The impact of self-set deadlines compared to externally imposed deadlines on task success and accuracy rates

(1)

University of Amsterdam

Faculty of Economics and Business Academic Year 2017-2018

Bachelor Thesis

__________________________________________________________

The impact of self-set deadlines compared to externally imposed

deadlines on task success and accuracy rates

__________________________________________________________________________

Bachelor Economics and Business

Specialization Economics

Name:

Lena Arndt

Student Number: 11103124

Supervisor:

dhr. S.D. (Stephan) Jagau MPhil

(2)

2 Statement of Originality

This document is written by Lena Arndt, who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

3

Introduction

The increasing degree of globalization has opened a door to unforeseen opportunities, but also challenges, including a rising level of international competition. Job structures are changing from highly routine, and supervised jobs to a more autonomous environment, in which employees are required to make independent decisions. Furthermore, work tasks have become more complex and need to be conducted in creative ways, and at a constantly rising pace. In order to keep up with developments and to stay competitive, employees need to be able to understand the potential further innovation brings, and use (technological) advancements in an efficient way (Ebert & Freibichler, 2017). This leaves knowledge worker productivity to be one of the biggest assets, but also challenges of the 21st century (Drucker, 1999), which explains the importance of understanding its influential factors. Amongst others, important indicators of high performance are the success rate, the ability to finish a job within a given time frame, and the task accuracy level, which evaluates the quality of the results. On the one hand, these factors are influenced by the worker’s abilities and values, as well as character traits. On the other hand, performance can be improved by introducing working conditions that enhance desired characteristics and behaviours. This is why various people from different backgrounds, including behavioural economists, university professors, and company leaders, are interested to learn which circumstances positively affect one’s task success and accuracy rate. One influential factor, which has been debated extensively in literature, is the impact of deadlines on people’s performance. Two important determinants are consistently mentioned as mediators of the relationship between deadlines and the task success and accuracy rates. These are one’s procrastination degree and the level of intrinsic motivation. While procrastination is a suboptimal performance reducing behaviour that should be avoided or limited, promoting intrinsic motivation is positively related to the performance indicators. In general, findings of previous researches tend to support deadlines as a productivity enhancing measure, as they are commitment devices that limit and therefore reduce suboptimal behaviours like procrastination (Cadena, Schoar, Cristea, & Delgado-Medrano, 2011). However, opinions diverge which type of deadline is the most effective concerning task success and accuracy rates. When comparing internal, self-set deadlines to externally imposed ones specifically, previous findings in literature report diverging effects on performance while considering corresponding procrastination and intrinsic motivation levels.

Overall, previous papers so far have identified indicators, influences, and aspects of procrastination, as well as intrinsic and extrinsic motivation. Furthermore, various links have been created between determinants, including autonomy, commitment, time-smoothing, and the corresponding procrastination or intrinsic motivation level. Thus, the impact of internal compared to external deadlines regarding separate influencing factors, their relation to the

(5)

5 mediators, and corresponding performance levels have been explained. Nonetheless, the established links are not sufficient yet to describe the overall effect of internal and external deadlines on task success and accuracy. A trade-off is found between the performance enhancing influence of internal deadlines on commitment and intrinsic motivation, in contrast to the negative effect of imperfect time-smoothing and possibly higher levels of procrastination. This trade-off has led to diverging results in previous research. With that in mind, this paper aims to introduce an experiment designed to explain the impact of self-imposed deadlines compared to external ones on accuracy and success rates of a task. In order to approach this research question, the paper is structured as follows.

First, findings in previous literature regarding this topic will be summarized. The main hypothesis of this paper, states that the accuracy and success rate of a task is higher under internal deadlines than external deadlines. Amongst others, this is due to the aspect of choice, which increases the level of self-determination and self-perceived autonomy, as well as the effort a person is willing to exert in a task. Higher effort levels then lead to higher performance. The main hypothesis of this paper furthermore relies on two sub-hypotheses that explain underlying connections between internal versus external deadlines on performance indicators regarding procrastination and intrinsic motivation. The first sub-hypothesis states that present biased decision makers, who are aware of their procrastination problem, show a higher accuracy and success rate under internal deadlines compared to external ones. This part of the literature review explains the overall procrastination phenomenon, and its influencing factors. It further shows that findings in previous research regarding the effect of self-imposed deadlines compared to external time limits on a procrastinator’s performance are divided. On the one hand, task success and accuracy rates are negatively related to internal deadlines, as they induce imperfect time smoothing. On the other hand, if a person is given the option to self-set a deadline, a higher level of commitment is expected, which enhances performance characteristics. This trade-off explains the ambiguous results regarding the hypothesis in previous literature. The second sub-hypothesis claims that the level of intrinsic motivation is higher under internal deadlines than external ones. Based on previous research results, the most influential factor supporting this hypothesis is a person’s increased level of perceived autonomy. By being able to make a decision regarding the task, like its deadline, people feel more connected to it. This leads to higher enjoyment and corresponding intrinsic motivation. Lastly, in the final part of the literature review, attention is drawn to the fact that an evaluation, regarding the overall impact of the separate factors and the trade-off they create, is still missing.

In order to fill the missing gap in literature, an experiment outline is constructed in the second part of the paper. It is designed to be conducted at the CREED lab of the Economics

(6)

6 department at the University of Amsterdam. Two different types of sessions are created, in which the participants need to solve the same task, but face either self-imposed deadlines or external time limits. Their task success and accuracy rates subject to the different conditions will be analysed. In contrast to previous research conducted, the experiment will aim to explain the overall connection between the deadline types and the performance indicators, while keeping in mind the mediating influence of procrastination and intrinsic motivation. Moreover, whereas most literature is focused on explaining factors that influence result quality, the experiment gives opportunity to consider the impact on task success and task accuracy rates separately.

Lastly, concluding remarks will be given. They summarize the ideas and strategies that underline the developed experiment design, possible limitations, and suggest ways to continue this research to explain the overall impact of the deadline types on performance.

Literature review

There has been an ongoing debate about the impact of deadlines, external compared to internal deadlines specifically, on workers’ performance. Based on previous literature, this section aims to explain the effect of deadline types on the procrastination level and intrinsic motivation in the work and academic environment. It will further point out that ambiguous results have been found regarding their overall impact on task completion and success. Based on described outcomes in literature, one main hypothesis and two sub-hypotheses have been created. In the following text the underlying connections that have led to a corresponding hypothesis will be explained after the hypothesis has been stated.

While few researchers, like Bisin and Hyndman (2014), argued that the presence of deadlines does not increase task completion rates, many have found that they improve a worker’s performance significantly (Ariely & Wertenbroch, 2002; Cadena, Schoar, Cristea, & Delgado-Medrano, 2011). However, there is disagreement when it comes to evaluating the effect of different deadlines types on the probability of accurate and successful task fulfilment. The impact of two different kinds of deadlines in particular has been debated recently, namely the influence of internal compared to external deadlines on a person’s performance. Internal deadlines describe time limits that are self-imposed, while external ones are set by an outside agent.

(7)

7 Hypothesis 1: The accuracy and success rate of a task is higher under internal deadlines than external deadlines.

Burgess, Enzle, and Rodney Schmaltz (2004) explained that there are different reasons to impose an internal deadline. On the one hand, a person may appoint sub-internal deadlines to himself to reach an external deadline stressfully. On the other hand, he may impose an internal time limit that is more stringent than the given external one. However, in order to create a direct comparison of their effects, this paper does not consider cases where both deadlines are imposed on the same person. A simple example of an internal deadline could therefore be represented by a student picking his own essay deadline, instead of his professor choosing the day it needs to be handed in (an example of an external deadline). Internal deadlines thus provide a choice, while external ones do not. To explain the impact of internally set time-limits, it is therefore important to understand how the aspect of choice influences a person’s behaviour, as well as their corresponding productivity and performance level. The ability to choose between options is a simple and effective way to strengthen one’s self-perceived autonomy (Meng & Ma, 2015). According to the self-determination theory, which states that individuals have three psychological needs including competence, relatedness, and autonomy, higher perceived independence results in an increased effort level when completing a task. The invested amount of effort on the other hand is positively correlated to the success level of demanding tasks. This shows that the aspect of choice and the corresponding higher degree of autonomy can result in an improved work performance (Meng & Ma, 2015). Thus, considering the impact of choice on self-determination, internally imposed deadlines are expected to enhance the accuracy and success rate of tasks. Burgess et al (2004) provided further arguments in favour of internal deadlines by naming performance diminishing effects of externally imposed ones. Firstly, they suggested that external deadlines cause a higher degree of pressure. This has several negative indications, including an increased stress level, emotional exhaustion, and inadequate strategies, as people are compelled to finish the task on time, regardless of other circumstances and rules. Moreover, Burgess et al (2004) stated that it can cause a lack of concentration, as people analyse the available information less critically. Overall, external deadlines can reduce the quality of results, which is indicated by lower accuracy rates.

Besides the positive impact of choice explained by the self-determination theory and the disadvantages of externally imposed deadlines due to increased pressure, the influence of two aspects on successful task completion are discussed extensively in literature. On the one hand, the level of procrastination is named as an important factor. Researchers agree that it is considered a performance diminishing trait (Brunnermeier, Papakonstantinou, & Parker, 2007; Gupta, Hershey & Gaur, 2012; Kaur, Kremer, & Mullainathan, 2010). Procrastination is seen

(8)

8 as a personal characteristic, that cannot be prevented and seldom influenced. However, several findings in literature indicate that certain task and workplace characteristics can enhance productivity of people with procrastination characteristics (Bisin & Hyndman, 2009; Lonergan & Maher, 2000; Van Eerde, 2003). A second relevant influence that explains the correlation between the different deadline types and work accomplishments, is the degree of intrinsic motivation. Overall, higher perceived autonomy induced by internal deadlines increases intrinsic motivation, which in return results in an improved performance level. As there are various factors that influence procrastination and intrinsic motivation, as well as the corresponding task success, two sub-hypotheses are used to explain the complicated underlying connections in more detail.

Hypothesis 1.1: Present biased decision makers, that are aware of their procrastination problem, show a higher accuracy and success rate under internal deadlines compared to external deadlines.

Procrastination can arise on many different domains, which include academic and work situations, but also leisure and family aspects of life (Lonergan & Maher, 2000). It is considered a form of self-regulatory failure, where an intended course of action is postponed voluntarily despite the expectation of being worse off due to the delay (Ngyuen, Steel, & Ferrari, 2013). It arises because people do not discount the future in a time consistent manner (Fischer, 2001). Instead, due to a disproportionate, present-bias effect on preferences, immediately available rewards are favoured over delayed ones (Ariely, & Wertenbroch, 2002). Therefore, procrastinators tend to focus on short-term objectives rather than using their available time and energy for more important long-term goals (Gupta et al, 2012). Moreover, when they are faced with an unpleasant task, present-biased decision makers underestimate the time necessary to complete it (Brunnermeier et al, 2007). This is often shown by missed deadlines. Nonetheless, Fischer (2001) explained that procrastination is not always characterized by missing deadlines or abandoning tasks. Instead, it can exhibit itself in an increasing workload towards the due date, as most actions are postponed until then. On the one hand, procrastination can be seen as an avoidance coping reaction (Van Eerde, 2003). By using a distraction, a short-termed action that is not only less important but also more pleasurable, emotional distress is lowered. However, Gupta et al (2012) argued that procrastination is mostly considered to be a disproved behaviour, which reduces the time people have to carefully evaluate a task and its outcome, thus leading to inefficiencies and suboptimal performance. Based on this argument they further reasoned that procrastination is seen as a foolish and harmful trait, which should be prevented or diminished. Not only can it lead to people falling short of their goals (Kaur et al, 2010), but also causes a vicious cycle, that increases time pressure and time famine (Van Eerde, 2003).

(9)

9 As procrastinators put off working on a task until shortly before its deadline, the workload becomes too high while the time to complete it becomes too little. Metin, Taris, & Peeters (2016) suggested that this leads to psychological discomfort, stress, poor health, and negative emotions.

Factors that influence the procrastination degree can be grouped in three dimensions. Firstly, there are intrapersonal aspects, which include personality characteristics. Through previous experiments conducted, researchers found a strong negative correlation of conscientiousness (Metin et al, 2016; Van Eerde, 2003), self-efficiency, and emotional stability to procrastination (Van Eerde, 2003). Moreover, Miao (2008) found that procrastination can be caused by temptation and low self-control. On a second dimension, the behaviour can be influenced by situational factors. A complicated family event could therefore increase the procrastination level temporarily (Gupta et al, 2012). Lastly, task and workplace characteristics have a significant impact on postponing actions. Kaur et al (2010) suggested that, in order to reduce procrastination at the workplace, immediate costs and benefits of an action need to be altered. Hereby, rewards for high effort expended to the task need to become more immediate, while penalties for laziness and low output should be introduced. In general, recent findings have shown that job enrichment can lower procrastination behaviour at the workplace (Lonergan & Maher, 2000; Metin et al, 2016). Job enrichment is defined by available work resources and demands. Job demands describe the physical and psychological effort or skills that are required for the completion of work tasks. Job resources include the aspects that promote work goals, which stimulate personal growth and development (Metin et al, 2016). Examples of job resources include variety, identity, and significance of the task, as well as feedback and autonomy. Little work demands and resources are connected to low motivation, frustration, resentment, and high levels of boredom at work, which are strongly associated with a raised degree of procrastination (Metin et al, 2016; Nguyen et al, 2013). On the contrary, an increase in a job resource can reduce this suboptimal behaviour. Lonergan and Maher (2000) stated that autonomy, the discretion in organizing and scheduling work, is an essential job characteristic that can increase job satisfaction, commitment, motivation, and performance. Moreover, their findings suggested that greater independence and its influence on other job components leads to higher performance of procrastinators at the workplace. However, Lonergan and Maher also admitted that highly autonomous jobs, that let their employees schedule their work and set their own deadlines, may give more room for procrastination because of missing supervision. If not supervised externally, a present biased person can try internal ways to overcome their behavioural problem. Firstly, he can improve his willpower to resist own temptations (Miao, 2008). Also, he can pre-commit to own behaviour by setting voluntary constraints on to avoid giving in to his future temptations (Ariely & Wertenbroch, 2002). Even though a rational economic agent with time-consistent preferences would not

(10)

10 impose constraints on his own choices, present-biased decision makers, that are aware of their procrastination problems, tend to seek binding deadlines (Bisin & Hyndman, 2014; Brunnermeier et al, 2007; Kaur et al, 2010). As they understand the benefit of immediately completing a task, they use commitment devices to start exerting effort right away. These commitment tools can be adapted through internal or external mechanisms. Bisin and Hyndman (2014) named examples of internal mechanisms, which include anticipatory planning, and external ones, like binding self-imposed deadlines or voluntary exposure to social pressure.

After analysing the procrastination phenomenon as well as its influencing factors in depth, the expected impact of internal compared to external deadlines is now comprehensible. As explained previously, commitment has been proven a useful device to reduce procrastinating behaviour (Cadena et al, 2011). Cadena et al (2011) explained that regardless of the type of deadline imposed, performance increases compared to open-ended tasks. However, they stated that people tend to set deadlines for themselves that require less than perfect smoothing of work over time. According to their findings, external deadlines thus have a higher potential to increase one’s performance level than self-imposed ones. Furthermore, Cadena et al explained that imperfect work allocation over time is a type of planning fallacy, as it leads to suboptimal decisions as well as outcomes. Indeed, findings of previous researches conducted suggest that students show a greater degree of procrastination when they set their own deadlines, compared to professor-paced classes (Lonergan & Maher, 2000). An experiment carried out by Ariely and Wertenbroch (2002) furthermore demonstrated that the provided flexibility can lead to lower grades instead of enhanced performance, if a student with self-control problems does not set optimal time-smoothing deadlines for himself. Ariely and Wertenbroch (2002) found that the best results are achieved with external deadlines, under the condition that they are optimal. Despite the risk of choosing a suboptimal deadline for oneself, the option to self-impose has overall proven to reduce one’s tendency to procrastinate (Lonergan & Maher, 2000). Even though the time smoothing under this condition may be less than perfect, internally set time limits are considered to be more binding than the external final deadlines, especially if a person’s willingness to commit is high (Bisin & Hyndman, 2009). Bisin and Hyndman (2014), further explained that self-imposed time limits are seen as a stronger commitment device, while external ones are perceived as an unchangeable circumstance controlled by outside parties. In order to cope with procrastination, measures that can make a task less threatening and more pleasant are needed. Out of the two evaluated deadline types, externally imposed ones make an assignment seem more demanding, and are therefore likely to increase stress and time pressure (Van Eerde, 2003). As a result, present biased workers specifically struggle with lower concentration levels and quality of work. Internally imposed deadlines on the other hand, can make a task seem more

(11)

11 pleasant through increased flexibility and autonomy. As explained previously, autonomy is considered a valuable job resource that can reduce one’s level of procrastination and therefore improve the overall work performance. As Bisin and Hyndman (2009) summarized, self-set goals can improve the result quality of procrastinators, if the present bias is large enough. Thus, it can be anticipated that present-biased decision makers show a higher task accuracy rate under internal compared to external deadlines. However, no previous researches have been conducted that enable the performance evaluation regarding task success and accuracy rate specifically and separately. Moreover, the trade-off between imperfect time-smoothing and commitment regarding the two performance indicators has not been assessed in further detail. The later designed experiment will fill this gap by giving an opportunity to evaluate the overall correlation of deadline types to procrastination, as well as task success and accuracy rates separately.

Hypothesis 1.2: The level of intrinsic motivation is higher under internal deadlines than external deadlines.

Intrinsic motivation is defined by performing an activity because of the enjoyment and satisfaction it contributes. The inspiration for acting can be found in the exercise itself. This stands in contrast to extrinsic motivation, where an action is carried out to achieve a desired external outcome, for example a salary. Intrinsic motivation has an important impact on achieving productive outcomes (Meng & Ma, 2015). Burgess et al (2004) argued that as intrinsically motivated people perceive a task as more interesting, they initiate activity engagement leading to higher productivity levels. In contrast, low intrinsically motivated people may not only grow disinterested in a job, but also in the subsequent performance feedback. Therefore, low levels of self-encouragement often lead to lower success rates and quality of work results (Meng & Ma, 2015). Furthermore, intrinsic motivation impacts procrastination behaviour. Nguyen et al (2013) found that procrastinators tend to have jobs that are lower in intrinsically rewarding qualities, while professions and tasks that require higher level of motivational skills are less likely to retain people with the behavioural problem. These results show that self-inspiring job characteristics can reduce procrastination while improving overall work performance. Nguyen et al (2013) supported this argumentation by stating that employee’s work strivings increase when individuals positively interpret the worthiness and meaningfulness of their job. Characteristics that enable intrinsic motivation to perform well include job clarity, assignment specification, interesting, and challenging tasks, as well as the ability to make decisions (Metin et al, 2016). Meng and Ma (2015) stated that autonomy, and corresponding decision-making responsibility, is the strongest predictor of intrinsic motivation and likely to enhance it.

(12)

12 When analysing the effect of the two different deadline types on intrinsic motivation, there is strong evidence that internal time limits will have a stronger positive impact than external ones. Burgess et al (2004) explained that external deadlines can pollute people’s interest in previously valued activities. Ariely and Wertenbroch (2002) further found that if people have to complete a task within a time frame imposed by outside agents, they tend to dislike the exercise more, than if they had the option to choose a deadline by themselves. Moreover, previous results in literature indicate a trade-off between extrinsic contingencies, which include rewards like money, but also constraints like deadlines, and intrinsic motivation. This proves that too high external incentives can undermine own ambitions and inspirations (Burgess et al, 2004; Meng & Ma, 2015). The results of an experiment conducted by Burgess et al (2004) indicated that people, who had the option to impose a deadline internally, spent more free-time choice engaged in the target task than their externally imposed counterparts. Providing a choice through letting people pick their preferred deadline is a simple and effective way to strengthen self-perceived autonomy (Meng & Ma, 2015). As intrinsic motivation is expected to be enhanced by events that encourage perceptions of personal independence (Burgess et al, 2004), it is anticipated that the completion of a task will bring more enjoyment under internal deadlines than external ones.

The previous literature review has summarized preceding research outcomes on the connection of internal compared to external deadlines to procrastination and intrinsic motivation, as well as general performance. Separate links regarding various influencing factors have been created, based on prior experiments conducted. However, there is an overall trade-off between the performance enhancing influence of internal deadlines on commitment and intrinsic motivation, in contrast to the negative effect of imperfect time-smoothing and possibly higher levels of procrastination. No research has been conducted yet that aims to explain the overall connection between the deadline types and performance indicators regarding that trade-off. This is why in the following section an experiment design is developed that enables a precise analysis of the previously stated hypotheses, thus closing a gap in literature.

(13)

13

Experiment Design

1. Participants and general design

The experiment is run at the CREED lab of the Faculty of Economics, University of Amsterdam. Its goal is to understand the impact of internal compared to external deadlines on the accuracy and success rate of a task. In the experiment, the exercise is represented by a pure-effort task used in previous experiments conducted (Weber & Schram, 2016). Hereby, participants are faced with two matrices that contain 100 random numbers valued between 1-100. In order to solve this exercise, students need to determine the highest value in each matrix, add them without using calculators, and report the sum in a box beneath the matrices. An example of the task is given below.

These matrices can be seen as pure effort tasks. This means that, even though a high level of mathematical literacy may be advantageous, solving them is possible for all participants with different levels of logical abilities, as long as they are willing to exert effort into the exercise. During the experiment, the students need to solve 30 matrices within a given time frame. Depending on which session they attend, participants will randomly either face an internally or externally imposed deadline. As part of the “internal group” (group 1), students will be able to choose their preferred time frame out of three given options. Statistically speaking, it is expected that each option will be chosen by one third of the participants. In order to create a sufficiently high-powered study, a minimum of 40 participants in each internal subgroup will be needed. Participants belonging to the “external group” (group 2) are randomly assigned to one of the three deadline categories (probability: 1/3 for each option), but can indicate which alternative they would have preferred at the end of the experiment. Group 2 is therefore divided into nine subcategories. Out of these, the three subcategories, where students indicated that they would have chosen their externally imposed deadline, need to have a minimum number of 40 participants. This number is required to evaluate the

(14)

14 hypotheses in a statistically significant way, as the task accuracy and success rates of these external deadline subgroups will be directly compared to the ones of the corresponding internal deadline categories. Assuming equally split probabilities, 480 participants, out of which 120 would be in the internal group, and 360 in the external one, are needed to fulfil this condition. However, based on the pilot results which could show an unequal probability distribution, this number may increase. Students receive 5€ as a show-up fee, as well as additional 10€ if they manage to successfully complete the task before their deadline There will be no minimum required accuracy rate. This will not be stated explicitly to avoid giving a wrong signal to the participants, as most would expect that the correctness of their answers matters. However, some students may become aware of the missing condition. It can be anticipated that strongly extrinsically motivated participants will then aim to finish the task as quickly as possible, and will not be bothered by the outcome of their answers. This is why performance will not only be measured by a participant’s ability to finish the task on time, but also by the proportion of correctly solved exercises. It can be predicted ex-ante that a potential higher success rate of this kind of participants will occur at the expense of their accuracy rate. The length of the experiment depends on the individually chosen or externally imposed deadline, and thus varies for each participant. In their conducted experiment, Weber and Schram (2016) showed that the participants managed to solve about 5 matrices in 8 minutes on average. Based on this result, it is expected that the average participant would need 48 minutes to complete 30 exercises. The corresponding deadline options are picked accordingly. One session furthermore includes an introduction phase with a maximal amount of 10 minutes, and a short questionnaire and pay-out stage which lasts about 10 minutes as well. Depending on the imposed deadline option, one session is thus estimated to take between 45 to 75 minutes overall.

A pilot session with 21 participants will be run for the internal deadline group to approximate the overall cost to conduct the entire experiment. This will be done by testing whether estimations regarding the length of the experiment, and the chosen deadlines are accurate. It will demonstrate if the three different deadline options are chosen equally, which would be in line with the statistical expectation, or whether one alternative is strongly preferred or disliked compared to the others. The outcome of the pilot session can lead to adjustments in the general design of the experiment, including the number of participants needed. Under the assumption that participants distribute themselves over the subgroups equally, it is estimated that 480 students will be needed. This would sum up to six sessions of 20 students each that are imposed to internal deadlines, while 18 sessions would be run where the participants will be given

(15)

15 a deadline externally. However, in order to ensure that the minimum requirement of 40 people per category if fulfilled, there is a possibility that this number will rise, if the pilot shows that participants prefer one deadline option over others. Additionally, results from the pilot may suggest further improvement measures and indicate a pattern in the task failure rate before the actual experiment is conducted.

2. Procedure

a. Instruction stage

The instruction phase will be the same for all participants. Firstly, students are given general information and directions on the experiment. They are made aware that they are not allowed to communicate with each other in a verbal or computerized way, and that they need to hand in their cell phones, which will be locked away until the end of the experiment. Furthermore, the goal of the experiment, solving 30 pure-effort matrices within a designated time limit, is described, and the students are informed of the pay-off structure. In the following step, the participants are introduced to the matrix tasks. By using an illustrating example, the correct way to solve the exercise is explained. Students further receive information on the average time people are given to solve 30 matrices, which is estimated at slightly above 45 minutes (Weber & Schram, 2016). After becoming acquainted with the task in theory, participants will be given two example matrices. Hereby it will be ensured that they understand how to solve the task, and how much time it takes them. After submitting a solution to one practice exercise, students will receive feedback if their answer is correct. This ensures that all participants understand the way the matrices need to be solved. However, students are also made aware that they will not receive any feedback on the accuracy of their results during the test stage, encouraging them to assure that they truly understand the exercise.

b. Test stage

In the second stage of the experiment the effect of internal compared to external deadlines on the success and accuracy of the task will be tested. In a preliminary amount of six sessions (that may change based on pilot results) students will have the option to choose their preferred time limit, while 18 sessions will be run where student are randomly, but equally assigned to one the three alternatives.

Self-imposed deadlines: Participants belonging to the “internal group” need to

(16)

16 choosing between the suggested competitive deadlines of 25, 35, and 45 minutes. They are aware that inability to finish all exercises on time will result in a lower payment. From an extrinsic motivational point of view, rational participants would thus choose to work for 45 minutes, as it is most likely to complete the task under this condition. However, the instructional page further notes that during the time the participant has chosen to work on the designated task, he is not allowed to use his mobile phone and has to devote all his attention to the exercise. Inability to predict the right amount of time needed to solve all matrices and finishing early would therefore result in waiting around in the CREED lab. This will be emphasized as a situation that the participants want to avoid. It is expected ex-ante that students will perceive waiting time as more tedious and boring than solving the exercises. This gives students an incentive to pick a more challenging deadline close to their estimated matrix skill level. After all participants have chosen their preferred time limit, they can start working on the exercises. Students will be presented with one task at a time, and can only get to the next one by solving the previous. Once submitted, an answer cannot be reviewed or changed. The participants do not receive feedback whether their solutions are correct - their task success rate is solely announced at the end of the experiment. Regardless of the speed at which the students solve the matrices, they are finished once their chosen time limit is reached. At that point they will be redirected to a new page that shows their percentage of correctly solved exercises and states whether or not they managed to successfully complete the task. After filling out a short questionnaire, participants are diverted to a final page that notifies them of the end of their experimental session. It also mentions that their information and results are being processed for about five more minutes, and that students should stay at their designated computer until instructed otherwise. They are offered the option to spend their waiting time with solving further matrices, that, in contrast to the previous procedure, give immediate feedback about the accuracy of an answer. The participants are uninformed that this waiting period is a part of the experiment. Moreover, students are unaware that the time they spend on further exercises instead of doing nothing is measured. Since there are no external incentives for them to continue the task, a preference for solving further exercises can be explained by intrinsic motivation, for example the will to test oneself or a general interest in the task. After five minutes of waiting time, participants’ table numbers are called out, they receive their payment and are free to leave the lab.

(17)

17

Externally-imposed deadlines: Participants of the external deadline sessions

face the same experimental procedure as the internal groups overall, with one defining difference. Instead of being able to choose a preferred alternative, students are randomly, with probability 1/3, assigned to a time frame of either 25, 35, or 45 minutes. Like the members of group 1, the external deadline participants are not allowed to leave the lab before their given time has run out. After being informed about their task success and accuracy rate, they also have the option to spend their end-of-experiment waiting time with solving additional matrices.

c. Questionnaire

To account for some personal characteristics of the participants in the result analysis, all students need to fill out a short questionnaire at the end of the experiment. Firstly, demographic characteristics regarding the student’s gender, age, nationality, field of study, and whether they have done an experiment in lab before are collected, as potential controls for the results analysis. Then, in order to be able to control for the aspect of choice in the result analysis, participants of the external deadline sessions are asked if they would have chosen a time frame of 25, 35, or 45 minutes is they had been given the option. Moreover, an important goal of the questionnaire is to evaluate the student’s tendency to procrastinate. The Academic Procrastination Scale–Short Form (APS- S), which has been validated by Yockey (2016), is used for this. For the five statements pictured below, participants need to indicate on a five- point scale how much they agree with the statement (1= Disagree to 5= Agree).

(18)

18 Furthermore, a sixth statement will be added to the questionnaire, where students can indicate how much they enjoyed the task in the experiment (1=did not enjoy to 5=enjoyed).

Lastly, they are given the opportunity to give an own comment or suggestion on the experiment.

3. Hypotheses

Here I explain how the data gathered through the previously designed experiment will be analyzed statistically in order to evaluate the experimental hypotheses.

Hypothesis 1: The accuracy and success rate of the task is higher under internal deadlines than external deadlines.

In order to evaluate this hypothesis, the success rate and level of result accuracy of the internal deadline session compared to the external one need to be analyzed and their significance tested statistically. To achieve meaningful results, the impact of internal vs external deadlines needs to be isolated by accounting for other result affecting factors. In the experiment, it is expected that students who are appointed to a more stringent deadline are less likely to complete the task on time. Therefore, it is only useful to compare the performance of participants appointed to the same time limit when evaluating the effect of the different deadline types. Moreover, variation in performance may occur due to differing logical and cognitive abilities of participants. To control for this, students that chose time limit x are matched with the participants that were externally imposed the same deadline, but also indicated that they would have chosen it in the questionnaire. By putting them into different categories and subcategories, it is more likely that the performance level of students with similar mathematical skills will be compared.

The performance level of participants will be evaluated on two accounts. Firstly, the success rate, the ability to finish a task by its designated deadline, is a positive performance sign. Therefore, the proportion of group members that managed to finish early or on time are compared within their controlled categories. A standard Z-test weighs H0: pOT1-pOT2=0 against H1: pOT1-pOT2<0. A significant p-value would show that

people are more likely to complete a task on time when they are facing an internal deadline. This result is expected, as previous findings in literature have shown that the aspect of choice created by imposed deadlines leads to a higher degree of self-determination due to increased levels of autonomy. Higher perceived autonomy has a positive impact on the success rate of the result, since it motivates people and affects their sense of responsibility to deliver a result on time.

(19)

19 Secondly, the proportion of correct answers of the number of submitted results is evaluated among the different control groups of the external and internal deadline sessions. In line with the above stated hypothesis, a standard Z-test would test H0: p1

-p2=0 against H1: p1-p2<0. Significant results would indicate that under internal

deadlines, there is a tendency to complete tasks with a higher level of concentration and care. One the one hand, this can be explained by the fact that external deadlines induce a higher amount of stress leading to lower levels of concentration. On the other hand, it is expected that internal deadlines enhance intrinsic motivation (tested in Hypothesis 1.2). This means that people care more about the quality of their outcomes, which results in higher task accuracy rates.

Moreover, the same two proportions tests explained above will be conducted, comparing the overall task success and accuracy rates for the entire internal group to all of the external session participants. The results from these tests will furthermore be analysed in contrast to previous outcomes to asses the significance of the used controls.

Test comparing all participants of the different session types:

Internal deadline session External deadline session Success proportion pSI pSE

Accuracy proportion pAI pAE

Hypothesis tests Success rate: H0: pSI-pSE=0 vs H1: pSI-pSE<0

Accuracy rate: H0: pAI-pAE=0 vs H1: pAI-pAI<0

(20)

20 Hypothesis 1.1: Present biased decision makers, that are aware of their procrastination problem, show a higher accuracy and success rate under internal deadlines compared to external deadlines.

While the different control groups remain as described for the previous hypothesis, a linear Ordinary Least Squares (OLS) regression will be run. This is done to evaluate the correlation between one’s degree of procrastination to the task success and accuracy rates under internal compared to external deadlines. Hereby, the restricted number of participants resulting from the previously explained chosen control groups will be used. Hence, students that chose time limit x are matched with the participants that were externally imposed the same deadline, but also indicated that they would have chosen it in the questionnaire. Since the correlation to and effects on task success and accuracy will be evaluated separately, two OLS regressions for each control group will be run, resulting in a number of six regressions overall. Task success and task accuracy each are the dependent variables in three regressions. The task success is a binary variable, where 1 represents the ability of a participant to finish all exercises on time, whereas 0 indicates that they have failed to make the deadline. It can take any value between 0 and 1, which represents the probability that the variable is 1. The task accuracy rate on the other side is a continuous variable, that shows the proportion of correctly solved exercised regarding the overall number of answers submitted. It can also take any value between 0 and 1, but in contrast to the task success rate, an interpretation regarding probabilities does not exist. Instead of looking at the average values that the dependent variables may reach, the main focus of the OLS regression will be on corresponding correlations between independent and dependent variables, as well as the influence of control factors. One important independent variable is the participant’s level of procrastination. Present-biased decision makers are expected to make up 80-95% of the students (Bisin & Hyndman, 2014). In order to determine the student’s procrastination degree, his answers to the Academic Procrastination Scale–Short Form in the questionnaire will be evaluated. Their procrastination level will then be calculated as follows: PL= 0.82*x1+ 0.78*x2+

0.73*x3+ 0.85*x4+ 0.86*x5, PL can range from 4.04 up to 20.2. This value, as well as

corresponding task success and accuracy rates, are used to estimate the correlation of the procrastination level to the dependent variable. However, several variables will be used to control this relation. Firstly, a binary variable is introduced that controls for the deadline type. 1 will hereby indicate an internally imposed time limit, while 0 stands for no internal, hence external, deadline. Furthermore, other added variables will control for the participant’s demographic information, collected in the questionnaire.

(21)

21 Regressions will be run to evaluate the correlations to the dependent variables, as well as well as connections between the independent and control variables. Significant correlations could then be analysed and explained in more detail. In line with the hypothesis, it is expected that under internal deadlines the correlation between procrastination level and task success as well as accuracy rate would give a lower negative number (closer to 0) than under external deadlines. Overall, as the hypothesis states, it is expected that procrastinators will be more likely to successfully complete a task on time under internal deadlines, as they see them as a stronger commitment device. It is also anticipated that they achieve higher levels of accuracy, because the higher perceived autonomy will motivate them to complete the task with more care. However, as previous findings in literature suggest, it is possible that internal deadlines will lead to imperfect time smoothing and therefore enhance procrastination behaviour. This in return could lead to lower task success and accuracy rates. As there have been no previous experiments in literature conducted to explicitly explain the trade-off between a higher sense of commitment but a suboptimal way of time-smoothing, it is possible that ambiguous results will occur. Significant correlations that support the above stated hypothesis on the other hand, could give a first indication as to which factor has a stronger influence on procrastinators and their overall performance.

Hypothesis 1.2: The level of intrinsic motivation is higher under internal deadlines than external deadlines.

This hypothesis can be evaluated using two different results from the experiment. Firstly, it can be judged by comparing the answers to statement 6 in the questionnaire by the internal and external deadline participants. The average reply to the assertion can be calculated for each group, and assessed statistically by a standard one-sided t-test. In line with the above mentioned hypothesis, H0: µq1-µq2=0 will be

tested against H1: µq1-µq2>0. Significant results would show that participants enjoyed

the task more during internal deadlines sessions, which can indicate that they also had a higher level of intrinsic motivation when solving the exercises.

Secondly, the level of intrinsic motivation during this experiment can be evaluated by assessing the time students spent on solving additional matrices during the waiting period at the end of the experiment. Here, the averages of the two groups are compared statistically by testing H0: µt1-µt2=0 against H1: µt1-µt2>0 using a standard

t-test. Significant results would explain that internal deadlines make a task seem less tedious than external ones, leading to a higher intrinsic motivation to do more exercises

(22)

22 in one’s “free time”. This is anticipated because internally imposed deadlines lead to increased observed autonomy levels. Since participants perceive that their opinion regarding the way they solve the task matters, it is anticipated that they feel more connected to it. This leads to higher levels of intrinsic motivation, that can be indicated by a positive response to the task, as evaluated in the first test conducted, or students spending their “free time” on the matrices instead of waiting, which is analysed in the second test.

In order to ensure that all information gathered in this experiment is used most effectively, it is useful to conduct an OLS regression to analyse overall correlations between internal versus external deadlines, intrinsic motivation, and task success and accuracy rates respectively. As explained previously, task success and accuracy are the chosen dependent variables. Six regressions would be conducted because of the different time limit control groups, since the aspect of time and mathematical ability can create a bias in the analysed correlations, if the entire pool of participants is used. The main independent variable is one’s level of intrinsic motivation, which is shown by the average “free time” that the participant spent on solving further matrices during the waiting time. However, it may be useful to also run separate regressions using the number indicated in the questionnaire statement as the determinant for intrinsic motivation. The value that gives more significant results could then be used for the analysis. Control variables are added to the OLS regression, which include a binary variable for the deadline type, as well as others, which account for the participant’s demographics. According to findings in previous literature, it is expected that the level of intrinsic motivation is significantly positively correlated to the task success as well as accuracy rate. This is because an intrinsically motivated person feels committed to a task and cares more about its outcome. Thus, he is more likely to finish it within a given time frame and deliver high quality results. By conducting these supplementary OLS regressions, further significant correlations, that cannot be predicted ex-ante, may emerge and could then be analysed and explained in more detail.

(23)

23

Conclusion

Previous literature has been able to describe separate performance influencing factors and their connection to procrastination and intrinsic motivation. But due to diverging experiment outcomes, with respect to procrastination behaviour specifically, the overall effect of internal compared to external deadlines has not been explained yet. This is why the paper aimed to introduce an experiment design that can demonstrate the impact of self-imposed deadlines compared to external ones on success and accuracy rates of a task. In order to reach this goal, an experiment outline, designed to be conducted at the CREED lab of the University of Amsterdam, has been constructed. The experiment is divided into two different kinds of sessions, where participants have to solve the same task, but are facing either self-set or externally imposed deadlines. In this way the overall effect of the two contrasting deadline types on task success and accuracy can be tested. The goal of the experiment is to evaluate three hypothesized statements. The main hypothesis claims that the introduction of internal deadlines improves task success and accuracy rates. In order to prove its validity, it has been proposed that proportion tests with respect to the two performance indicators could be conducted. A restricted portion of the overall number of participants would be used in the test to control for outside influential factors, including the difference in mathematical ability of students. A way to assess the validity of the two sub-hypotheses, which state that internal deadlines enhance intrinsic motivation and the performance level of a procrastinator, has furthermore been demonstrated. It has been suggested that they can be evaluated by creating OLS regressions, that indicate and explain significant correlations between dependent, independent, and control variables.

One limitation of the designed experiment is the possibility that relevant control variables may be missing, as they have not been accounted for. However, this cannot be proven ex-ante as actual data and results are needed to analyse possible influences further. The same applies for additional aspects of the experiment design that have remained unclear so far, including the required number of participants. Additional relevant factors as well as currently vague details can be determined more precisely after the pilot session is run. As that session will give further indications regarding next steps that are required to realize the designed experiment, conducting it is the closest measure that is needed to continue research on this topic. Moreover, it is expected that further research ideas can be revealed once real data has been gathered regarding the current subject.

(24)

24

Bibliography

Ariely, D., & Wertenbroch, K. (2002). Procrastination, deadlines, and performance: Self-control by precommitment. Psychological science, 13(3), 219-224.

Bisin, A., & Hyndman, K. (2014). Present-bias, procrastination and deadlines in a field

experiment. National Bureau of Economic Research.

Brunnermeier, M. K., Papakonstantinou, F., & Parker, J. A. (2008). An economic

model of the planning fallacy. National Bureau of Economic Research.

Burgess, M., Enzle, M. E., & Schmaltz, R. (2004). Defeating the potentially

deleterious effects of externally imposed deadlines: Practitioners’ rules-of-thumb. Personality

and Social Psychology Bulletin, 30(7), 868-877.

Cadena, X., Schoar, A., Cristea, A., & Delgado-Medrano, H. M. (2011). Fighting

procrastination in the workplace: An experiment. National Bureau of Economic Research.

Drucker, P. F. (1999). Knowledge-worker productivity: The biggest challenge. California management review, 41(2), 79-94.

Ebert, P., & Freibichler, W. (2017). Nudge management: applying behavioural science to increase knowledge worker productivity. Journal of Organization Design, 6(1), 4.

Eerde, W. V. (2003). Procrastination at work and time management training. The

Journal of psychology, 137(5), 421-434.

Fischer, C. (2001). Read this paper later: Procrastination with time-consistent preferences. Journal of Economic Behavior & Organization, 46(3), 249-269.

Gupta, R., Hershey, D. A., & Gaur, J. (2012). Time perspective and procrastination in the workplace: An empirical investigation. Current Psychology, 31(2), 195-211.

Lonergan, J. M., & Maher, K. J. (2000). The relationship between job characteristics and workplace procrastination as moderated by locus of control. Journal of Social Behavior

and Personality, 15(5), 213.

Meng, L., & Ma, Q. (2015). Live as we choose: the role of autonomy support in facilitating intrinsic motivation. International Journal of Psychophysiology, 98(3), 441-447.

Metin, U. B., Taris, T. W., & Peeters, M. C. (2016). Measuring procrastination at work and its associated workplace aspects. Personality and Individual Differences, 101, 254-263.

Miao, J. (2008). Option exercise with temptation. Economic Theory, 34(3), 473-501. Nguyen, B., Steel, P., & Ferrari, J. R. (2013). Procrastination's Impact in the

Workplace and the Workplace's Impact on Procrastination. International Journal of Selection

and Assessment, 21(4), 388-399.

Weber, M., & Schram, A. (2017). The Non‐equivalence of Labour Market Taxes: A Real‐effort Experiment. The Economic Journal, 127(604), 2187-2215.

Yockey, R. D. (2016). Validation of the short form of the Academic Procrastination Scale. Psychological Reports, 118(1), 171-179.