The Illusion of Control and Personal Beliefs: A University Report

This report details a psychology study conducted at the University of Adelaide investigating the illusion of control (IOC) and its relationship to personal beliefs and superstitions. The study involved 532 participants who completed a computer-based task and an online survey. Participants were randomly assigned to either a naturalistic or analytical condition, and the total number of button presses, subjective ratings of task controllability, and the use of strategies were recorded. Psychological measures included the Drake Beliefs About Chance (DBC) Inventory and the Everyday Illusion of Control measure. The study hypothesized that participants in the naturalistic condition would report greater control, press buttons more frequently, and be more likely to use a strategy. Furthermore, it was hypothesized that higher scores on the belief scales would correlate with greater button presses, perceived controllability, and strategy use. The results showed no significant difference in task controllability ratings or button presses between the conditions, but participants in the naturalistic condition reported feeling more in control. A positive correlation was found between the two belief scales, and the scores on the belief scales were positively correlated with the control ratings. The study provides valuable insights into the cognitive biases that influence decision-making and the role of personal beliefs in shaping perceptions of control.

Running head: ILLUSION OF CONTROL AND PERSONAL BELIEFS 1
The illusion of control: Relationships between people’s beliefs, superstitions and
performance on simple decision-making tasks
The University of Adelaide
2200 words
The illusion of control: Relationships between people’s beliefs, superstitions and
performance on simple decision-making tasks

ILLUSION OF CONTROL AND PERSONAL BELIEFS 2
The illusion of control (IOC) is a bias whereby individuals overestimate their actions
and develop inappropriate expectations of personal success in chance-based situations
(Langer, 1975). Evidence suggests that IOC stems from individuals’ “need for control”,
which can protect self-esteem and psychological well-being (Bandura, 1989; Langer, 1975;
Weiner, 1979). An example of IOC is when individuals believe that they recover from
disease by unnecessary means, just because a certain medication or treatment coincides with
spontaneous remissions (Matute & Blanco, 2014). Another example is when individuals win
lotteries, they believe that they have control over the outcome of the lottery because they
choose the ticket themselves (Langer, 1975). In fact, IOC is generally seen as a factor
strongly related to the aetiology of problem gambling (Joukhador, Blazczynski, &
Maccallum, 2004). Accordingly, modulating factors of IOC should be explored because IOC
might pose a challenge to self-esteem and influence one’s financial, medical and political
decisions.
IOC has been comprehensively investigated in laboratory environments, which
involve simple decision-making tasks where participants press buttons to obtain an outcome
such as stopping a noise. However, the outcomes are pre-determined, such that a percentage
of outcomes occurs anyway regardless of participants’ actions. For instance, Jenkins and
Ward (1965) found that participants enthusiastically searched for strategies to press buttons to
light up a panel and were confident that they would succeed. Similarly, several studies
discovered that participants tended to report that a certain combination of buttons can
generate desired outcomes and perceive a sense of control over the outcomes (Alloy &
Abramson, 1979; Matute, 1996). Specifically, when button pressings are frequently followed
by desired outcomes, people are motivated to exert more control to attain that outcome.
Therefore, they are likely to press the buttons more often and experience IOC (Delfabbro et
al., 2020; Stefan & David, 2013).

ILLUSION OF CONTROL AND PERSONAL BELIEFS 3
The social environment, as emphasised by Langer (1975), is playing a definitive role
in modulating IOC. Instructions can be regarded as an essential feature of the social context
(Glenn, 1987). In this sense, existing research suggests instructions stating that the task might
be based on chance can regulate IOC (Presson & Benassi, 1996). A notable study examining
the influence of instructions was Matute (1996)’s, where participants were allocated to two
conditions and given different sets of instructions. In naturalistic conditions, participants were
instructed to simply obtain an outcome. While in analytical conditions, they were told to
behave scientifically to assess the degree of control they have over the outcome. Results
revealed that participants in the naturalistic condition responded on nearly all trials and
displayed stronger IOC. More recently, research has also provided evidence that instructions
influence both the response probability and perceived controllability (Benvenuti et al., 2018).
Individual differences have also been demonstrated in the development of IOC
(Langer, 1975). Some individuals are more prone to IOC due to the difference in the locus of
control, which is the degree to which individuals believe that they have personal control over
various life events (Rotter, 1966). However, the validity of locus of control instruments
(placing people on an internal-external continuum) has been debated in measuring differences
in people’s susceptibility to IOC (Lefcourt, 1966), because both internal and external control
could arguably be associated with greater IOC. In order to provide a clearer picture of the
link between people’s beliefs and IOC, the present study will utilise two specific measures to
capture people’s beliefs about the nature and causes of events in the world, which are the
Drake Beliefs About Chance (DBC) Inventory (Wood & Clapham, 2005) and Everyday
Illusion of Control measure (Delfabbro et al., 2020). Higher scores on both measures would
associate with stronger IOC. Notably, as Everyday Illusion measure is a newly developed
scale, its correlation with the DBC inventory will be examined; a high correlation would lend
itself to evidence of convergent validity.

ILLUSION OF CONTROL AND PERSONAL BELIEFS 4
This study aimed to investigate the relationship between people’s beliefs and
superstitions and their performance in IOC tasks. From the research examined, it was
hypothesised that:
H1: People in naturalistic conditions would rate the task as more controllable and report more
control than those in analytical conditions.
H2: Number of button presses would be higher among people in naturalistic conditions.
H3: People in naturalistic conditions would be more likely to report using a strategy.
H4: Higher scores on the Drake and Everyday beliefs about control scales would be
associated with greater total button presses, controllability ratings and found among people
who report using a strategy.
H5: Scores on the two belief scales would be positively correlated.
Method
Participants
The sample consisted of 532 participants, including 267 females and 265 males, with
ages ranging from 13 to 78 years, and a mean age of 33.79 years (SD=15.78). The data was
based upon collection by final year psychology students from the University of Adelaide over
a period of two years. Each university student recruited one or two participants to undertake
the experiment and this yielded the total sample size.
Measures
Experimental measures. After participants completed computer-based tasks, the
experimenter recorded the total button presses. Using a scale of 0 to 100, participants were
asked to rate their perception of how controllable the computer task was and how much
control they perceived themselves to have. Participants also reported whether they had a
strategy for pressing the buttons on a yes or no scale.

ILLUSION OF CONTROL AND PERSONAL BELIEFS 5
Psychological measures. An 11-item Superstition subscale from the Drake Beliefs
About Chance Inventory (Wood & Clapham, 2005) was included in the questionnaire. The
scale involved various superstitions that people often have about chance events. Participants
rated statements, such as “I have a special system for picking lottery numbers”, on a 5-point
Likert-type scale (from 1=strong agree to 5=strongly disagree). Items on the Drake Belief
Scale were reverse-scored and summed. The possible total scores ranged from 11 to 55, with
high scores reflecting greater levels of superstition. This scale has good internal reliability
with Cronbach’s alpha varied between .88 and .90 (Delfabbro et al., 2020; Edgerton, Melnyk,
& Roberts, 2015).
IOC was measured by the Everyday Illusion of Control measure, which included 12
items that captured common behaviours in the general population such as driving and simple
gambling tasks (Delfabbro et al., 2020). Items such as “I prefer to choose my own tickets in
raffles” were rated on a 5-point rating scale. The scores on each item were summed to
produce total scores, such that a high score indicated stronger IOC. Cronbach’s alpha for this
scale was .70.
Demographics and gambling behaviour. Participants were asked to indicate their
age, gender and the frequency of gambling behaviour.
Procedure
The study involved two phases: (1) participants completed a computer-based task and
(2) a short online survey. The overall process of the study was illustrated in Figure 1.
Ethics approval from the Human
Research Ethics Committee at the School
of Psychology, University of Adelaide.
Participants completed an
informed consent form

ILLUSION OF CONTROL AND PERSONAL BELIEFS 6
Figure 1. Flow Diagram of the Study Procedure.
Results
Table 1.
Participant 1 was read a set
of Naturalistic instructions
Participant 2 was read a set
of Analytical instructions
Computer-Based task
 Participants were required to press a combination of buttons (A, B
and C) based on the given instructions (Naturalistic or Analytical)
to prevent the trend line from dropping down into the RED or “Lose
Zone”
 The line was plotted automatically every 1.5 seconds until it
finished
 Participants were required to complete a total of 4 plot lines, each
with its own pre-determined pattern
 The number of button presses that appeared to influence the line
movement was recorded
Participants responded to questionnaires online, which included:
(1) Subjective rating of the task
(2) The Drake Belief Scale and Everyday Illusions of Control Scale
(3) Demographics and the frequency of gambling behaviour
(3)

ILLUSION OF CONTROL AND PERSONAL BELIEFS 7
t-test results comparing the ratings of how controllable the task was perceived (task
ratings), how much control the participant perceived themselves to have (control ratings)
and the number of button presses across the two instructional conditions.
Measures
Naturalistic
Condition
Analytical
Condition t-test df p Cohen’s
d
M SD M SD
Task ratings 29.03 27.64 24.51 28.37 1.86 527.85 >.05 .16
Control ratings 28.44 27.37 21.37 26.29 3.04 529.99 <.05 .26
Number of button
presses
469.97 516.51 414.87 477.02 1.27 517.43 >.05 .11
Hypotheses 1 and 2
Hypotheses 1 and 2 were investigated using independent samples t-tests, and the
results are displayed in Table 1. There was no significant difference in the ratings of the task
as more controllable between people in the naturalistic condition and those in the analytical
condition. Furthermore, people who were allocated to the naturalistic condition reported that
they exerted more control in the task than those in the analytical condition. This difference
was statistically significant. Concerning hypothesis 2, there was no significant difference in
the number of button presses between the two instructional conditions. Cohen’s d effect size
for these analyses was small. These results were illustrated in Figure 2.

ILLUSION OF CONTROL AND PERSONAL BELIEFS 8
Figure 2. Mean proportion representing the number of button presses, the rating of how
controllable the task was perceived (task rating) and how much control the participant
perceived themselves to have (control rating) in naturalistic and analytic conditions. Error
bars denote the 95% confidence interval.
Hypothesis 3
A 2x2 Chi-squared test showed that there was no significant association between the
instructional conditions and whether participants chose to use a strategy (X2(1)=3.21, p>.05.
The proportion of participants who chose to use a strategy in the two instructional conditions
was similar, with 58.3% of participants in the naturalistic condition and 50.2% in the
analytical condition reported using a strategy. The effect size is negligible (Cramer’s V=
0.08).

ILLUSION OF CONTROL AND PERSONAL BELIEFS 9
Hypotheses 4 and 5
Table 2.
Correlations between the two belief scales and experimental measures (the number of button
presses and the rating of how much control the participant perceived themselves to have).
1 2
2. Drake Belief Scale Pearson’s r - .49*
df - 184
4. Everyday Illusions Scale Pearson’s r .49* -
df 184 -
6. Number of button presses Pearson’s r .03 .09
df 518 176
8. Control ratings Pearson’s r .39* .28*
df 530 184
Note. * p < .05
As Table 2 shows, Pearson’s correlations revealed a moderate positive relationship
between the scores on the Drake Belief Scale and Everyday Illusions Scale, which was
consistent with hypothesis 5. This relationship is statistically significant. Regarding
hypothesis 4, results indicated that the scores on the two belief scales and experimental
measures are positively correlated. Specifically, there was a negligible correlation between
scores on the two measures and the number of button presses. Furthermore, people who
reported that they exerted more control in the task tended to score higher on the two belief
scales; a moderate correlation for the Drake Belief Scale and a weak correlation for the
Everyday Illusions Scale. These correlations were statistically significant.

ILLUSION OF CONTROL AND PERSONAL BELIEFS 10
Table 3.
t-test results comparing the scores of two belief scales between participants who reported
using a strategy and those who reported no strategy.
Measures
Strategy
t-test df p Cohen’s dYes No
M SD M SD
Drake Belief Scale 25.79 8.63 22.33 8.83 4.55 510.27 <.05 .40
Everyday Illusions Scale 36.44 5.91 33.70 6.59 2.93 161.97 <.05 .44
Additionally, as displayed in Table 3, independent samples t-tests revealed that people
who reported using a strategy scored significantly higher on both Drake Belief and Everyday
Illusions scales than those who reported no strategy. Both have a moderate Cohen’s d effect
size.
Other Analyses
Table 4.
t-test results comparing males and females on scores of two belief scales.
Measures
Males Females
t-test df p Cohen’s dM SD M SD
Drake Belief Scale 22.53 8.82 25.87 8.64 -4.42 529.59 <.05 .38
Everyday Illusions
Scale
35.28 6.65 35.21 6 .08 183.57 >.05 .01
Gender differences in the two belief scales were investigated using an independent
samples t-test and the results are presented in Table 4. Females reported higher levels of

ILLUSION OF CONTROL AND PERSONAL BELIEFS 11
Drake superstition scores than males and the Everyday Illusions of control scores were
comparable across gender.
Discussion
The present study, in contrast to hypotheses 1, 2 and 3, demonstrated that the ratings
of how controllable the task was perceived, the total button presses and the proportion of
people who reported using strategies were similar across the two instructional conditions. The
results contradict the previous research that concluded people in naturalistic conditions score
higher on these experimental measures (Alloy & Abramson, 1979; Jenkins & Ward, 1965;
Matute, 1996). One reason for this is that previous research largely dealt with the estimates of
personal control over the outcome, while this study classified the controllability perception
into two types, which were whether the entire task was perceived as chance-based or skill-
based and the extent to which participants perceived their actions were influencing the
outcome.
Participants might be inclined to provide midpoint responses that are potentially
advantageous (König, Mura, & Schmidt, 2015), irrespective of the instructional conditions.
By rating the task as comprising both chance-based and skill-based elements, they are able to
take credit for successes and attribute failures to chance, enhancing their self-esteem.
Therefore, the task ratings were similar across both conditions. In contrast, participants in the
naturalistic condition perceived greater controllability over the plot movement, suggesting
that hypothesis 1 was partially supported. However, both conditions had similar response
rates (button pressing and strategy searching), which are inconsistent with the explanation
that the response probability is larger among people who perceive themselves to have more
control (Matute, 1996). Accordingly, the difference between controllability ratings over the
task or outcome is of importance. Future research should explore which rating could precisely
represent IOC and predict the response rates in different instructional conditions.

ILLUSION OF CONTROL AND PERSONAL BELIEFS 12
Hypothesis 4, which indicated that the scores on the two belief scales are positively
associated with experimental measures, was supported. This suggests that people who report
higher controllability ratings and likelihood to discern strategy tend to score higher on the
two belief scales. Thus, this finding supports the claim that individual differences play a
relevant role in IOC (Langer, 1975; Rotter, 1966). However, the total button presses were not
significantly related to the scores of belief scales. Earlier studies proposed that the act of
pressing buttons would only increase if it is followed by desired outcomes (Alloy &
Abramson, 1979; Matute, 1996). It is probable that, when desired outcomes are not
perceived, superstitious people blame their luck (Wiseman & Watt, 2004), instead of the
program itself. Consequently, they are not motivated to press the buttons. This situation is
further complicated by the fact that women held more superstitious beliefs than men.
Nevertheless, the present study sample was fairly balanced in terms of gender. The current
study confirms that IOC lies at the core of real-life superstitions and ungrounded beliefs;
thus, people’s beliefs about how events occur in the world can determine their susceptibility
to IOC.
Results also supported hypothesis 5 stating that scores on the Drake belief and
Everyday Illusions scales are correlated well with each other. Therefore, the newly developed
Everyday Illusions scale could be included in individual differences research concerning
IOC. However, more research is needed to comprehensively assess their inter-item
correlations to further validate the scale.
A possible limitation of this study was the lack of ecological validity. The computer-
based task was arguably long and boring and appeared to differ in many ways from real-life
illusions and superstitions. According to cultural studies, superstitions are often negative:
certain behaviours will result in undesired outcomes (Aeschleman, Rosen, & Williams,
2003). The pervasiveness of real-life negative illusions is also manifested in the most widely