D O P S U Y PADUASOY, fourth in the series.

 

The first 10 anagrams used for this task are listed in Table 1, p. 19. A total of 165
anagrams were available to keep participants engaged in the task so that those who tended
to stay with each anagram for longer periods of time would not be distracted by people
finishing early and leaving the room. Two to four insoluble anagrams were included in
each block of 10. Although participants were not informed of the time limit, a maximum
time of 40 minutes was set for the entire anagram-solving task, after which the computer
presented the next task. Due to a glitch in the MEL program, the maximum time actually
varied from 40 to 41.17 minutes.

The computer recorded the time spent on each anagram. In the case of the practically
insoluble anagrams, this provided a measure of nonproductive persistence. The computer
also recorded the number of anagrams attempted and solved correctly, which, after
partialing out intelligence, provided a measure of productive persistence.

To test whether persons lower in constructive thinking were more concerned about
evaluation and the impression they were making on others, and, therefore, more
conscientious in following the direction to "do your best to solve each anagram" than were
those high in constructive thinking, 96 participants were presented with a different

instruction, identical to the preceding except for the first paragraph :

The anagrams you will be given vary in difficulty. It is important that you do your
best to solve as many anagrams as you can. As soon as it becomes clear that you
cannot solve an anagram, go on to the next one by pressing the F10 key, so as to

attempt and solve as many as possible.

19

Table 1. Descriptive Statistics for 233 Students' Anagrams Task Performance
Measures

M. £82)
Total number of anagrams solved 7.43 (3.77)
Total seconds on anagrams task 2247 (405)

First 10 anagrams Time in seconds: Percentage of persons who
M (Q) solved each
1 AMPST (stamp) 19.1 (14.3) 95.7
2 YWRFEEA (freeway) 42.1 (42.2) 69.1
3 ACMORRSUU (macrurous) 107.5 (114.7) 0.0
4 AADOPUSY (paduasoy) 137.3 (151.5) 0.0
5 AEFKLNOSW (snowflake) 91.8 (117.3) 54.5
6 CHNORSTUU (cothurnus) 167.6 (163.3) 0.0
7 ACEENNRT (entrance) 144.3 (117.0) 4.7
8 AFFLOOTI‘ (flatfoot) 97.0 (81.3) 41.6
9 AERRSTUY (treasury) 136.8 (117.7) 7.3
10 ADEHIMRTY (diathermy) 139.5 (141.2) 0.0

Blocks of 10 anagrams

Time in seconds:

Number solved in each

M(§12) block: M(s_r))
1-10 1082 (647) 2.73 (1.06)
11-20 574 (354) 2.64 (1.69)
21-30 271 (250) .70 (1.06)
31-40 115(158) .41 (.85)
41-50 63 (100) .38 (.83)
51-60 40 (69) .20 (.54)
61-70 22 (51) .04 (.23)
All remaining items (71-165) 73 (173) .34 (.97)

20

prersons low in constructive thinking are more compliant or more conscientious in
following directions than those high in constructive thinking, it was expected that they
would be less inclined to persist on the practically insoluble items under this second "go on"
instruction than under the original instruction, and that the relationship between
Constructive Thinking scores and time spent on insoluble anagrams would be attenuated.

The main anagrams task was followed by a request that participants rate each of a
series of 10 new anagrams for perceived difliculty. In the latter series, all were seven
letters long and are listed in the order of presentation with the mean rating of each in
Appendix C, p. 38. Five were relatively easy to solve, and five were comparable in
dificulty to the "essentially insoluble" items of the main anagrams task. Each anagram was
exposed for 13 seconds before being rated as either "very difficult" or "less diflicult". The
number of seconds was selected after a pilot run of the experiment which determined the
minimal time necessary to solve the soluble anagrams in the set. To ensure that participants
would be unlikely to actually solve the anagrams before rating them, the interval was made
shorter than this minimal time. This provided a very stringent measure of the ability to
discern early on, perhaps in an intuitive manner (Hart, 1965; Shaughnessy, 1979; Zajonc,
1980), when persistence might and might not pay ofl‘.

The study ended with the posttask questionnaire. Emotion items from the pre- and
posttask questionnaires (Appendix B, p. 37) were scored on a Likert scale ranging from 1
(strongly disagree) to 5 (strongly agree). Scores for change in emotion were determined by
subtracting the pretask rating of each emotion fiom the posttask rating. Composite scores
for changes in emotion, either positive or negative, were calculated by aggregating the
relevant change scores. At the end of each session, participants were debriefed and given

information about the experiment as set forth in Appendix D (p. 39).

CHAPTER 4
RESULTS
Performance measures on the anagram-solving task appear in Table 1 (p. 19).

Participants averaged 38 minutes on the anagram-solving task, with most persons working
to the maximum time limit of approximately 40 minutes. Some persons finished early by
skipping through the anagrams, but 90% of the participants spent at least 27 minutes on
this task. Participants tended to spend most of their time on earlier items, with few
persons spending more than a few seconds on items beyond number 30. The mean number
of anagrams solved was 7.43, ranging fiom 1 to 22. The mean time spent on each of the
first ten anagrams and the percentage of persons solving each are listed in Table 1, along
with the mean time and average number of anagrams solved per each group of ten through
item 70. Descriptive statistics for the other central measures appear in Table 2 (p. 22).
The mean Wonderlic score of 25.5 is equivalent to a WAIS-R Full Scale IQ of 112, which
appears reasonable for a sample from a college population. On average, negative emotions
increased notably (p < .001) from pretask to posttask, while positive emotions decreased
(p < .001). Most effect sizes for these changes are large (Cohen, 1992), the two
exceptions being the changes in ratings of depression and anxiety, which were small and
medium, respectively (Table 2, p. 22).

Two-tailed tests were used to determine the statistical significance of all correlations.
The Constructive Thinking scale was unrelated to the Wonderlic Personnel Test
(W onderlic 1989) which estimates IQ (r = -.01), but did correlate substantially with the
Rosenberg (1965) Self-Esteem Scale (1 = .60; p < .001, two-tailed test). The CTI's global
scale did not correlate meaningquy with the number of anagrams solved, neither before
partialing out intelligence (r = .01) nor after (1 = -.02). The correlation of Constructive
Thinking with time spent on solved anagrams was also not statistically significant (1 = .02).

A search of all present variables which might be conceivably related to the number of

21

22

Table 2. Descriptive Statistics for 233 Students' Scores on Principal Meflrres and Related
t-teStS

Pretask Posttask Pre- to Posttask Changesc Effect size
M (S_D) M (.82) M (S2) t-value of
changes

Constructive Thinking 93.8 (11.8)

Rosenberg Self-Esteem Scale 32.7 (4.4)

Wonderlica 25.5 (4.8)

Emotion measures b
Ifeel upset. 1.96 (.99) 2.61 (1.12) .66 (1.11) 9.05 .51
Ifeel angry. 1.89 (.96) 2.65 (1.24) .76 (1.22) 9.52 .53
I feel frustrated. 2.18 (1.02) 3.70 (1.18) 1.52 (1.36) 17.06 .75
Ifeel depressed. 2.14 (.99) 2.41 (1.00) .27 (1.04) 3.92 .25
I feel anxious. 2.97 (1.08) 3.47 (1.11) .50 (1.21) 6.28 .38
Total negative 11.15 (3.73) 14.84 (4.17) 3.70 (3.66) 15.40 .71
I feel relaxed 3.95 (.79) 2.59 (1.05) -1.36 (1.13) -18.46 .77
I feel happy 3.67 (.84) 2.80 (1.03) -.87 (1.05) -12.63 .64
I feel confident 3.82 (.78) 2.85 (.92) -.98 (.94) -15.97 .72
I enjoy word puzzles 3.17 (.99) 2.33 (1.20) -.84 (1.01) -12.62 .64
Total positive 14.62 (2.42) 10.58 (3.31) -4.04 (2.93) -21.05 .81

lam good at solving word 3.05 (.80) 2.14 (1.05) -.91 (.94) -14.84

puzzles.b

I did my best to solve the 3.86 (1.06)

anagrams

It was important to me to do 3.14 (.98)

well.

 

a Wonderlic N = 211; Wonderlic score of 25.5 is equivalent to a WAIS-R FSIQ of 112.
b 1 = Strongly disagree; 2 = Disagree; 3 = Neutral; 4 = Agree; 5 = Strongly agree.
C p < .001 for t 2 3.29.

m: Effect sizes were calculated from t values using Rosenow and Rosenthal's (1988) formula, p. 206.

23

anagrams solved yielded only one statistically significant correlation with the participants'
pretask rating of their word puzzle ability (r = .17; p < .01). The same rating also had a
nearly significant association with posttask ratings of how important it was to do well on
the task (1 = .13, p = .051), but was unrelated to the time spent on any anagram, except for
the first (_r_ = -. 18; p < .01), or to the rating of the extent to which participants tried their
best (I = -.04; ns). Solving the fourth anagram, which immediately followed the first two
insoluble ones, was not significantly correlated with measures of constructive thinking, self-
esteem, or intelligence (all ['5 < .06). The correlation between Constructive Thinking
scores and the intuitive rating of anagrams for difficulty was not significant (r = -.06).

Accepting Cohen's ( 1990) suggestion that the p < .10 level is acceptable for exploratory
studies, the CTI's Constructive Thinking scores did correlate significantly in the anticipated
negative direction with time spent on the first insoluble anagram (r = -.11; p < .10, two-
tailed test), whereas neither self-esteem (r = -.04) nor intelligence (I = .00) did. However,
when one notably outlying score associated with the most extreme length of time spent on
this anagram (8.7 standard deviations above the mean) was removed from the analysis, this
correlation was no longer significant (1 = -.06). Logarithmic transformation of the
positively skewed distribution of times, as recommended by Tabachnik and Fidell (1989),
slightly increased this value, but not to the point of statistical significance (1 = -. 10). Since
the logarithmic transformation used in other calculations yielded the same result (slight
increases, but none to the point of statistical significance), only correlations performed
without the transformation are reported. Considering only the group of persons who
received the original instruction to "do your best to solve each anagram," no significant
relationship between Constructive Thinking scores and time spent on the first insoluble
anagram was found (r = -.09; ns). The correlation of the CTI'S global scale with the second
insoluble anagram was not significant (r = -.07). Neither the time spent per anagram nor
solution rates (whether or not each was solved) for the first two soluble anagrams

correlated significantly with the Constructive Thinking scale (all [S < .05).

24

The groups receiving different instructions did not differ significantly in their
correlations between Constructive Thinking scores and time spent on the first essentially
insoluble anagram. The correlation for the subgroup of 96 persons receiving the "go on"
instruction (1 = -. 16; ns) was only slightly higher than that of the 137 persons receiving the
original instruction (r = -. 10; ns). Removing the most extreme score for time spent on the
essentially insoluble anagram reversed this pattern, yielding a slightly lower correlation for
those receiving the "go on" instruction (r = -.03, ns) as compared to the group receiving
the original instruction (1 = -. 10, ns). The difference between these correlations was not
statistically significant.

However, although there was only a weak main effect for Constructive Thinking scores
in relation to time spent on the first insoluble anagram (r = -.11; p < .10), there was a
statistically significant main effect for instruction condition on time spent on this puzzle
(_1; = -. 15, p <05), and also a significant interaction between Constructive Thinking and
instruction condition for time spent on the insoluble anagram, E (2, 229) = 3.59, p < .05
when the outlying score for time spent on the insoluble anagram was removed. After
splitting this sample's Constructive Thinking scores at the median, the instruction was
found to have a greater effect on persons low in constructive thinking than on persons high
in constructive thinking as depicted in Figure 1 (p. 25), as well as a similar impact on
persons low in self-esteem as compared to those high in self-esteem E (2, 229) = 3.54,

p < .05, again splitting scores at the sample's median (Figure 2, p.25). Neither interaction
nor the main effect for instruction condition was statistically significant when the outlying
score for time spent on the first insoluble anagram was included.

Beginning with a multiple regression equation in which self-esteem scores and
instruction condition are previously included as independent variables in relation to time
spent on the first insoluble anagram (R = .17), the addition of Constructive Thinking scores
adds little to the predictive power of the equation (R = .18). Similarly, adding the self-

esteem measure to the multiple regression of Constructive Thinking and instruction

25
160 T
140 A

120 d-

100 --
moantineon
firstinsoluble '0 ._

 

 

 

 

 

 

anagrln
seconds
‘ ) so . ., + High 611
+ Low CTI
“ d D
20 s '-
“do best" '90 on'

instruction condition

Figu_re 1. Interaction of Constructive Thinking scores with instruction condition in relation to time spent on the
first insoluble anagram.

160 --
14o --

120 ‘L

Meantime 100 J.

 

 

 

 

 

 

on first
insoluble so «-
anagram

(monds) 60 ‘1' + High SE

+ Low SE
40 4-
20 ..

0 i 4
”do best“ '90 on"
instruction condition

Eiguie_2. Interaction of Rosenberg self-esteem scores with instruction condition in relation to time spent on the
first insoluble anagram.

26

condition onto time spent on the first insoluble anagram (R = .17) adds very little (R =
.18). These results were obtained without the outlier value for time spent on the insoluble
anagram. When the outlier was included, the predictive power of the equation with only
the self-esteem and instruction condition variables (R = .09) was improved somewhat by
the addition of Constructive Thinking scores (R = .15), but the equation including only
Constructive Thinking and the instruction condition @ = .15) was not improved by the
addition of the self-esteem measure (R = .15).

Table 3 (p. 27) shows all correlations of the Constructive Thinking and self-esteem
measures with scores for pre- and posttask ratings of emotion and the related change
scores for emotion, and the results of t—tests for the difference in the strength of the
correlations of each of the emotion measures with either the global CTI scale or the
Rosenberg. These t—tests indicate in what instances the strength of the relationship of
constructive thinking with change in emotion is significantly different from the strength of
the relationship of self-esteem with change in emotion. Bonferroni adjustments seemed
unnecessary since 7 of the 11 tests (64%) for differences in the strength of the correlations
are statistically significant if two at the p < .10 level are included, a percentage far greater
than the 10% that can be expected on the basis of chance alone.

Table 4 (p. 28) shows all intercorrelations of emotion measures and the correlations of
emotion measures with the Constructive Thinking scale and the Rosenberg (1965) Self-
Esteem Scale. The correlations of the Constructive Thinking scale and the Rosenberg with
emotion measures generally differed more at posttask than at pretask, as can be seen by
comparing the first two sets of columns in Table 4. This is also illustrated by the greater
difference in correlations of the Constructive Thinking scale and the Rosenberg with the
two overall m_sttask emotion measures as compared to the lesser difference in correlations

of each with the two overall pLetask emotion measures. Correlations of the Constructive

27

Table 3. Correlations of Change in Emotion Ratings (Pretty; Minus Posttask) with the
Global Constructive Thinking Scale and the Rosenberg Self-Esteem sme (N = 233

Constructive Thinking Rosenberg Self-esteem t-test of
difference

Upset -.05 .07 2.61**
Anger -. 17* .00 2.94"
Depression .05 . 16* 1.90T
Frustration .11 .17* 1.04
Anxiety .05 .08 n.s.
Total negative .00 .16* 2.76**
Happiness .20** .04 2.79“
Relaxation . 14* .04 1,72T
Confidence -.03 -.07 n.s.
Enjoyment . 14* .09 1.53
Total positive . 16* .04 2.07*

 

* p < .05, two-tailed test. ** p < .01, two-tailed test. T p < .10, two-tailed test.

Note. Differences between these correlations, which share a common variable, were
computed using McNemar's (1963) formula, p. 158.

28

 

.3. 8.3. v38 .3. NTS. vase .NN. ~38. v30".
K. Nw 3 £- mw ew co en on we 3. we NT wm- eN- o7 Nm- 31 eN- wm- mm- NV. em- ww- 3- mm. mm- we on we 2 av EV 5:25 .mom
mm- wm- ww- w.m- om- :1 em- ev- w? :- mm- wN- av we «m we we ew 3 8 en E. mw bw :- mw nw 2 N- ew- ee ev- Nm- 8:25 .wuZ
- wv me Nm wm mm on N we 3 3-313- mNme- :1 2-91:1 cm. he. ev- wN- hm- wm- we N I” E ev mm serge—om
wv me Nm - 3.“ cm wv mm 3 mm 2. hm- E- 2. ON. em- wT em- Nv- :1 Vm- wm- 0N- mm- Si we Nm cm eN mv em mmofigam
wm mm on em on we - N on NN me- MN- e7 w_- N- ew- m_- em- 91 N- NV- mm- mm- 3. mv- he. Nv wm me- Nv Vm 352-280
AN 9. 3 mm 9 mm MN on NN - ve-eN-E- Ne 2- we. 2- N. N. wT wm- N. EN. 3”- em- we 3 oe 3 N eN Eofizomcm
3-313- 2-5m- 3- me- MN- 2. we- eN- E- - ev hm .m we N _N 2 am w. 9 mm «N E 8-:- me N. In- house
N. on- :1 Z- N- em- w_- N- ew- Ne-mT we- ev hm 2 - ee 2 Q ev ev 5m 3“ mm we 2 NT em- Z N. :1 :ozamem
2-91:1 wT em- NV- mT em. 3- :- N- N. we N _N ee 2 Q - eN wm Q .N E em 3 em- ov- me ev- mv- 5:38qu
em. :1 eq- 21 3. wm- eN- Nv- mm- w? wm- N. m. on w. ev 0v hm eN Nm mv - S E. 2. ee N. .m- C- N.»- mm- Bw=<
wN-hm-wm- 0N- mm- 31 mm- .m- 3- VN-Vm- em- 2 mm N 2 mm we «N G Vm S vs on - no N. ev- me- Nv- NV- 6mm:
we AN 3.. we Nm E“ be. Nv wm oe 2 me C no-2- 2 N73. 2 em- ev- ee "N. F.- N-e oN-ev. - ee mmmm
E ev mm eN Q on me- Nv em E N eN me N. .m- 2 MN- :1 me ev- mw. C- NV- mm- we- Nv- Nv- ee - Eb
:ocaxsom mmoEn—aaz 853280 EoEAoEm 52x5 confide-Em :oqmmoaofl 892 632 mmmm Eb

 

 

32:50 2388 EASE-8 m8

 

new we SEQ Ea xmmzmom c.2305 m need 5525 were Sconce—elm? 2355. 03883480 Mdofi< Scale—880 :08me .v 055-

29

Thinking scale and the Rosenberg were quite similar at pretask for both negative
emotions (r = -.52 and g = -.46, respectively; I = .95, ns) and positive emotions (r = .47
and r = .45, respectively; 1 = .34, ns). However, the two measures differed significantly at
posttask for both negative emotions (r = -.46 and r = -.27, respectivel)’; 1 = 4.21, p < .001)
and positive emotions (r = .49 and r = .36, respectivel)’; I = 2.41, p < .05). The
Wonderlic (1989) did not correlate significantly with any of the emotion scores. No
significant correlations were found between scores of change in emotion with performance

(anagrams solved) or with time spent on the anagram-solving task.

CHAPTER 4

DISCUSSION

Constructive Thinlgi_11g_and Persistence

The lack of a positive relationship between constructive thinking and productive
persistence was surprising, since it is contrary to Epstein's theory of how constructive
thinking and persistence ought to be related (Epstein, 1992). It may be, however, that the
task was not engaging enough to make persistence on any anagram seem productive to
participants, since on average they reported being "neutral" on the importance of doing
well on the task. Perhaps a monetary incentive for solving anagrams would increase their
motivation and clarify this finding. Alternatively, there may be better measures of
productive persistence than the number of anagrams solved. Perhaps a measure with
clearer external validity, such as earning a college degree (controlling for intelligence),
would be significantly associated with constructive thinking.

The fact that participants' pretask rating of their word puzzle ability was significantly
and positively associated with the number of solved anagrams appears to reflect the
accuracy of their self-assessment, rather than a link with effort. Although it was nearly
significantly related to ratings of the importance of the task, this ability rating was not
related to the rating of doing one's best to solve the anagrams, nor to the time spent on any
anagram but the first. The negative relationship with time spent on the first anagram
supported the accuracy of the ability ratings, since those who rated their ability higher took
less time to solve this relatively easy puzzle than those who rated their word puzzle ability
lower.

The lack of relationship between self-esteem and persistence was also surprising, given
the prior reports of such a relationship (Brockner et al., 1983; Orbach & Hadas, 1982;
Shrauger & Sorman, 1997). In contrast to those works, however, the present investigation
did not provide performance feedback, which may be a necessary condition for self-esteem
to register influence on persistence. Without this feedback, participants may be less self-

30

31

focused and less aware of their "failure" on the task, and may more easily attribute
difficulties to the nature of the task rather than themselves.

The first insoluble anagram appeared to best assess nonproductive persistence because
all participants worked on this anagram without frustration from a previous insoluble
anagram, whereas subsequent anagrams were not addressed by some participants due to
the 40-minute time limit for this task. The absence of a significant correlation of the
Constructive Thinking scale with either solutions or time spent on the first two (soluble)
anagrams suggests that experience with these two anagrams did not create a confound.

On the basis of time spent on the first insoluble anagram, constructive thinking
appeared weakly associated with avoiding nonproductive persistence, whereas self-esteem
and intelligence appeared unrelated to nonproductive persistence. However, this
relationship relied heavily on the very extreme outlier, and even with this score, the
association was much weaker than that expected based on the results of my pilot study
(Ettema, 1993).

One possible reason for this is that the nature of the task may have been changed too
drastically in efforts to ensure sufficient time commitment on the part of participants. Pilot
study participants knew at the outset that there were only five anagrams in the task. The
task was changed to a larger and unspecified number of anagrams because of an apparent
temptation on the part of some pilot study participants to rush through the five anagrams,
without really attempting to solve any, leading to a loss of usable data. Among those who
made at least a reasonable attempt to solve the puzzles, however, the time spent on the first
insoluble anagram varied widely, ranging up to 2012 seconds (33.5 minutes), as opposed to
the present study's ceiling of 1321 seconds (22 minutes). The tendency of some persons to
persist for extremely long periods of time in order to prove their competence may have
been more exaggerated when they knew that there were only five anagrams, as failure on
any one anagram would reduce their potential overall success rate by 20%. The relative

restriction of range in the time spent on the insoluble anagram in the present study may be

32

responsible in part for the attenuation. Another explanation for the smaller correlation may
lie in the larger groups of participants used for the present study. Perhaps those with a
need to prove their competence by solving the anagrams were less inclined to do so in
larger, more anonymous groups, since their own failure to solve an anagram would seem
less noticeable to the experimenter. Finally, it may also be that the pilot study, involving
only nine participants, caught random fluctuations in a way that misleadingly suggested a
relationship.

The interaction between Constructive Thinking scores and instructions (Figure 1, p. 25)
suggested that persons lower in constructive thinking may be more inclined to follow
directions carefully than persons higher in constructive thinking , perhaps in efforts be
"good subjects" and to please the experimenter. This is consistent with previous findings
which indicated that persons low in constructive thinking were more concerned about
evaluation when working on a task (Epstein, 1992) and more concerned about the
impression they make on others (Katz & Epstein, 1991) than persons high in constructive
thinking. The similar interaction found between the instruction condition and self-esteem
(Figure 2, p. 25) suggests that, as found in the study by McFarlin and colleagues (1984),
persons higher in self-esteem may be less inclined than low scorers to follow "advice," here
in the form of instructions. Since the Constructive Thinking scale is strongly related to
self-esteem (r = .60), this explanation can apparently be extended to high constructive
thinkers, as they also were virtually unaffected by the difference in instructions. We may
speculate that, had the manipulation involved information as to the insolubility of some
items, as in the studies by Sandelands et al. (1984) and McFarlin (1985), it would have had
a greater impact on those higher in self-esteem and constructive thinking, since they would
have been more able to maintain their sense of personal competence and control in using
this information, rather than submitting to being told what to do. The present findings
support the view that high constructive thinkers are more independent in their thinking and

behavior than those lower in constructive thinking.

33

The addition of either Constructive Thinking or self-esteem scores to a multiple
regression equation, which already included the other measure and the instruction condition
in relation to time spent on the first insoluble anagram, had a negligible efi‘ect in each case,
particularly when the outlier value for time spent on the first insoluble anagram was
excluded fi'om the analysis. This suggests that the interactions of both the Constructive
Thinking scale and the Rosenberg with the instruction condition involve the overlapping
aspects of each measure reflected in their correlation (r = .60).

Since Constructive Thinking scores did not correlate significantly with scores for
accuracy in rating anagrams for difficulty, whatever led those higher in constructive
thinking to be slightly more inclined to disengage fiom the insoluble anagram sooner than
those lower in constructive thinking, it was apparently not a keener intuitive knowledge of
the difliculty of the anagrams, at least as measured in the present study. It may be,
however, that the measure of this intuitive knowledge was too stringent, and that exposure
of the anagrams for more than 13 seconds would have yielded a different result. The
problem with greater exposure times would be distinguishing correct intuitive ratings from
anagrams solved by some participants.

While this study offered no support for the hypothesized positive relationship between
constructive thinking and productive persistence, it did yield some very limited support for
the negative relationship between constructive thinking and nonproductive persistence.
Future studies with improved measures of the persistence constructs may provide more
conclusive evidence for these relationships. The results also suggest that constructive
thinking may involve greater independence and relative freedom from the need to please
others.

The Nature of Constructive Thinking

Constructive Thinking was unrelated to estimated IQ scores. This finding supports the

conclusion of Epstein and Meier (1989) that the Constructive Thinking Inventory, which is

intended to measure practical intelligence, is not a measure of academic intelligence.

34

It is not surprising that constructive thinking correlated significantly with self-esteem,
since the same aflect-laden experiences which are thought to determine constructive
thinking (Epstein & Meier, 1989), are likely to affect self-esteem as well. When the two
measures are entered in a multiple regression equation, however, they maintain sufficient
tolerance to suggest that the overlap is not so great as to render one or the other
meaningless in the joint prediction of dependent variables (Tolerance = l - r2 = 1 - .36 =
.64; Tabachnik & Fidell, 1989).

The findings regarding emotional reactions support the hypothesis that constructive
thinking is associated positively with positive emotional reactions to the task and inversely
with negative emotional reactions. This task was apparently quite aversive for most
participants, admittedly more aversive than intended, as most reported increased negative
and decreased positive emotions. The Constructive Thinking scale was associated with a
greater number of measures for change in emotions than was either the Rosenberg or the
Wonderlic. It is not surprising that the Wonderlic, used here to estimate IQ scores, did not
correlate with any of the measures of change in emotion. But interestingly, while both the
Rosenberg and the Constructive Thinking scale correlated significantly with some measures
of change in emotion, only the Constructive Thinking scale was related in the manner
expected for a measure which involves the management of emotion. Considering all
significant correlations of Constructive Thinking scores with change in emotion (Table 3, p.
27), higher Constructive Thinking scores were associated with lower change scores for
anger and higher change scores for happiness, relaxation, enjoyment of word puzzles, and
overall positive emotions. Higher self-esteem scores, on the other hand, were associated
with greater increases in depression, frustration, and overall negative emotion, comprising a
pattern which is opposite to that which would be expected of a measure which captures the
ability to manage emotional reactions. The fact that the correlations of the overall positive
and negative emotion ratings with the Constructive Thinking scale and the Rosenberg

differed more at posttask than at pretask suggests that the differences in the correlations of

35

these measures with change in emotions reflect differences in reaction to the task, rather
than differences in baseline levels of emotion.

Epstein and Meier (1989) have cited ample evidence that the ability to effectively
manage emotions affects thinking and behavior in ways that can influence success in
everyday living. The interaction of Constructive Thinking scores with differences in
instructions indicates that this measure may involve different approaches to problem-
solving and different reactions to following advice or directions when working on a
problem, as does self-esteem. While constructive thinking and self—esteem are both
theoretically and statistically related, constructive thinking appears to involve the
management of emotions as stipulated by Epstein's CEST theory (Epstein & Meier, 1989)
whereas self—esteem apparently does not. This, along with the extended evidence for the
Constructive Thinking scale's independence from academic intelligence, fiirther supports
this measure's potential value for predicting success in everyday living in areas in which

standard measures of intelligence fail to do so.

APPENDICES

~

36
APPENDIX A

RESEARCH CONSENT FORM

I have freely consented to take part in a scientific study
being conducted by: Janelle Cayo Ettema
Under the supervision of: John R. Hurley, Professor, Department of Psychology.
This experiment involves solving some problems and puzzles and answering questions
about thoughts and feelings.
Participation in this experiment usually takes no more than 1.5 hours.

The study has been explained to me and I understand the explanation that has been
given and what my participation will involve.

I understand that I am free to discontinue my participation in the study at any time
without penalty.

. I understand that I will receive three (3) research credits in my psychology course for

my participation in this experiment. I also understand that if I prefer not to participate in
this or any other research project, other options are available for fulfilling this
requirement, and that the instructor of my course will inform me as to the details of such
options.

. I understand that the data will be coded to ensure my confidentiality and that no

identifying information will be disclosed. Within these restrictions, results of the study
will be made available to me at my request.

. I understand that my participation in the study does not guarantee any beneficial results

tO me.

. I understand that, at my request, I can receive additional explanation of the study afier

my participation is completed by contacting
Janelle Ettema, (517) 548-7429.

You indicate your voluntary agreement to participate by completing and returning this
form.

Signed:

 

Date:

 

3 7
APPENDIX B

ITEMS OF THE PRE- AND POSTTASK QUESTIONNAIRES

These 2 items were presented only in the Pretask Questionnaire:
What is your gender? male female
Is English your first language? yes no

The following items were presented in both the Pretask and Posttask Questionnaires and
were rated by participants and scored as follows:

a. Strongly agree. (5 points)
b. Agree.(4 points)

c. Neutral. (3 points)

d. Disagree (2 points)

e. Strongly disagree. (1 point)

I feel relaxed.

I feel upset.

I feel anxious.

I feel depressed.

I feel frustrated.

I feel confident.

I feel angry.

I feel happy.

I enjoy word puzzles.

I am good at solving word puzzles.

These 2 items were presented only in the Posttask Questionnaire, and were rated on the
same 5-point Likert scale described above:

I did my best to solve all the anagrams.

It was important to me to do well.

APPENDIX C

Table 5. Anagrams Rated for Difficulty

Anagram

EINPRRT
AABHLTY
ABFIILR
FIMNORU
BIOOSUV
ADJLMOR
CEPSSTU
EFFFINO
ACFLRUU

AEEGMSS

Solution

printer
bathyal
bifilar
uniform
obvious
jarldom
suspect
infeoff
fiircula

message

38

AN AGRAMS RATED FOR DIFFICULTY

Mean rating ( SD )_3

1.55 (.50)
1.72 (.45)
1.76 (.43)
1.76 (.43)
1.79 (.41)
1.80 (.40)
1.64 (.48)
1.72 (.45)
1.83 (.37)

1.57 (.50)

 

a 1 = relatively easy; 2 = very difficult

39
APPENDIX D

FEEDBACK SHEET

The purpose of this experiment is to look at persistence behavior as related to a
measure called the Constructive Thinking Inventory (Epstein & Meier, 1989). The CTI
contains items that describe constructive and counterproductive automatic thoughts that
people report having in everyday life, and is considered to reflect patterns of thinking
which are linked to one's past emotional experiences. The experiment also involves
measures of self-esteem and general cognitive ability, and explores how these constructs
might be related to constructive thinking.

All of the anagrams in this experiment were actual words in the English language, but
some of them were so obscure as to make them practically impossible to solve. The
experiment explores how people who differ in "constructive thinking" might also differ in
persistence behavior -- on both the soluble anagrams and those that are essentially
insoluble. It is hypothesized that those higher in constructive thinking will persist longer
on the soluble anagrams than those lower in constructive thinking, but not as long on the
practically insoluble ones.

The experiment also explores the nature of constructive thinking as it may be related to
intelligence, self-esteem, and emotional reactions to this rather fiustrating task.

Ifyou have any questions or comments about the experiment, feel free to call the

experimenter, Janelle Ettema at 1-548-7429.

LIST OF REFERENCES

LIST OF REFERENCES

Abramson, L. Y., Seligman, M. E. P., & Teasdale, J. D. (1978). Learned helplessness
in humans: Critique and reformulation. Journal of Abnormal Psychology, 87, 49-74.

Atkinson, J. W., & Litwin, G. H. (1960). Achievement motive and test anxiety
conceived as motive to approach and motive to avoid failure. Journal of Abnormal and
Social Psychology, 60, 52-63.

Beck, A. T. (1976). Cogaitive therapy and the emotional disorders. New York:
International University Press.

Brockner, J., Gardner, M., Bierman, J ., Mahan, T. Thomas, B., Weiss, W., Winters,
L., & Mitchell, A. (1983). The roles of self-esteem and self-consciousness in the
Wortman-Brehm Model of Reactance and Learned Helplessness. Journal of PersonLity
and Social Psychology, 45, l99-209.

Brockner, J. & Rubin, J. Z. (1985). Entrapment in escalating conflicts: A social
psychological analysis. New York: Springer-Verlag.

 

Byrne, B. M. (1983). Investigating measures of self-concept. Measurement and
Evaluation in Guidanceal 6, 1 15-126.

Cohen, J. (1990). Things I have learned (so far). American Psychologist, 45, 1305-
1312.

Cohen, J. (1992). A power primer. Psychological Bulletin. 112, 155-159.

Cronbach, L. J ., & Meehl, P. E. (1955). Construct validity in psychological tests.
Psychological Bulletin. 52, 281-302.

Dodrill, C. B. (1981). An economic method for the evaluation of general intelligence
in adults. Journal of Consultingand Clinical Psychology, 49, 668-673.

Dodrill, C. B. (1983). Long term reliability of the Wonderlic Personnel Test. Journal
of Consultingand Clinical Psychology, 51, 316-317.

 

4O

41

Edinger, J. D., Shipley, R. H., Watkins, C. E., & Hammett, E. B. (1985). Validity of
the Wonderlic Personnel Test as a brief IQ measure in psychiatric patients. Journal of
Consultingand Clinical Psychology, 53, 937-939.

Ellis, A. (1962). RflscL and emotion in psychotherapy New York: Lyle Stewart.
Ettema, J. C. (1993). Constructive thinking and persistence: An application of the
constructive thinking inventogy. Unpublished manuscript, Michigan State University, East

Lansing.

Epstein, S. (1992). C0ping ability, negative self-evaluation, and overgeneralization:
Experiment and theory. JoumJaI of Personalig and Social Psychology, 62, 826-836.

Epstein, S., & Katz, L. (1992). Coping ability, stress, productive load, and
symptoms. Journal of Personality and Social Psychology, 62, 813-825.

 

Epstein, S., & Meier, P. (1989). Constructive thinking: A broad coping variable with
specific components. Journal of Personality and Sociafisychology, 57, 332-350.

Feather, N. T. (1962). The study of persistence. Psychological Bulletin, 59, 94-115.
Feather, N. T. (1982). Actions in relation to expected consequences: An overview of

a research program. In N. T. Feather (Ed), Expectations and actions: Expectancy vaLue
models inpsychology (pp. 53-95). Hillsdale, NJ: Lawrence Erlbaum Associates.

 

Felsman, J. K., & Vaillant, G. E. (1987). Resilient children as adults: A 40-year study.
In E. J. Anthony & B. J. Cohler (Eds), The invulnerable child (pp. 289-314). New York:
Guilford Press.

Fischhoff, P. S., & Slovic, P. (1980). A little learning...: Confidence in multicue
judgment tasks. In R. S. Nickerson (Ed), Attention and performance (pp. 779-800).
Hillsdale, NJ: Lawrence Earlbaum Associates.

Frisch, M. B., & Jessop, N. S. (1989). Improving WAIS-R estimates with the
Shipley-Hartford and Wonderlic Personnel Tests: Need to control for reading ability.
Psychological Reports, 63, 131-134.

Gurin, P., Gurin, G., Lao, R. C ., & Beattie, M. (1969) Internal-extemal control in the
motivational dynamics of Negro youth. Journal of Sociai Issues, 25, 29-53.

Hagborg, W. J. (1993). The Rosenberg Self-Esteem Scale and Harter's Self-
Perception Profile for Adolescents: A concurrent validity study. Psychology in the
Schools 30 132-136.

 

42

Hart, J. T. (1985). Memory and the feeling-of-knowing experience. Journal of
Educational Psychology 56, 208-216.

Harter, S. (1988). Manual for the Self-Perception Profile for Adolescents. Denver,
CO: University of Denver.

Hawkins, K. A., Faraone, S. V., Pepple, J. R., & Seidman, L. J. (1990). WAIS-R
Validation of the Wonderlic Personnel Test as a brief intelligence measure in a psychiatric
sample. Psychological Assessment, A 198-201.

Hurley, J. R. (1990). Constructive thinking and elevated ratings of self in
interpersonal groups. Journal of Psychology, 124, 563-573.

Hurley, J. R. (1991). Related measures of constructive and rational thinking. Journal
of Psychology, 125, 229-236.

Janoff-Bulman, R., & Brickman, P. (1982). Expectations and what people learn from
failure. In N. T. Feather (Ed), Expectations m Actions: Expectancy Value Models in
Psychology (pp. 207-23 7). Hillsdale, NJ: Lawrence Earlbaum Associates.

Kassinove, H, Crisci, R., & Tiegerrnan, S. (1977). Developmental trends in rational
thinking: Implications for rational-emotive school mental health programs. Journal of
Community Psychology, 5, 266-274.

Katz, L., & Epstein, S. (1991). Constructive thinking and coping with laboratory-
induced stress. Journal of Personality and Social Psychology, 61, 789-800.

Klinger, E. (1975). Consequences of commitment to and disengagement from
incentives. Paychological Review, 82, 1-25.

McFarlin, D. B. (1985). Persistence in the face of failure: The impact of self-esteem
and contingency information. Personality and Social Psychology Bulletin, 1 1, 153-163.

McFarlin, D. B., Baumeister, R. F., & Blascovich, J. (1984). On knowing when to
quit: Task failure, self-esteem, advice, and nonproductive persistence. Journal of

Personality, 52, 138-155.

McNemar, Q. (1963). Psychological Statistics (3rd ed). New York: Holt.

Mednick, S. A. (1962). The associative basis of the creative process. Psychological
Review 69 220-232.

 

Meichenbaum, D. (1977). Cagnitive behavior modification. New York: Plenum
Press.

43

Miller, N. E. (1948). Studies of fear as an acquirable drive: 1. Fear as motivation
and fear-reduction as reinforcement in the learning of new responses. Journal of
Experimental Psychology, 38, 89-101.

Orbach, 1., & Hadas, Z. (1982). The elimination of learned helplessness deficits as a
function of induced self-esteem. Journal of Research in Persoaality, 16, 511-523.

Roistacher, R. C. (1974). A microeconomic model of sociometric choice.
Sociometry, 37, 219-238.

Rosenberg, M. (1965). Societv and the adolescent self-image. Princeton, NJ:
Princeton University Press.

Rosnow, R. L., & Rosenthal, R. (1998). Focused tests of significance and effect size
estimation in counseling psychology. Journal of Counseling Psychology, 35, 203-208.

Sandelands, L. E, Brockner, J ., & Glynn, M. A. (1988). If at first you don't succeed,
try, try again: Effects of persistence-perforrnance contingencies, ego involvement, and self-
esteem on task persistence. Joumfi of Applied Psychology, 73, 208-216.

Seligman, M. E. P., Nolen—Hoeksema, S., Thornton, N., & Thornton, K. M. (1990).
Explanatory style as a mechanism of disappointing athletic performance. Psychological
Science 1 143-146.

 

Shaughnessy, J. J. (1979). Confidence-judgment accuracy as a predictor of test
performance. Journal of Research in Personality, 13, 505-514.

Shipley, W. C. (1946). Shipley Institute of Living Scale for measuring intellectuafl
impairment: Manual. Hartford, CT. Institute of Living.

Shrauger, J. S., & Sorman, P. B. (1977). Jourpal of Consultingand Clinical
Psychology, 45, 784-795.

Silber, E., & Tippett, J. (1965). Self-esteem: Clinical assessment and measurement
validation. Psychological Reports, 16, 1017-1071.

Stake, J. E. (1994). Development and validation of the six-factor self-concept scale
for adults. Educational and Psychological Measurement, 54, 56-72.

Tabachnik, B. G., & Fidell, L. S. (1989). Using multivariate statistics. California
State University, Northridge: Harper Collins.

Vaillant, G. E. (1977). Adaptation to life. Boston: Little, Brown.

44

Violato, C., & Travis, L. D. (1988). An application of generalizability theory to the
consistency-specificity problem: The transsituational consistency of behavioral persistence.
The Journal of Psychology, 122, 389-407.

Wechsler, D. (1955). Manual for the Wechsler Adult Inteflgence Scale QN AIS).
New York: Psychological Corporation.

 

Wechsler, D. (1981). Manual for the Wechs_ler Adult Intelligence Scale-Revised
(WAIS-R). San Antonio, TX: Psychological Corporation.

 

Weiner, B., Heckhausen, H., Meyer, W. U., & Cook, R. E. (1972). Causal
ascriptions and achievement behavior: A conceptual analysis of effort and reanalysis of
locus of control. Journal of Personality and Social Psychology, 21, 239-248.

Wonderlic, E. F. (1989). Wonderlic Personnel Test. Northfield, E: E. F. Wonderlic
Personnel Test, Inc.

Zajonc, R. B. (1980). Feeling and thinking: Preferences need no inferences.
American Psycholggist, 35, 151-175.

HICHI

13

can STATE UNIV. LIBRQRIES
WWW"Will"”WWIIHIHIIH"WWW
1293014096279