lNW)!NHH‘FLWWNHH[\IHUIHHMUEHI

l

  

3% MW

416

ZOO-7

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the
thesis entitled

The Effect of Relevant and Irrelevant Cognitive Effort on
Deception Detection Ability

presented by

Lauren M. Hamel

has been accepted towards fulfillment
of the requirements for the

Master of degree in Communication
Arts

 

 

    

/

' Major Professo‘ 5 Signature
?I:-: 1'71 [2 j; 7 6C 7
Date

MSU is an affinnative-action, equal-opportunity employer

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/CIRCIDateDue.indd-p.1 .

 

THE EFFECT OF RELEVANT AND IRRELEVANT COGNITIVE EFFORT ON
DECEPTION DETECTION ABILITY

By

Lauren M. Hamel

A THESIS
Submitted to
Michigan State University
in partial fitlfillment of the requirements
for the degree of
MASTER OF ARTS

Communication

2007

ABSTRACT
THE EFFECT OF RELEVANT AND IRRELEVANT COGNITIVE EFFORT ON
DECEPTION DETECTION ABILITY
By
Lauren M. Hamel

Research has found that high levels of cognitive effort result in a decreased
ability to distinguish truth from deception. Research has also used very artificial
methods to induce high cognitive effort (number memorization etc.).To improve upon
this research, an experiment was conducted in which different forms of cognitive load
were induced before participants completed a judgment task requiring them to detect
deceptive and truthfiil interview responses. College student participants were assigned
to one of four conditions. The conditions intended to induce high cognitive effort
required them to process and retain (1) task-relevant information, (2) task-irrelevant
information, or (3) series of numbers that were an irrelevant task. A fourth control
condition omitted the extra information and task. It was hypothesized that task-
relevant cognitive effort would aid detection accuracy, whereas irrelevant information
and an irrelevant task would hinder accuracy.

Results were not significant, however, the predicted pattern did emerge with
participants whose cognitive effort was relevant to their detection task having an
increased detection ability. In addition, the low cognitive effort condition, which
usually has the highest accuracy, had the lowest accuracy in this case.

Implications for how people believe and process information are discussed.

DEDICATION

I would like to dedicate this thesis to my wonderful, loving family, my
mother, my father, and my little sister Amanda. They have always encouraged me to
go after my dreams and to stay strong in the face of adversity. Their undying love and

support were my foundation during this experience.

iii

ACKNOWLEDGEMENT

I would like to gratefirlly acknowledge the support and guidance of my
advisors, Dr. Judee K. Burgoon and Dr. Sandi W. Smith. Their never-ending
enthusiasm for this project kept my spirits high at all times and their passion for their
craft keeps me hopefiJl and in anticipation of my future.

I would also like to acknowledge my fellow graduate students whose love,
support, and advice made this work possible. Specifically, Hillary Shulman, Jennifer
Maginnis, and Edward Glazer, thank you.

Finally, I am forever grateful to my loving family who has always supported
my academic endeavors and always give me the confidence to accomplish whatever I

set out to do.

iv

TABLE OF CONTENTS

LIST OF TABLES .............................................................................. vi
LIST OF FIGURES ............................................................................. vii
CHAPTER 1
INTRODUCTION .............................................................................. 1
Literature Review ..................................................................... 2
Belief Debate .................................................................. 2
Gilbert’s Three Stage Model of Information Processing. .. ..............3
Heuristic Systematic Model .................................................. 6
Three-Stage Process Model and Heuristic Systematic Model .......... 8
Type of Effort ................................................................. 9
Conceptual Definitions ............................................................... 9
Hypothesis and Research Question ................................................. 10
CHAPTER 2
METHOD ...................................................................................... 12
Pilot .................................................................................... 12
Independent Variables ............................................................... 12
Main Experiment ..................................................................... 14
Participants ........................................................................... 14
Procedure .............................................................................. 15
Manipulation Check ................................................................. 15
Dependent Measures ................................................................ 15
CHAPTER 3
RESULTS ....................................................................................... l7
Reliability of Measures .............................................................. 17
Manipulation Check .................................................................. l7
Hypothesis 1 .......................................................................... 18
Research Question 1 ................................................................. 23
CHAPTER 4
DISCUSSION ................................................................................. 25
Gilbert’s Three Stage Information Processing Model and Heuristic Systematic
Model .................................................................................. 27
Alternative Explanation ............................................................. 28
Belief Debate ......................................................................... 29
Limitations ............................................................................ 30
Future Research ...................................................................... 31
REFERENCES ................................................................................ 33

LIST OF TABLES

Table 1 Manipulation Check ................................................................. 18
Table 2 Detection Accuracy .................................................................. 19
Table 3 Accuracy Proportions for Each Statement ....................................... 21

vi

LIST OF FIGURES

Figure 1 Truth and Deception Accuracy per Statement .................................. 23

vii

Decades of research concerning deception in communication have made
apparent not only that deception exists in everyday life (DePaulo, Kashy, Kirkendol,
& Epstein, 1996) but also that most peOple are ill equipped to detect it (Bond &
DePaulo, 2006). Bond and DePaulo’s (2006) meta-analysis, which included 184
studies concerning the ability of individuals to detect deception, concluded that on
average people correctly identify truth 61 % of the time and deceit 47 % of the time.

Many researchers have tried to determine variables that, depending on their
level of influence, could inhibit or facilitate deception detection. Among the variables
that have been found to influence detection accuracy are interacting with the deceiver
(Burgoon, Buller, Ebesu, & Rockwell, 1994; Burgoon, Buller, White, Afifi, & Buslig,
1999), priming of suspicion (McComack & Levine, 1990b), attention to particular
communicative elements (DePaulo, Lassiter, & Stone, 1982), relationship to deceiver
(Comedena, 1982; Gagne & Lydon, 2004), training in deception detection (Levine,
Feeley, McComack, Harms, & Hughes), probing the deceiver (Stiff & Miller, 1986;
Buller, Strzyzewski, & Comstock, 1991a) and level of cognitive effort (Gilbert,
Pelham, & Krull, 1988; Gilbert, Krull, & Malone, 1990).

The current study concerns the latter--the effect of cognitive effort on
deception detection. More specifically, it is concerned with the manner in which
previous research has induced cognitive effort and attempts to improve upon past
efforts to make some more refined conclusions about how cognitive effort influences
deception detection. It will attempt to do so by inducing cognitive effort in a more
situationally relevant manner and comparing its effects to those of the traditionally

used methods. These traditional methods, which will be explained in greater detail in

the following section, generally consist of number memorization or arithmetic

problems as a means to induce greater cognitive effort in experimental participants.
Literature Review

Belief Debate

Before a review of previous research regarding cognitive effort is given, a
background on the theoretical debate of how individuals process and accept incoming
information will be outlined. This argument is important to understand because it is
what has helped start and develop much of the research that has thus far been
conducted concerning cognitive effort and many of its implications.

The process of belief is divided into the Cartesian viewpoint and the
alternative Spinozan View. The Cartesian viewpoint states that incoming information
is judged as true or false before it can be comprehended and stored in one’s memory.
An individual can hold a piece of information, make a judgment conceming its
veracity, and then accept it as either true or false (Gilbert, 1991). In other words, an
individual does not have to accept a piece of information as accurate before he or she
can judge whether it is indeed true or, after firrther deliberation, is actually false.

Alternatively, the Spinozan view of information comprehension states that
when an individual is exposed to new information, it will be initially accepted as true.
This initial belief is a necessary precursor for comprehension. Only later if the
individual realizes that this new information is inconsistent with previously held
understanding will any judgment of truth or deceit be made (Gilbert, 1991).

At first blush, the above argument may appear a bit trivial, however, its

implications involving deception and its detection are applicable to everyday life.

More specifically, if people function within the Spinozan view, and are predisposed
to believe all information they encounter before they can judge it, both lies and deceit
will initially be believed, and only uncovered if one has the ability to make that
judgment correction. It is possible for any number of deception attempts to go
undiscovered if that final stage of correction does not occur. The current research
takes the Spinozan view and a long line of similar research to determine if cognitive
effort has an effect on individuals’ ability to make that final determination. Moreover,
it attempts to make a narrower distinction of what effects certain types of cognitive
effort could have on deception detection.
Gilbert’s Three Stage Model of Information Processing

A major researcher in the area of cognitive effort and its effects on
information processing and deception detection is Gilbert (Gilbert et al., 1988, 1990).
His work in this area has led him and his colleagues to outline three steps of cognitive
processing. The first is categorization, the second is characterization, and the third is
correction. While in the categorization stage of processing, individuals will recognize
what someone is doing (Amanda is telling her professor that she missed her last exam
because her grandmother passed away). While in the characterization stage
individuals will make inferences about what trait that person is demonstrating through
that particular action (Amanda is a good student because she has a legitimate excuse
for missing her exam). Last, while in the correction stage individuals will consider
what possible situational constraints or prior knowledge may help form a more
accurate perception of the person’s action (Amanda has already used this excuse two

times earlier this semester, so she must be lying) (Gilbert et al., 1988).

It is believed that the first two stages of categorization and characterization
occur automatically and with very little effort (Gilbert et al., 1988). However, it
appears that the final stage of correction is much less automatic and requires more
cognitive resources to complete than the previous two stages (Gilbert et al., 1988). To
refer back to the previous of example of the lying student, it is easy to see what
Amanda is doing and to make the obvious inferences. However, it takes an increased
amount of cognitive effort and resources to find and apply the extra situational
information that changes the perception of the interaction entirely. Distinguishing
among these three stages is important here because it provides a framework for what
occurs when people take in and process new information, even when that information
may not be truthful.

Gilbert et al. (1990) examined the consequences incompletion of this final
stage has on deception detection. Through experimental manipulation of cognitive
effort and deception, they found evidence that if people are being deceived, they will
detect it in the correction stage of information processing. In this study, participants
were given a list of propositions regarding the translation of English words into Hopi
Indian words. Some of these propositions were true and some were false and
participants were instructed to judge their veracity. To determine the consequences of
the incompletion of the final stage of correction, some participants were interrupted
during their judgment tasks and forced to make their decisions more quickly. As
predicted, participants’ ability to accurately determine truthful statements was not
affected by the interruption (55 % when uninterrupted vs. 58 % when interrupted),

but did reduce their ability to detect deception (55 % when uninterrupted vs. 35 %

when interrupted). This is evidence that when people are cognitively busy, they will
not enter into the correction stage and therefore be unable to detect the deception
attempt.

To further examine this three stage process of information processing, Gilbert
and Osborne (1989) conducted an experiment to see if and how participants who
make inaccurate judgments of others are able to correct those inaccuracies.
Participants either watched a muted video tape of a person discussing an anxious
topic or a mundane topic. They were all kept cognitively busy by rehearsing an eight-
digit number. Once the tape was over, participants stopped their cognitive busyness
by recalling the number for the experimenter. Next, half of the participants were
asked to think about the videotaped person in a variety of mundane and anxiety-
provoking situations (other-thought) and the other half were told to think of
themselves in the same situations (self-thought). All participants then rated the video
taped person’s trait anxiety. Next, subjects who had previously taken part in other-
thought now took part in self-thought and vice versa. Finally, participants rerated the
video taped person’s trait anxiety. As predicted, participants who took part in other-
thought were more likely to correct their previous erroroneous trait attributional
assumptions when taking situational contexts into consideration. This is further
evidence that not only does cognitive effort reduce judgment accuracy, but the final 4
stage of correction does not inevitably follow the reduction of the cognitive effort.
This indicates that the final stage is much more elfortfirl in comparison to the first

two stages of categorization and characterization.

Previous research has made strides in determining how the induction of
cognitive resources will effect information processing, and more specifically
deception detection, however, there are some issues regarding the ecological validity
of the manner in which cognitive effort has been induced. In past research, cognitive
effort has usually been induced by having individuals engage in cognitive rehearsals
(Gilbert et al., 1990), memorize long numbers (Gilbert & Osborne, 1989; Crisp,
Perks, Stone, & Farr, 2004; Weary, Vaughn, Stewart, & Edwards, 2004), work under
time pressure (Silvera, 2000), count backward by a consistent interval from a high
number (F eeley & Young, 2000), or work on arithmetic problems (Millar & Millar,
1997). These cognitive effort manipulations are irrelevant to the task of determining
tnrthfulness and might be producing results that will not generalize to the effects of
more message-relevant forms of cognitive effort. In other words, there may be a
difference between cognitive effort that acts as more of a distractor and cognitive
effort directly related to the task in which the participants are involved. Put in the
context of deception, the ability to detect deception may not be the same for those
who are induced with relevant versus irrelevant cognitive effort. Thus the current
investigation seeks to expand upon previous research concerning the effect cognitive
effort has on deception detection ability by comparing relevant cognitive effort and
irrelevant cognitive effort.

Heuristic Systematic Model
The Heuristic Systematic Model (Chaiken 1980, 1987) of information

processing will serve as a fi'amework to guide the current research. Its inclusion of

dual modes of processing will allow for the different types of cognitive efforts being
studied.

The HSM posits that individuals can process information through two
different routes, a systematic route and a heuristic route. The systematic route of
processing involves a purposeful scrutinization of incoming information. All new
information is carefirlly evaluated until the individuals can make an efficient decision
or form an educated opinion on what is being communicated to them. In addition,
those who process systematically will differentiate between strong and weak
arguments and are unaffected by heuristic elements of the message that do not pertain
directly to the content of the message. Conversely, the heuristic route of information
processing is based on peripheral contextual information. Decisions and opinions are
formed by the influence of heuristics such as the sources of the message or the length
of the message (Todorov, Chaiken, & Henderson, 2002).

The HSM posits that the processing route used to assess incoming information
will be determined by an individual’s motivation and ability, which is determined by
available cognitive resources. In other words, an individual who is motivated to listen
to and understand new information will be more likely to engage in systematic
processing when compared to someone who is not motivated to process the
information. Also, a person who has the available cognitive resources to conduct this
type of extensive processing necessary for the systematic route will be more likely to

do so than someone who does not have the required resources available.

Three-Stage Process Model and Heuristic Systematic Model

Gilbert’s three-stage model of information processing and the HSM will be
used jointly to inform the current research. The first two stages of the three-stage
model, categorization and characterization, are thought to occur automatically that is,
an individual will make judgments of what another is doing and infer something
based on those actions with little effort. However, the third stage of correction, which
serves as a time for any erroneous judgments to be corrected, may not occur if an
individual is cognitively busy. The HSM corroborates this three-step process by
outlining two routes to information processing: systematic and heuristic. As
previously explained, the systematic route will be utilized when there is an ideal
amount of cognitive capacity available to scrutinize the information. Conversely, the
heuristic route will be dominate when very little cognitive capacity is available. The
heuristic route can be viewed as the occurrence of the first and second stage of
Gilbert’s three step model and the systematic route as the entire three-step process. In
other words and based on previous research, people who are cognitively busy will
default to the heuristic route of information processing and only pass through the first
two stages of categorization and characterization and will therefore, be less likely to
distinguish between truth and deceit. However, those who are not cognitively busy
will be able to use the systematic route, reach the final step of correction in their
information processing, and will be more likely to pick out truth from lies. The
current research, however, seeks to determine if this relationship between cognitive
effort and information processing continues to hold if cognitive effort is relevant to

the situation of deceit

Type of Eflort

As previously discussed, the current research is concerned with the past
operationalization of cognitive effort. Specifically, past research has induced
cognitive effort in a very artificial and ecologically invalid manner. The current
research is improving on past efforts by offering a more relevant alternative of
cognitive effort. In addition, because past research has offered consistent findings in
the differences between high and low cognitive effort (Gilbert, Krull, & Malone,
1990; Gilbert, Krull, & Pelham, 1988; Gilbert & Osborne, 1989; Pontari & Schlenker,
2000; Silvera, 2000) the current research is only concerned with high levels of
cognitive effort (relevant and irrelevant) because that is where it is anticipated the
most important effects will be found. In other words, low cognitive effort relevant or
otherwise will not hinder information processing or deception detection ability. For
this reason efforts were focused on the different types of high cognitive effort and
their effects on detection accuracy.
Conceptual Definitions

To clarify among the different types of cognitive effort being induced in the
current research, a conceptual definition of each is provided. Relevant cognitive effort
is mental load induced through an increased amount of information directly related to
the subject the possible truths and lies are regarding. Irrelevant cognitive efl'ort is
mental load induced through an increased amount of information unrelated to the

subject the possible tnrths and lies are regarding.

Hypothesis and Research Question

Based on past research, it was predicted that high irrelevant cognitive effort
inhibits deception detection. However, based on the assumptions of the HSM, it was
predicted that high relevant cognitive effort will act as a detection facilitator. It was
thought that those who are induced with relevant cognitive effort will take the
systematic route of information processing because the increased load will not act as
a distracter but as a mechanism for increased higher-order processing. This prediction
that high relevant cognitive effort will increase deception detection accuracy is
counterintuitive to previous thinking, however, it is theoretically driven. Specifically,
if relevant cognitive effort acts as predicted, participants with this type of load will
process incoming information systematically because this type of load is directly
relevant to the task the participants are required to partake in. With the increased
relevant cognitive effort and the expected increase of systematic processing,
participants will be better equipped to distinguish truths from falsehoods.

The current research is predicting that high cognitive effort will not always be
detrimental to deception detection if it is relevant to the context of the deceit. This
prediction is based on the two factors the HSM postulates determine which route of
processing is used: ability and motivation. The current research will hold motivation
constant and vary ability by inducing either relevant or irrelevant cognitive effort.

Past research has shown that increased cognitive effort inhibits deception
detection. However, it is predicted that when high cognitive effort is paired with
relevant information, the opposite will be true. In other words, those who are less

cognitively taxed will be better at detecting deception except compared to those are

10

cognitively taxed but by information that is task-relevant. Increased cognitive effort
will facilitate deception detection in those who receive task relevant information
because they will process this information in a more systematic manner. Because this
systematically processed information is relevant to their task, the higher level
processing will increase participant knowledge and focus and allow for an increased
ability to detect any subsequent deception. When these two types of cognitive effort —
relevant and irrelevant- are compared, the following is hypothesized:

H1: Relevant cognitive efi'ort will produce higher deception detection

accuracy than irrelevant cognitive effort.

In addition to comparing the differences between relevant and irrelevant
information on deception detection accuracy, an irrelevant task manipulation was
included to compare the current research with previous research designs. To
determine if irrelevant information and an irrelevant task induce equal levels of
cognitive effort and consequent decrements in detection accuracy, the following
research question is posed:

R1: Does irrelevant information produce diflerent eflects on detection

accuracy than an irrelevant task?

11

Method
Pilot

A pilot effort was implemented to create the stimuli to be used in the main
experiment. Ten female participants completed a task concerning how best to survive
in the desert called Imperative Information: Surviving in the Desert. The information
for this task was compiled from several websites and resources on desert survival,
including Ganci (1991), Kraus (197 8), and Nelson and Nelson (1997). The desert
survival task was a stimulus used in past research as a prompt for discussion between
study participants (Burgoon, Bonito, & Hamel, 2006). The task involved participants
prioritizing what items would be most needed if stranded in the desert. These 10
participants were put into pairs to discuss and create their priority lists. One member
of each dyad was privately instructed to be deceitful about her true opinions regarding
the prioritizing of salvable items. These discussions were audio recorded and one
honest and one deceptive utterance was selected fi'om each dyadic conversation for
use as the stimuli in the main experiment.
Independent Variables

Relevancy of Information. One-fourth of the participants were randomly
assigned to the high information relevancy condition. They read the background
information of the Imperative Information: Surviving in the Desert document. This
document was used as background information regarding the truthful and deceptive
messages the participants would eventually be exposed to. Another fourth of the
participants were assigned to the low relevancy of information condition. They

instead read a document about desert flora and fauna, a topic that was tangentially

12

related to the desert scenario but had nothing to do with the desert survival discussion
they would eventually listen to. The remaining half of the participants did not receive
any background information.

Cognitive Eflort Three fourths of participants were induced with high
cognitive effort. The desert survival and desert flora and fauna background
information were written at a collegiate reading level to induce high cognitive effort.
To ensure that equivalent reading levels and length of document were obtained, the
documents were evaluated with Word Perfect’s Grammatix tool which provides
measures of lexical diversity and the F leish-Kincaid readability index.

Once participants in the previously described conditions read their respective
instructions, they listened to conversational excerpts from the pilot. The conversations
were segmented into individual utterances. At the end of each segment, participants
rated how credible and truthfirl they felt each person appeared to be.

To maintain a high level of cognitive effort, participants were told that they
would be quizzed at the end of the session. They were told they would have to recall
five of the 10 statements they listened to. This was meant to utilize participants’
cognitive resources, but to do so in a manner that is relevant to their task.

Participants in the irrelevant task condition were not given any background
information to read. Instead they were instructed to memorize five three-digit
numbers. This condition was implemented to compare the current research’s
manipulation of relevant cognitive effort to previous research methods for inducing
cognitive effort. This third condition was added due to the nature of the comparison

of relevant and irrelevant information. A complete 2X2 design (which is typical for

13

this type of research) could not be created because there does not exist a relevant task
in the same way there is an irrelevant task. In other words, the relevant information
condition requires information that is related to the subsequent truthfirl and deceptive
statements. There cannot be a relevant task condition that involves the same crucial
components of the relevant information condition and can still be a task like the
irrelevant task condition.

Control An offset control condition was included to compare participants in
the other three conditions to participants who were not induced with any cognitive

effort and only asked to judge the veracity of each utterance.

 

 

 

 

 

Conditions
Level of Effort Relevant Information Irrelevant Irrelevant Task
Information

High -High reading level -I-ligh reading level -No background
-On-topic Off-topic information
information information -Memorize five
-Memorize five -Memorize five three-digit numbers
statements statements

Control

-No background

information

-No memorization

 

Main Experiment

A 1 (amount of effort) x 3 (relevancy of information/task) with an offset
control design was implemented to test how high cognitive effort and relevancy of
information effect the ability to detect deception.
Participants

One hundred and four undergraduate students from a large Midwestern

university were recruited to participate in a deception detection study as a fulfillment

l4

for a requirement for course credit. Participants also had the opportunity to earn a $25
cash prize for being the most accurate at distinguishing between truth and deceit.
Procedure

Participants were brought into a lab and randomly assigned to one of the four
conditions. They were told by an experimenter that they were going to engage in a
deception detection and memorization task and that their level of accuracy at each
task would result in a reward once they have completed their task. The participants
were instructed to read the directions on the packet of materials that was in front of
them. These materials differed based on condition.

They were asked to read the document relevant to their condition or to
memorize a number and then judge the veracity of the same 10 conversational
excerpts from the pilot.

Manipulation Check

After participants were finished with the experiment, they were asked to
respond to three items regarding how relevant they felt the information was to the
conversation they listened to and 18 items regarding how much cognitive effort they
felt they had to expend to read through the initial instructions and to memorize the
necessary information. These items were included as a manipulation check to ensure
cognitive effort and relevancy of information were manipulated as intended.
Dependent Measures

All subjects, regardless of condition, were asked to judge each conversational
excerpt on a dichotomous measure of truth or deception. In addition, they were

instructed to give perceived credibility ratings for each excerpt. The veracity of each

15

statement was known by the experimenter so an objective right or wrong answer
could be calculated. Separate percentage correct judgments were calculated for truth

and deception.

16

Results

Reliability of Measures

Reliability checks of all the measures taken produced high reliabilities. The
scale used to measure cognitive effort was composed of 18 items which assessed
cognitive effort with direct items (i.e. how much mental activity was required to
complete this task?) and indirect items (i.e. how tense were you while completing this
task?) The direct items were measured on an 11-point scale, the indirect items were
measured on a seven-point scale and yielded a reliability of a = .96. The perceived
relevance of the document read by the participants in the relevant and irrelevant
information conditions was measured with three items on seven-point scales which
had a reliability of a = .91. The credibility of the document was assessed with six
items on seven-point scales with a reliability of a = .85. Last, the self reported
motivation of the participants was assessed with three items and had a reliability of a
= .70.
Manipulation Check

To ensure that cognitive effort was manipulated properly, a planned contrast
‘ compared the control condition to the three experimental conditions. Results
indicated that those in the control condition experienced significantly less cognitive
effort than those in the relevant information condition ,the irrelevant information
condition, and the irrelevant task condition (see Tablel), t(99) = 2.71, p <.01. In
addition, contrast tests revealed that the irrelevant task did not differ from the two
information conditions (p > .20) and the relevant and irrelevant information

conditions did not differ from one another (p > .20) (See Table 1).

l7

A manipulation check was also conducted to determine if the level of
participant motivation was comparable across conditions. A one-way analysis of

variance produced no significant difference in motivation among the four conditions,

F(3, 98) = .08, p = .97, n2 = .002, relevant information, irrelevant information,

irrelevant task, and the control (see Table 1).

Last, a manipulation check was conducted to ensure there was a comparable
level of memory accuracy among the three conditions that required participants to
either memorize facts or numbers. A one-way analysis of variance among the three

experimental conditions found no significant difference in memory accuracy was

found among the four conditions (see Table 1), F(2, 74) = 1.14, p = .32, r]2 = .03.

Table 1 Manipulation Check

 

 

 

 

 

 

Cognitive Effort Motivation Memorization

Accuracy ,

M SD M SD M SD
Relevant 5.10 1.31 4.92 1.11 2.56 .51
Information
Irrelevant 4.66 1.05 5.04 .85 2.20 1.08
Information
Irrelevant 4.83 1.62 4.98 1.14 2.37 .84
Task
Control 4.06 1.27 5 .04 1.04
Hypothesis 1

The hypothesis predicted that participants induced to expend relevant
cognitive effort would be more accurate at detecting deception than those induced to
expend irrelevant cognitive effort. Hypothesis 1 was tested with multiple contrast

analyses. The relevant information condition was compared to the remaining three as

18

an aggregate and then the three experimental conditions (relevant information,
irrelevant information, and irrelevant task) were all compared to each other using a
Dunnett’s comparison. No significant differences were found in any of the
comparisons. See Table l for means and standard deviations of each condition.
Inspection of the means reveals that accuracy was higher (though not significantly so)
with both the relevant information and irrelevant task than with the irrelevant
information; the latter condition was on a par with the control condition. These
patterns are contrary to what was hypothesized.

Table 2 Detection Accuracy

Total Accuracy Truth Accuracy Deception

 

 

 

 

 

 

Accuracy

M SD M SD M SD
Relevant Information .62 .13 .65 .18 .60 .22
Irrelevant Information .59 .16 .63 .17 .54 .23
Irrelevant Task .63 .15 .66 .17 .60 .18
Control .58 .20 .59 .23 .57 .21
Overall .61 .16 .63 .20 .58 .21

 

To examine firrther the relationship between the different types of cognitive
effort and the ability to distinguish truthfirl statements from deceptive statements,
separate one- way analyses of variances were conducted for overall accuracy for

truthfirl statements and deceptive statements. Neither the truthfirl statement analysis

[(3, 100) = .629, p = .60, n2 = .02 (see Table 2) nor the deceptive statement analysis

[(3, 100) = .412, p = .745, r]2 = .01 (see Table 2) yielded significant differences

among the cognitive effort conditions. Lowest deception detection accuracy scores

again accrued in the control and irrelevant task conditions, and the control condition

19

also yielded the lowest truth detection accuracy, although the differences failed to
achieve statistical significance. The patterns did not conform to expectations.

As can be determined fi'om the overall accuracy proportions as well as the
truth and deception accuracy proportions, these data do not conform with what has
previously been reported concerning the ability to distinguish between truthfirl and
deceptive statements. Bond and DePaulo’s (2006) meta-analysis reported an overall
accuracy rate of 54 % and separate accuracy rates for truth and deception detection
accuracy of 61 % and 47 % respectively. The data presented here show an overall
accuracy proportion of 61 %, a truthful accuracy proportion of 63 %, and deception
accuracy proportion of 58 %. The truth accuracy percentage is not significantly
different from what was found in the meta-analysis t(103) = -1.52, p = .132 but
deception accuracy is t(103) = 2.68, p =.009. A possibility that could account for this
11-point discrepancy is the stimuli chosen for this research.

In line with previous research, these data show that accuracy was generally
higher among truthfirl statements when compared to deceptive statements. However,
accuracy for deceptive statements was higher than what is typically reported. A closer
examination of the accuracy rates on a per question basis indicates that several
questions were consistently judged correctly across all conditions (T1, T5, and D5).
In addition, although not significantly different from the accurately judged questions,
several questions were also consistently judged incorrectly across all conditions (T2,
T3, and D4). This suggests that the results may be due to a poor choice in stimuli. In
other words, participants were consistently able to discriminate between truth and

deception for some statements, but for others the stimuli appear to be either too easy

20

or too difficult to offer any real possibility for cognitive effort to play a part in
inhibiting or facilitating accurate judgments.

Table 3 Accuracy Proportions for Each Statement

 

 

 

 

 

 

 

 

 

 

 

 

 

Condition
Truthful Category Relevant Irrelevant Irrelevant Control Overall
Statements Information Information Task
T 1 Accurate .88* .84* .89* .85* .87*
Inaccurate .12 .16 .11 .15 .13
T 2 Accurate .48 .44 .37 .22* .38
Inaccurate .52 .56 .63 .78 .62
T 3 Accurate .40 .44 .67 .44 .55
Inaccurate .60 .56 .33 .56 .45
T 4 Accurate .84* .56 .63 .44 .55
Inaccurate . 16 .44 .37 .56 .45
T 5 Accurate .64 88* .74* .78* .76*
Inaccurate .36 . 12 .26 .22 .24
Deceptive
Statements
D 1 Accurate .76* .72* .81* .78* .77*
Inaccurate .24 .28 . 1 9 .22 .23
D 2 Accurate .52 .36 .59 .63 .62
Inaccurate .48 .64 .41 .37 .38
D 3 Accurate .68 .68 .59 .67 .65
Inaccurate .32 .32 .41 .33 .35
D 4 Accurate .40 .40 .41 .30 .38
Inaccurate .60 .60 .59 .70 .62
D 5 Accurate .64 .56 .59 .48 .57
Inaccurate .36 44 .41 .52 .43

 

* Significantly different within condition at p < .05

In addition, a simple tally of how many statements resulted in a majority of
inaccurate judgments shows that, although indicated as the most accurate condition in
previous research, the control condition resulted in the most statements that were
judged inaccurately. The control condition contained five statements judged
inaccurately by the majority whereas the relevant information, irrelevant information,
and irrelevant task conditions resulted in three, four, and two statements misjudged,

respectively, by the majority.

21

An item analysis also confirmed that although a few statements were
consistently judged either correctly or incorrectly, those who had the highest overall
accuracy were also the participants who consistently judged statements correctly,
even if those statements were generally judged incorrectly. In the same vein, the item
analysis also confirmed that those who have low overall accuracy, judged statements
incorrectly even when those statements were generally judged correctly.

Consistent with previous research, a main effect was found for accuracy. That

is, accuracy was significantly higher for truthful statements when compared to

deceptive statements F( 1, 100) = 5.521, p = .021, r]2 = .052. The estimated marginal

mean of truthful accuracy was M = .63, SD = .02 and deception accuracy was M =
.58, SD = .02.
An interaction effect was also found between accuracy and the specific stimuli

used. Specifically, the main effect for accuracy was not consistent across all ten

statements F(1, 100) = 7.867, p < .000, n2 = .245. See Table 4 and Figure 1 for

accuracy means for all ten statements. In Figure 1, the five statements represent the

five consecutive stimuli that were truthfirl or deceptive.

22

Figure 1 Truth and Deception Accuracy per Statement

1-

 

 

 

 

 

 

E

O

o- \

go's " / -—-Truthful
E 0.6 a ~/_2._/ Accuracy
2: 0.4 V ‘~ ------- Deception
‘3 02 __ _ _.;A£EEEY_.
g .

< o I 77 I r fl

 

 

Statement

Figure 1 demonstrates that some statements (T1 and D1, for example) are
consistently judged accurately. However, other questions (T2 and D4, for example)
are consistently judged inaccurately. This finding is a problem in regard to testing the
effect cognitive effort has on deception detection. These findings pertain across all
conditions, so it is apparent that regardless of the effect of the different types of
cognitive effort induced, participants were consistently accurate or inaccurate on the
statement being judged in certain cases. Although this interaction effect does not
pertain directly to the hypothesis being tested, it does provide evidence that testing
different types of cognitive effort should not end here.

Research Question 1

The research question asked whether cognitive effort induced by irrelevant
information produces different effects on detection accuracy than cognitive effort
induced by an irrelevant task. In other words, did those whose increased cognitive
effort was due to reading the document regarding desert plant life differ in their
ability to detect deception when compared to those whose increased cognitive effort

was due to number memorization? An independent sample t test reveals that these

23

two conditions did not differ in ability for (a) overall accuracy of judging truthfirl and
deceptive statements t(50) = -.97, p = .338, (b) judging truthfirl statements t(50) = -
5.7, p = .571, or (c) judging deceptive statements t(50) = -.96, p = .341. These
findings are reassuring regarding the current research because, as previously
explained, this is the first time cognitive effort was induced using reading level.
Previous research has relied on number memorization tasks (similar to the irrelevant
task condition in this study), complicated math problems, or counting backward from
a higher number using difficult intervals. Despite the fact that significant accuracy
differences were not found among the experimental conditions, the current research
can offer firture endeavors the choice of inducing cognitive effort in a way that

appears to be effective and comparable to previous methods.

24

Discussion

The purpose of this research was to extend previous knowledge of the effects
of cognitive effort on the ability to distinguish between truth and deceit. Past
endeavors in this area of research have induced cognitive effort in a manner irrelevant
to the task of detecting deception and have generally found this type of effort inhibits
ability to distinguish between truth and deception. The current research compared the
effects of relevant cognitive effort to those of irrelevant cognitive effort.

The data from the current research was not consistent with the proposed
hypothesis. Overall, relevant cognitive effort did not facilitate deception detection as
the current research predicted. In addition, cognitive effort in general did not inhibit
deception detection as previous research has indicated. In fact, results were the
reverse of what is generally obtained fi'om deception detection research involving the
manipulation of cognitive effort. Specifically, it is usually the case that when
compared to participants who undergo inductions to elevate cognitive effort, those in
the effortless control conditions generally exhibit increased ability to distinguish
between truth and deception. That was, however, not the case here. The control
condition was actually the lowest at detecting truth from deception. Despite the fact
that the results were not consistent with (a) previous research or (b) the current
research predictions, the idea of relevant versus irrelevant cognitive effort should not
be dismissed. The results may not have been a result of only the experimental
manipulation of the type of cognitive effort. Several factors may also have played a

part in the results of this experiment.

25

After examining the data on a per statement basis, it is apparent that there was
a high amount of variability among the stimuli themselves. It is possible that the
relative ease of some statements and extreme difficulty of others did not allow for
enough discrimination of the observers. More ambiguous stimuli would produce a
more reasonable challenge and valid test of the ability to distinguish truth from
deception when under cognitive load. Future research is needed to rule out the
possibility that the current results represent an artifact of the specific stimulus set.

In addition to the variability of the stimuli given, judgments were based solely
on audio recordings. Observers only had a short amount of monologue fi'om which to
form their judgments. All ten statements were less than 20 seconds of talk time. If
participants were presented with stimuli with more channels of communication, such
as visual cues, and periods of communication of a few minutes, results may have been
different than what was obtained with the current stimuli.

It is also possible that, although the cognitive effort conditions reported higher
levels of cognitive effort compared to the control conditions, not enough cognitive
effort was induced to either inhibit or facilitate deception detection. Perhaps a more
cognitively taxing task is needed to occupy more of the participants’ mental resources
which, depending on what type of effort is induced, will either draw attention away
from the detection task or allow for greater focus on it.

Another factor that may have had an effect on the results of this study is the
sample size. The number of participants allowed for enough power to make
reasonable inferences and conclusions, yet it is possible that with more people, the

differences among the cognitive effort conditions would become more apparent.

26

Despite that the results failed to achieve conventional levels of significance.
Generally speaking, accuracy was higher in the relevant information condition when
compared to the irrelevant information and irrelevant task condition. These findings
may indicate a finer difference that requires a larger sample size to detect this effect.
Gilbert ’s Three Stage Information Processing Model and Heuristic Systematic Model

Although the current data did not fall in line with what was predicted, what
has been found here can lend to further knowledge regarding information processing.
The three experimental conditions experienced greater cognitive effort than the
control group, but these participants were still able to process systematically and
therefore had an increased ability to distinguished truth from deceit.

Previous research has reasoned that control groups in deception detection
research involving cognitive effort are generally more successful due to the fact that
those in this group are able to put their firll focus on the possibly deceptive
communication. However, due to the innovative manner in which the current research
induced cognitive effort, an unforeseen effect may have occurred. Specifically, it is
possible that those who were induced with relevant and irrelevant cognitive effort
were inadvertently induced to be more attentive generally to the task at hand. This,
then, could have resulted in higher awareness and greater focus on the issue of
determining truth from deception and therefore an increased ability in doing so.

To recap the three stages of Gilbert’s model of information processing, when
new information is obtained, the first two stages of categorization and
characterization are automatic. These stages allow those who are processing

information to evaluate what the information is and then make basic inferences about

27

where the information has come from. The third stage of correction, however, is less
automatic and requires higher level thinking before any mistakes in the previous
stages can be corrected. Based on what was found here, although cognitive effort was
induced, it did not appear to inhibit that final stage of correction as has been
suggested previously. This undiminished ability is even more pronounced when the
three high effort conditions are compared to the low effort control condition.
Participants in the control condition fared the worst at distinguishing between
truthtellers and deceivers.

What can account for this inconsistent finding? One possibility is that those in
the high effort conditions had an increased awareness of their task when compared to
participants in the control condition. It could be the case that those in the high
cognitive effort conditions had to concentrate harder on their task than those in the
control condition and thus, more systematic processing occurred in these conditions.
The control condition, however, did not require the same amount of concentration.
Not being required to read or memorize a list of facts or numbers could have allowed
these participants the luxury of peripheral processing and therefore diminished their
ability to detect truth and deception. This lack of systematic processing would have
left this group susceptible to biases that caused their poorer performance.

Alternative Explanation

These data are evidence of what Langer (1978) has proposed as an alternative
conceptualization of cognitive capacity. In her viewpoint, cognitive capacity is not
necessarily finite. In fact, when given a cognitively effortfirl task, cognitive awareness

may actually increase rather than decrease. That increased awareness then results in

28

increased mindfulness and an enhanced ability to perform certain tasks. The increased
performance of the high cognitive effort conditions may have been a result of
increased mindfulness.

Belief Debate

To review the debate over belief described earlier in this paper, the Cartesian
viewpoint states that new information can be held suspended without evaluation
while an individual judges it as either truthful or deceptive. After this judgment
occurs, the individual can then comprehend the new information. Conversely, the
Spinozan point of view argues that one must accept new information as true before
comprehension can occur. Only once the new information is accepted as true and then
comprehended can an evaluation of its veracity be made. If the previous interpretation
of the results is correct, then the data are somewhat consistent with the Spinozan
viewpoint. The results from the three high cognitive effort conditions are consistent
with this point of view. Despite an increase in cognitive effort, these participants
achieved higher overall accuracy that was due to an increase in systematic processing.
This higher level of information processing allowed the third and final stage of
correction to occur and then resulted in an increased ability.

If it is indeed increased awareness and the resultant systematic processing that
facilitated increased ability at distinguishing between truthfirl and deceptive
statements in the high cognitive effort conditions, then this lends partial support to the
Spinozan argument of belief. This improved ability is predicted by both the HSM and
Gilbert’s three—stage information processing model. However, the data do not fall

exactly in line with what either model predicts. Ifthe Spinozan View held true across

29

all conditions then the control condition should have defaulted on a tnrth more often,
resulting in high truth accuracy but low deception accuracy. This, however, was not
the case. The control group does not offer overarching support for either side of the
belief debate, but with a larger pool of statements to judge more decisive evidence
could become apparent. These two findings offer some support that systematic
processing and its higher degree of cognitive effort help facilitate the ability to
accurately determine when one is telling the truth or lying but more research is
needed, especially in the area of low cognitive effort, to provide any more concrete
conclusions.
Limitations

As previously mentioned, the sample size of this experiment may have been a
limitation. The sample size used was large enough to obtain an effect size of d = .5,
but only if the relevant information condition was compared to the other three
conditions as a group. A larger sample will be needed to increase the power enough to
allow for individual comparisons of each group. Ifa larger sample were used, the
differences among the high cognitive effort conditions and the control may have been
more pronounced. Although the differences between the high cognitive effort
conditions and the control conditions are interesting in terms of level of accuracy,
without statistical significance, only speculations can be drawn at this point.

Another component of the experiment that could be improved upon is the
stimuli used. The recorded stimuli were naturally flowing from conversation but the
deceit within was sanctioned by the experimenter. Although difficult to obtain,

naturally occurring deception is ideal for research involving deception detection.

30

Another limitation of the stimuli is that they were only audio examples of
truthful and deceptive communication. Ifobservers were presented with truthful and
deceptive stimuli that also allowed a visual channel of communication, the data may
have been more informative.

Lastly, the stimuli were relatively short. All audio segments were less than 20
seconds. This limited amount of information may not have allowed observers enough
time to acclimate themselves to the potential truthtellers or deceivers speech patterns.
It could be the case that given more baseline information regarding communication
patterns of those being observed, more informed judgments could be made.

A tangentially related limitation to the issue with limited communication is
the problem of lack of conversation. Observers were presented with a few sentences
given by one person. Stimuli involving two interlocutors could also provide firrther
information that would alter veracity judgments.

With the previously mentioned changes to the experiment, the already existing
pattern of results which include an increased ability of detecting truth fiom deception
in the high effort relevant information condition, may reach levels of significance. In
addition, because the control condition did not offer a consistent pattern of results,
this condition of low cognitive effort needs the recommended changes as well to
allow for a more definitive comparison to the high cognitive effort conditions.
Future Research

The idea presented here of different types of cognitive effort should continue
to be explored so that a more nuanced understanding of the influence of cognitive

effort can be achieved. No significant differences in accuracy were obtained,

31.

however, preliminary patterns show that the relevant cognitive effort condition did
tend to aid in accurately determining when someone was telling the truth or lying.
Future research should take the suggested improvements--including stimuli that
mimic real life conversation and larger sample size-~to further examine the different

types of cognitive effort and their effects on deception detection.

32

REFERENCES

Bond, C. F. & DePaulo, B. M. (2006) Accuracy of deception judgments. Personality
and Social Psychology Review, 10, 214-234.

Buller, D. B., Strzyzewski, K. D., & Comstock, J. (1991a). Interpersonal deception I:
Deceivers’ reactions to receivers’ suspicion and probing. Communication
Monographs, 58, 1-24.

Burgoon, J. K., Bonito, J. A., & Hamel, L. M. (2006, November). Efi’ects of
communication modality on communication and decision quality:
Does cognitive load matter? Paper presented at the meeting of the National
Communication Association, San Antonio, TX.

Burgoon, J. K., Buller, D. B., Ebesu, A. S., & Rockwell, P. (1994). Interpersonal
deception V: Accuracy in deception detection. Communication Monographs,
57, 219-230.

Burgoon, J. K. Buller, D. B., White, C. H., Afifi, W., & Buslig, A. L. (1999). The role
of conversational involvement in deceptive interpersonal interactions.
Journal of Personality and Social Psychology Bulletin, 25, 669-685.

Buxton, P. A (1955). Animal life in deserts: A study of the fauna in relation to the
environment. London: Edward Arnold Ltd.

Chaiken, S. (1980). Heuristic versus systematic information processing and the use of
source versus messages cues in persuasion. Journal of Personality and Social
Psychology. 51, 1032-1043.

Chaiken, S. (1987). The heuristic model of persuasion. In M. P. Zanna, J. M. Olson,
& C. P. Herman (Eds), Social Iryluence: The Ontario Symposium 5 (pp. 3-
39). Hillsdale, NJ: Lawrence Erlbaum.

Comadena, M. E. (1982). Accuracy in detecting deception: Intimate and fiiendship

relationships. In M. Burgoon (Ed) Communication Yearbook 6 (pp.
446-472). Beverly Hills, CA: Sage.

Crisp, R. J ., Perks, N., Stone, C. H, & Farr, M. J. (2004). Cognitive busyness and the
processing of evaluative information in intergroup contexts. Journal of Social
Psychology, 144, 541-544.

DePaulo, B. M., Kashy, D. A., Kirkendol, M. M., & Epstein, J. A. (1996). Lying in
everyday life. Journal of Personality and Social P.9Ichology, 70, 979-995.

33

DePaulo, B. M., Lassiter, G. D., & Stone, J. I. (1982). Attentional determinants of
success at detecting deception and truth. Personality and Social Psychology
Bulletin, 8, 273-279.

Feely, T. H. & Young, M. J. (2000). Self-reported cues about deceptive and truthful
communication: The effects of cognitive capacity and communicator veracity.
Communication Quarterly, 4, 101 - 191.

Gagne, F. M. & Lydon, J. E. (2004). Bias and accuracy in close relationships: An
integrative review. Personality and Social Psychology Review, 8, 322-33 8.

Ganci, D. (1991). The basic essentials of desert survival. Merrillville, IN: [C S Books,
Inc.

Gilbert, D. T. (1991). How mental systems believe. American Psychologist, 46, 107-
119.

Gilbert, D. T., Krull, D. S., & Malone, P. S. (1990). Unbelieving the unbelievable:
Some problems in the rejection of false information. Journal of Personality
and Social Psychology, 59, 601—613.

Gilbert, D. T. & Osborne, R. E. (1989). Thinking backward: Some curable and
incurable consequences of cognitive busyness. Journal of Personality and
Social Psychology, 5 7, 940-949.

Gilbert, D. T., Pelham, B. W., & Krull, D. S. (1988). On cognitive busyness: When
person perceivers meet persons perceived. Journal of Personality and Social
Psychology, 54, 733-740.

Kraus, J. (1978). Alive in the desert. Library of Congress.

Langer, E., Blank, A., & Chanowitz, B. (1978). The mindlessness of ostensibly
thoughtfirl action: The role of “placebic” information in interpersonal
interaction. Journal of Personality and Social Psychology, 30, 63 5-642.

Levine, T. R., Feeley, T., McComack, S. A., Harms, C., & Hughes, M. (2005).
Testing the effects of nonverbal training on deception detection accuracy with
the inclusion of a bogus training control group. Western Journal of
Communication, 69, 203-218.

McComack, S. A. & Levine, T. R (1990b). When lovers become leery: The

relationship between suspicion and accuracy in detecting deception.
Communication Monographs, 5 7, 219-230.

Millar, M. G. & Millar, K. U. (1997). The effects of cognitive capacity and suspicion
on truth bias. Communication Research, 24, 556-570.

34

Nelson, D., & Nelson, S. (1997). Desert survival. Glenwood, NM: Tecolote Press.

Silvera, D. H. (2000). The effects of cognitive load on strategic self-handicapping.
British Journal of Social Psychology, 39, 65-72.

Stiff, J. B. & Miller, G. R. (1986). “Come to think of it. . Interrogative probes,

deceptive communication, and deception detection. Human Communication
Research, 12, 339-358.

Todorov, S., Chaiken, S., & Henderson, M. D. (2002). The heuristic-systematic
model of social information processing. In J. P. Dillard & M. Pfau (Eds) The
Persuasion Handbook: Developments in Theory and Practice, (pp. 195-21 1).
London: Sage.

Weary, G., Vaugh, L. A., Stewart, B. D., & Edwards, J. A. (2004). Adjusting for the
correspondence bias: Effects of causal uncertainty, cognitive busyness, and
causal strength of situational information. Journal of Experimental Social
Psychology, 42, 87-94.

35

 

lllll‘ll”Hilll

3 1293 02956 1523