• No results found

The impact of multiple deadlines compared to a single deadline on the performance of both a mathematical and a verbal task

N/A
N/A
Protected

Academic year: 2021

Share "The impact of multiple deadlines compared to a single deadline on the performance of both a mathematical and a verbal task"

Copied!
34
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

The impact of multiple deadlines compared to a single

deadline on the performance of both a mathematical and

a verbal task

Aletta Verberg1 10193162

Supervisor: dhr. dr. A. van der Veen Second Examiner: dhr. dr. J.B. Engelmann December 15, 2017

Abstract

Nowadays, more assignments and exams are implemented in the study program of students. The increase in the number of deadlines is supposed to improve the performance of students. However, time pressure may arise when deadlines are implemented. This can lead to choking under pressure. Therefore, it is not yet clear whether the implementation of the number of deadlines is improving performance. In this paper, the effect of a single deadline is compared to the effect of multiple deadlines on the performance of a task. Past research has shown that men and women do not always perform equally in different types of tasks. Therefore, two types of tasks were used, a mathematical task and a verbal task. The main result is that the performance under multiple deadlines was significantly lower than under one single deadline for the mathematical task. For the verbal task, no differences were found. Since the results for the two types of tasks differ, research should not solely experiment with mathematical tasks, but also with verbal tasks.

1 University of Amsterdam: Faculty of Economics and Business. Master of Economics, track Behavioural Economics and Game Theory (15 ECTS).

(2)

Statement of Originality

This document is written by Student Aletta Verberg, 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)

Table of contents

STATEMENT  OF  ORIGINALITY  ...  2  

TABLE  OF  CONTENTS  ...  3  

1.  INTRODUCTION  ...  4  

2.  LITERATURE  REVIEW  ...  5  

2.1.  EFFECTS  OF  DEADLINES  ...  5  

2.2.  EFFECTS  OF  COMPETITIVE  INCENTIVES  IN  PAYMENT  SCHEMES  ...  7  

3.  EXPERIMENTAL  DESIGN  ...  11   3.1.  THE  TREATMENTS  ...  11   3.2.  REWARDS  ...  13   3.3.  HYPOTHESES  ...  14   4.  RESULTS  ...  17   4.1.  DESCRIPTIVE  STATISTICS  ...  17  

4.2.  THE  EFFECT  OF  DEADLINES  ...  18  

4.3.  GENDER  DIFFERENCES  ...  21  

4.4.  EFFECTS  OF  CONTROL  VARIABLES  ...  22  

5.  CONCLUSION  AND  DISCUSSION  ...  24  

CONCLUSION  ...  24   LIMITATIONS  ...  25   DISCUSSION  ...  25   REFERENCES  ...  27   APPENDICES  ...  29   APPENDIX  1:  INSTRUCTIONS  ...  29  

APPENDIX  2:  CONTROL  QUESTIONS  ...  31  

(4)

1. Introduction

To improve the study results of students, midterms are often used. This way, students will not have to revise all the material at once for the finals, but will study already for the midterm. The midterm is a mean to induce an extra deadline and eventually increase the performance of students. But what is the effect of such an extra deadline?

Ackerman and Gross (2003) found that students who have less free time to study, actually perform better than students who have more free time. As the students with less free time to study face more deadlines, it may be the case that setting multiple external deadlines is improving performance. Ariely and Wertenbroch (2002) already investigated the effect of self-imposed deadlines compared to deadlines set by an external entity. They found that these self-imposed deadlines are not as effective as external deadlines. But when multiple external deadlines are set to improve performance, the time pressure may cause choking under pressure and affect performance negatively.

There can only be an increase in performance when the positive effect of setting deadlines is greater than the negative effect of choking under pressure caused by time pressure. To find out whether this is in fact true, this paper tries to determine whether a situation with multiple external deadlines results in a higher performance compared to a situation where there is only one external deadline.

To answer this question, an online experiment is conducted to collect data about the performance under a single deadline and on the other hand multiple deadlines. Additionally, due to gender differences in performance in mathematical and verbal tasks, especially under competitive incentives, both these types of tasks are included.

The main result is that the negative effect of multiple deadlines in the mathematical task is larger than the negative effect of one deadline in the mathematical task. For the verbal task, there was no significant difference between the performance under multiple deadlines compared to one deadline. Gender differences in performance were negligible.

The content of the paper is organized in the following manner. In the second section, literature about deadlines and time-pressure is reviewed. The experimental design of this paper is described in section 3 and the data of the experiment is analysed in the fourth section. This thesis will end with a discussion and conclusion in section 5.

(5)

2. Literature review

Before the data is analysed, the most important literature on deadlines is reviewed. This involves evaluating the effects of time constraints and time pressure. Finally, literature on gender differences in performance under deadlines, competitive incentives and different type of tasks is discussed.

2.1. Effects of deadlines

When looking at deadlines, there are two important terms to define; time pressure and time constraints. Time pressure involves putting individuals in a situation where stress can occur when there is the possibility that there is not enough time to finish a task. A time-constraint on the other hand is simply a deadline (Ordonez & Benson, 1997). It may be possible that the time-constraint that is given for a task, gives more time than is actually needed to complete a task. Therefore, setting a deadline does not always create time pressure.

When looking at the effect of deadlines, Maule, Hockey and Bdzola (2000) mention that deadlines may increase the involvement and effort of participants, resulting in higher performances. To examine the effect of time pressure on the performance of individuals, De Pauola and Gioia (2016) ran a field experiment with students taking an actual exam. Under time pressure, the final grades of the students were significantly lower than when there was no time pressure2. The exam in the experiment of De Pauola and Gioia (2016) consisted of both a verbal and a numerical part. The negative effect of time pressure was slightly higher3 for the numerical task, but both effects were negative. De Pauola and Gioia (2016) also looked at gender differences. They find that females are stronger negatively affected by time pressure than men. This negative effect is significant for the verbal task, even though females performed significantly better in the verbal task than men.

There are some explanations for the lowered performance under time pressure. One of them is the change in the decision-making process of an individual. When a decision is made while having limited time, the strategies used to make decisions are simplified or decisions are taken faster than when there is no time-constraint (Ordonez & Benson, 1997). The reasoning process is affected, leading the brain to ignore relevant information and to fall back

2 Since the exam already had a time-restriction, De Pauola and Gioia (2016) created a time scheme to create time pressure. This way, controlling for time pressure was possible.

3 De Pauola and Gioia (2016) do not mention statistically significant differences between the effect of the verbal and the numerical task.

(6)

on cognitive heuristics (Verplanken, 1993; Rieskamp & Hoffrage, 2008). When relevant information is missing and time pressure is disrupting the decision-making process, the brain makes decisions with minimal cognitive effort based on heuristics, which Lau and Redlawsk (2001) call ‘cognitive shortcuts’. The heuristics are defined as problem-solving strategies. Making decisions based on heuristics might be faster, but may also result in biased decisions.

Another effect of time pressure, found by Chuderski (2016), is the loss in relational learning. Chuderski (2016) executed an experiment with treatments with high time pressure, relaxed time pressure and no time pressure, in combination with prior experience. When time pressure is high, performance did not increase when prior experience was present, while in treatments without time pressure, or relaxed time pressure, performance did increase when prior experience was present. The overall performance of the treatments without time pressure was higher than the performance of the treatments with time pressure. This is often called ‘choking under pressure’.

Beilock and Carr (2005) find that in cognitive tasks, the working memory is most affected by time pressure. Due to this loss in working memory, the performance decreases. The authors imply that workers that are most qualified will be affected most, since the working memory affected is the asset that makes it possible to perform better than others. This might be an explanation for choking under pressure. The reasoning of Beilock and Carr (2005) can also explain the result found by De Pauola and Gioia (2016), that even though females perform better in verbal tasks, the negative effect of time pressure is larger than for men.

The effects of time pressure and time constraints have been evaluated above, but possible explanations for the differences in the effects of imposing one deadline or multiple deadlines for a task have not yet been addressed. In the case of having one deadline, the concept procrastination may be relevant. Procrastination arises because people have reversed preferences for effort and benefits. Benefits are preferred immediately, while costs are preferred as late as possible (O’Donoghue & Rabin, 1999). The optimal moment to complete a task depends on these preferences. The findings of Ackerman and Gross (2003), where students with more free time (and therefore less deadlines) perform worse than students with less free time, can be explained by the concept of procrastination, where tasks are postponed to a moment closer to the deadline. Students who do not have a lot of free time don’t have the possibility to procrastinate. People may also procrastinate because they get tired after

(7)

performing tasks, increasing the cost of exerting effort to complete the task (Bisin & Hyndman, 2009). This leads to a different optimal moment to complete the task.

With one deadline, it should be possible to allocate time by preference. Because people are aware of their procrastinating behaviour, they try to optimally use their time by setting self-imposed deadlines. As they were not evaluated as effectively as external deadlines (Ariely & Wertenbroch, 2002), self-imposed deadlines may not completely correct for the effect of procrastination.

However, Burger, Charness and Lynham (2011) investigated if the problem of procrastination could be solved by implementing multiple external deadlines and found contradicting results. Burger et al. (2011) find that the implementation of multiple external deadlines was not lowering procrastination. They state that the level of procrastination depends on incentives and task descriptions.

2.2. Effects of competitive incentives in payment schemes

When competitive incentives4 are introduced, a worker can either increase labour effort or cognitive effort, according to Bracha and Fershtman (2013). The labour effort is simply working harder, while cognitive effort is working smarter. The authors examine whether these efforts are increasing under competitive incentives. The labour effect is supposed to increase since the worker wants to work harder under the competitive incentives, but the effect on the cognitive effort is harder to evaluate. One of the problems with the cognitive effort, is the phenomenon ‘choking under pressure’, which arises when there is a lot at stake for an individual. Choking under pressure makes it harder to think and therefore to perform well.

Bracha and Fershtman (2013) executed a lab experiment where the participants had to do a cognitive task and a ‘labour’ task, within ten minutes. The cognitive task was rewarded with more money per correct answer, but was more difficult (filling in number sequences). The ‘labour’ task involved evaluating a number as either ‘odd’ or ‘even’. The participants could divide the time spent on either of the two tasks themselves. The main results were that the level of performance was lower under competitive incentives and that less time was spent on the cognitive task. Additionally, the success rate in the first 7 minutes was significantly higher than in the last 3 minutes of the experiment. This is striking, since there was no

4 Competitive incentives are incentives that motivate people to compete with other people (Niederle & Vesterlund, 2007). Usually, these incentives are monetary rewards based on the performance of not only the individual, but also on the performance of other individuals. The monetary reward is therefore depending on the relevant performance of an individual.

(8)

pressure effect found in the treatment without competitive incentives, while the time for the task was similar. When the participants are divided in high and low performance groups, the lower performance in the competitive incentivized treatment was only significant for low performers. This in contradiction to the time spent on the cognitive task, which was only significantly lower for high performers.

When dividing the participants on the basis of their gender, there is a significant difference in the time spent on the cognitive task. Women spent more time on the labour task than on the cognitive task, while the success rates were similar between the labour task and the cognitive task. The performances between the competitive incentivized and the piece-rate payment weren’t significantly different for both men and women. Bracha and Fershtman (2013) explain the gender difference in time spent on the cognitive task by differences in risk-aversion and confidence.

Niederle and Vesterlund (2007) also investigated the difference in performance between men and women for a mathematical task, under non-competitive and competitive payment schemes. The participants were asked to do some math problems, consisting of adding sums of five numbers for five minutes. There were no significant differences found in performance between men and women. The same results were found by Gneezy, Niederle and Rustichini (2003), who executed an experiment with solving mazes.

After the task that was rewarded with a piece-rate scheme in the research of Niederle and Vesturlund (2007), the participants were asked to participate in a tournament, where the payment scheme was relative to the other participants. In the last task, the participants could choose between a tournament and the piece-rate reward scheme. Men would choose the tournament relatively more than women. When in the tournament, women performed equally compared to the task with the piece-rate payment scheme. However, men performed significantly higher when they were participating in a tournament. Contradicting this result, when men and women were only competing against their own gender, the difference in performance between men and women is not significant anymore (Gneezy et al., 2003). This implies that women also increase their performance under a competitive payment scheme, but when put together with men, women will not perform better.

In another research of Niederle and Vesterlund (2011), the preference of men to compete in tournaments is investigated. One of the traits that may explain this preference is the confidence level of men that is relatively higher than for women (Niederle & Vesterlund, 2011). Another aspect they mention is that women are more risk averse than men, which

(9)

explains why men prefer taking risk more. The risks are higher in a tournament, therefore the lower risk aversion in combination with the increased confidence levels could be an explanation for the increased participation of men in tournaments.

In addition to the level of confidence and risk aversion, Shurchkov (2012) also adds the influence of task stereotypes and time constraints as an explanation for the gender gap in performances. She thereby assumes that men and women have equal skill levels, but react differently to competitive circumstances. Additionally, Shurchkov (2012) assumes that there may be different outcomes for different tasks. This comes from the observation that women choose other types of careers and therefore may be better at other tasks than usually tested in experiments. To test this statement, Shurchkov not only executes a mathematical task, but also a verbal task.

The tasks that were used were conducted in a way that both quantity (number of answered questions) and quality (number of correct answered questions) could be tested. The tasks were executed under a non-competitive payment scheme (piece-rate payment) and a competitive payment scheme (tournament). Lastly, high and low time pressure treatments were added to the various treatments.

Shurchkov (2012) does not find any differences in performance under the non-competitive payment scheme with time pressure for the mathematical task, but in the tournament treatment, men score significantly higher. Yet when looking at the verbal task under high time pressure, women score better than men, even though the difference (14.19 versus 12.91) is not significant under the piece-rate scheme. In the tournament treatment the difference is significant, but only at a ten percent confidence level.

Under low time pressure, the overall performance in the mathematical task is higher compared to the high time pressure treatment, in both the tournament as piece-rate payment. Women perform significantly better than men in the tournament under low time pressure, while there is no difference in the piece-rate treatment. Shurchkov (2012) considers motivation to keep competing the most plausible explanation for the gender difference. This motivation is higher in the verbal task, possibly caused by a higher level of confidence in the verbal task.

Taking all the literature into account, the effect of a time constraint may be positive when it is increasing task involvement. However, when the time constraint causes time pressure, the performance may be lowered. When people experience a high level of time-pressure, the

(10)

cognitive skill level of individuals typically decline due to the ‘choking under pressure’ effect, resulting in a lower level of performance. Setting self-imposed deadlines may help, but literature find that these deadlines are not as effective as external deadlines.

Typically, a tournament structure (a competitive payment scheme) increases the performance of individuals compared to the performance under a non-competitive payment scheme. However, there are some variations between the performances of men and women. Gender differences in performance under competitive incentives, specifically competitive payment schemes, are explained by risk aversion and confidence (Niederle & Vesturlund, 2011; Bracha & Fershtman, 2013). Shurchkov (2012) adds time-constraints and stereotyping in tasks as possible explanations for gender differences and finds somewhat contradicting results for men and women between a mathematical and a verbal task.

Since the effects of competitive incentives on verbal tasks is less tested, a research to this topic may give new insights to the effect of competitive incentives on the performance of individuals. Moreover, current literature is focused on tournament reward schemes and time-pressure, but is not agreeing on the effect of multiple deadlines yet. Therefore, the effect of one deadline compared to multiple deadlines has not fully been determined. This paper tries to contribute to the overview of the effects of deadlines and time pressure in combination with competitive incentives in general, and for task context and gender specifically.

(11)

3. Experimental design

This section is focused on the experimental design. To gather the data for this thesis, an online experiment was conducted with the use of Qualtrics, an online tool to create, execute and analyse surveys. In this section, the tasks and the setup of the experiment will be explained. Additionally, the reward scheme is explained. This section ends with the hypotheses that are tested in the fourth section.

3.1. The treatments

There were 66 students participating in the experiment. All of the participants were students at the University of Amsterdam, enrolled in the first year of the bachelor in either Business Administration or Economics and Business Economics. In the experiment, each participant executed either a mathematical or verbal task. The participants were randomly selected for one of the two types of tasks, so either the mathematical or the verbal task. Both tasks consisted of two parts, part A and part B. The parts consisted of 9 questions each, so the entire task consisted of 18 questions. In part A, the baseline, the participants would answer the 9 questions without any deadlines. After part A, the participant answered 9 similar questions again in part B, but now deadlines were implemented. In part B, the participant was randomly selected to answer the questions in either a situation with one deadline or in a situation with multiple deadlines. An entire overview of the tasks in combination with the separate parts is given in Table 1. After the tasks, a few control questions were asked.

Table 1: Overview of the treatments in the experiment

Part A Part B Mathematical

task (N=28) Baseline: no deadline

Single deadline (N=13) Multiple deadlines (N=15)

Verbal task (N=38) Baseline: no deadline Single deadline (N=18) Multiple deadlines (N=20) Notes: The total number of participants was 66. From the total of 28 students participating in the mathematical task, 13 were in the single deadline treatment and 15 in the multiple deadlines treatment in part B. The 38 participants in the verbal task in part B had a distribution of 18 students in the single deadline treatment and 20 students in the multiple deadlines treatment.

In the single deadline treatment, all the 9 questions were displayed on the screen, together with a clock that was counting down starting from 6 minutes. After these 6 minutes, the participant was automatically forwarded to the control questions, which was the end of the

(12)

experiment. The clock was prominently shown on the screen to induce the feeling of time pressure. On average, the participant had 40 seconds to answer each question. However, because all the questions were shown at the same time, the participant could choose to skip a question, to return to it later. Also, it was possible to allocate the total time spent on a question to the participants’ own preference. This way, if a question took less time than 40 seconds to answer the questions, the participant had time left to spend on questions that were considered more difficult. If the participant did not provide an answer to one of the questions, this question was evaluated as incorrect.

Allocating time was not possible in the multiple deadlines treatment. Now each question was shown on a separate page, with a countdown of 40 seconds per question. After the participant had filled in their answer and clicked to the next page, it was not possible to return to a question5. If the participant was still working on a question after 40 seconds, the page was

automatically forwarded to the next question. If there was no answer provided by the participant, the question was marked as a false answer.

The design of the experiment was constructed with a baseline (the situation where there was no deadline), which allowed controlling for the abilities of the subjects. A between-subjects design was used to compare the treatments with one deadline or multiple deadlines. This was also the case for comparing the mathematical task to the verbal task. The experiment was constructed in such a way to prevent any learning that could occur when a participant would face all the treatments, since the participant would learn that there were deadlines involved in part B of the task. Moreover, conducting both the tasks in one experiment could result in an experiment that would be considered too long. The participants could lose motivation while doing the experiment when the experiment takes too much time to complete.

The mathematical task

The mathematical questions were based on the questions posted by Fibonicci (2017). The participants were asked to give an answer to nine numerical sequences. To make sure that the participants understood the experiment before the tasks started, instructions were given and there was an example question shown. All the instructions of the experiment can be found in

5 To control for effects caused by displaying the single deadline and multiple deadlines treatment differently, the baseline (part A) was also divided in two types of displays. In part A1, all the questions were shown on one page, while the questions were shown on separate pages in part A2. Again, returning to questions was not possible with the separate pages display. The participants were randomly selected for part A1 or part A2.

(13)

Appendix 1. The example question given to the participants was to complete the sequence of five numbers: 2 – 4 – 9 – 11 – 16. There were four possible answers to choose from, 21, 19, 18 and 23. To make sure that the participant understood the task, feedback about the given answer was provided to the participant. The correct answer is 18, since the sequence alternates between +2 and +5. The correct answer and the explanation of how to find the correct answer were provided to the participant for the example question.

The verbal task

The verbal questions were also based on questions by Fibonicci (2017). The verbal questions were word analogies. The sentence that was provided had two missing words. To find the right words to fill in the blank spaces, the participant had to find the pattern between the two words that were given in the sentence. In the example question (in this case: … stands to cow as wool stands to …), the pattern between cow and wool is important. There is a word that stands to cow, as wool stands to another word. In this case, the answer would be that milk is provided by a cow as wool is gained from a sheep. To come up with the answer, the participant had to find the logic behind the sentence and find the pattern between the two words that were shown in the question. There were 5 possible answers for both the blanks. These answers were grouped in columns, the first column included the five choices for the first blank, the second column provided the five possible answers to the second blank. After the example question was answered, feedback about the answer and the explanation of the answer was given.

Control questions

The control questions provided information about the age, gender, home country, mother tongue, track of study and how the participants evaluated some statements about deadlines and time pressure. The participants were asked to scale in what degree they agreed to the statement. The scale consisted of 7 possible answers: strongly disagree, disagree, somewhat disagree, neither agree nor disagree, agree and strongly agree. The exact statements can be found in Appendix 2.

3.2. Rewards

The participants received 1 point for every correct answer in both parts. When the participant did not provide an answer, or answered the question falsely, the participant received 0 points for that question. The monetary reward of the participant depended on the relative ranking of

(14)

the participant. If the participant scored the highest amount of points of all the participants and finished first place, his potential reward was 20 euros. If the participant ended in 25th place, his potential reward was 5 euros. An overview of the reward scheme is given in Table 2. Only four participants were randomly selected for payment after the experiment. This way, the payment scheme was a tournament scheme and therefore a competitive incentive payment scheme. Since the amount of points was the total points obtained in the entire experiment, the payment incentives were equal among all treatments.

Table 2: Reward scheme

Overall ranking participant Reward in euros

1 20 2 18 3 15 4-10 10 11-20 8 >20 5

Notes: The participants received one point for every correct answer and zero points for false answers. When a question was not provided with an answer, zero points were given to the participant. The overall ranking of a participant was depending on the total amount of points. The participant with the highest amount of points was ranked in first place and would be rewarded with 20 euros, if and only if this participant was one of the four participants selected for payment.

3.3. Hypotheses

The main focus of this paper lies on the effect on the performance of a task under two different situations, either with a single deadline or multiple deadlines. When time pressure plays a role in the execution of a task, choking under pressure often occurs. The time constraint of 40 seconds per question is equal in both treatments, but the feeling of time pressure is higher in the multiple deadlines treatment due to the predetermined allocation of time. Due to the higher feeling of time pressure, the effect of choking under pressure is assumed to be higher in the multiple deadlines treatment than in the single deadline treatment. On the other hand, more time pressure may also increase the focus and motivation of the participants, increasing performance.

The one deadline treatment gives the participants the opportunity to choose their own order of completing the questions and allows skipping questions. This way, the participants can decide the allocation of time, in line with their preferences. The participant now has the ability to spend time that was left from an easier question to a harder question, possibly resulting in a higher performance than under the multiple deadlines treatment. However, the

(15)

personal allocation of time might also result in ‘self-imposed’ deadlines to stay within the overall deadline of 6 minutes. Since these deadlines are less effective, participants might lose track of time or fail in self-control and spend too much time on a question, leaving too little time for the remaining questions and finally in less correctly answered questions.

The performance in the multiple deadlines treatment is only larger than the single deadline treatment when the total effect of choking under pressure combined with the motivational factor is less negative (or more positive) than the total effect of ‘self-imposed’ deadlines bundled with the potentially more optimal allocation of time.

Since choking under pressure is assumed to play a bigger role in the execution of the task in the multiple deadlines treatment than the motivational factor, thus the total effect is expected to be negative. In the single deadline treatment, the overall effect is presumably positive, since the personal allocation of time is assumed to have a bigger positive effect than the effect of failing in self-control in the self-imposed deadlines. Taken together, it is more likely that the performance is higher in the single deadline treatment compared to the performance in the multiple deadlines treatment. Together, the following hypothesis is conducted:

Hypothesis 1: Multiple deadlines within a task decrease the performance compared to a single deadline for the same task.

When looking at the overall effect of deadlines compared to a situation without deadlines, the effect of the implementation of deadlines depends mainly on the time restriction. If the time restriction is set in a way that time pressure arises, choking under pressure may affect the performance negatively, for both the mathematical and the verbal task. Even though procrastination can play a part in a situation without deadlines, it is assumed that this effect is lower than the effect of choking under pressure. Therefore, the second hypothesis is as follows:

Hypothesis 2: Implementing deadlines does not increase the performance on a task.

Most literature agrees on the fact that under piece-rate payment schemes, men and women have similar performances (for both mathematical and verbal tasks). But when the payment is under a tournament payment scheme, literature suggests that men are most likely to perform better than women in mathematical tasks. Since the payment scheme in this experiment is a

(16)

tournament scheme, it can be assumed that men will perform better than women. Moreover, the negative effect of time pressure on women is greater than for men. Therefore, the following hypothesis about the performance in the mathematical task is set:

Hypothesis 3: Men outperform women in the mathematical task under no deadline, a single deadline and multiple deadlines.

For the verbal task, there are results to assume that women perform better than men in the verbal task in tournament form, under low time pressure. For high time pressure this is not found. This results in the following hypothesis on the performance of women in the verbal task:

(17)

4. Results

In this section, the data of the experiment is analysed. Firstly, the effect of one deadline is compared to the effect of multiple deadlines in both the verbal and the mathematical task. Additionally, the effect of deadlines is determined. Lastly, gender differences are evaluated. 4.1. Descriptive statistics

What immediately stands out in Table 3 is the difference between the means6 of the overall scores between the mathematical task (score of 12.82) and the verbal task (score of 7.13). The scores for the baseline also differ, 7.68 and 3.89 respectively. Possible explanations for the difference in the mean is the level of difficulty of the task. The verbal task may have been more difficult than the mathematical task. The scores in the treatments are lower than the score in the baseline for the mathematical task. However, this is not the case for the verbal task. An overview of the remaining descriptive statistics is presented in Table 8 in Appendix 3.

Table 3: Overview scores in treatments

Task Observations Mean of scores (standard deviation)

Mathematical task 28 12.8214 (3.1157) Baseline: no deadline 28 7.6786 (1.2781) Single deadline 13 6.4615 (1.7134) Multiple deadlines 15 4.0000 (1.9272) Verbal task 38 7.1316 (3.2314) Baseline: no deadline 38 3.8947 (1.9561) Single deadline 18 3.1111 (1.4507) Multiple deadline 20 3.3500 (2.0072)

Notes: For each task and treatment, the number of observations and the mean of the scores are presented. Since all the participants executed the task without a deadline, the observations were 28 for the mathematical task and 38 for the verbal task. The mean of the scores is the mean of the number of points achieved in the experiment by the participants in the specific task and treatment. For the mathematical task and the verbal task, the maximum of the score was 18 points in total. For the baseline, this was 9 points and for both the single deadline and the multiple deadlines treatment, the maximum was also 9 points. The standard errors of the mean of the scores are presented in parentheses.

6 Testing for differences in medians between the mathematical and the verbal task using the Wilcoxon rank-sum test gives a p-value of 0.000. Therefore, the scores in the mathematical task were significantly higher than in the verbal task. The output can be found in Table 8 in Appendix 3.

(18)

4.2. The effect of deadlines

Hypothesis 1: Multiple deadlines within a task decrease the performance compared to a single deadline for the same task.

To find out potential differences between the effect of multiple deadlines compared to a single deadline, the treatment with one deadline is compared to the treatment with multiple deadlines. This comparison is done by using a Wilcoxon rank-sum test. The rank-sum test is used to determine if two separate sets of observations (in this case, the observations in the single deadline and the multiple deadlines treatment) come from the same population (Stata, 2017). If this is true, the two sets will have the same distribution. Since the number of observations is not large enough and there are no other leads to assume that the data is normally distributed, the non-parametric rank-sum test suits the data.

The means of the scores of part A and B of the two treatments for both the mathematical and the verbal tasks are shown in Table 4. Looking at the outcomes of the rank-sum test in Table 4, the tests differ in outcome between the mathematical and the verbal task. The rank-sum test on the mathematical treatments shows a significant result at a 1 percent level7. The result indicates that the performance under a single deadline is higher than in the multiple deadlines treatment8. Additionally, a rank-sum test has been performed on the difference between the scores in the baseline and in the treatment, to control for the ability of the participants. This test also results in a significant difference between the single deadline treatment and the multiple deadline treatment. For the verbal task, no significant test results are found. For the mathematical task, there is some evidence that supports the hypothesis that multiple deadlines decrease the performance compared to a single deadline for the same task9.

For the verbal task, there are no results to form the same conclusion.

7 Computing an Ordinary Least Squares Regression, similar results were found.

8 The two display types used in the two treatments show no significant effect on the performance and cannot explain the difference between the two treatments in the mathematical task. The output of the test on the display types can be found in Table 10 in Appendix 3.

9 To control for possible differences between the ability of the participants in the single deadline treatment and the multiple deadline treatment, the scores in the baseline are compared. No significant result has been found. Therefore, there is no reason to assume that there were ability differences between the participants in the treatments. The output can be found in Table 11 in Appendix 3.

(19)

Table 4: Comparing the performance in the single deadline and multiple deadlines treatment (1) Treatment (2) Observations (3) Mean of scores (4)

Single deadline versus multiple deadlines

z-value (p-value)

(5)

Single deadline versus multiple deadlines Controlled for scores

in baseline z-value (p-value) Part A Part B Mathematical task 28 Single deadline 13 7.6923 6.4615 -2.862 (0.0042**) -3.655 (0.0003**) Multiple deadlines 15 7.6667 4.0000 Verbal task 38 Single deadline 18 3.3889 3.1111 0.299 (0.765) -1.172 (0.2410) Multiple deadlines 20 4.3500 3.3500

Notes: In the first column, the treatments are described. In the second column, the observations per treatment are shown. In the third column, the mean scores) of the participants in the experiment (split into parts A and B) are provided. In the fourth column, a rank-sum test is used to determine if there is a significant difference between the scores of the participants in the two treatments. In the fifth column, the scores in the treatments are corrected by deducting the score in the baseline (part A). For the mathematical tasks, significant differences are found. Significance is indicated as follows: +if p < 0.1, * if p < 0.05, ** if p < 0.01.

When looking more closely to the performance in the mathematical and verbal task, it is straightforward to assume that the results differ between the mathematical and verbal task. Not only are different skills asked to perform well in the tasks, but one of the tasks may also be more challenging to complete. Also, the mathematical task had four possible answers per question, while there were 25 possible combinations in the verbal task. This made guessing less favourable in the verbal task compared to the mathematical task. The difference of the scores in the different types of tasks was significant, implicating that the distribution of the overall performance (the total scores of the participants in the baseline and the treatment) in the verbal task is lower than the distribution of the mathematical task.

Combining the difference in difficulty with the fact that participants reacted differently to the implementation of deadlines (both single and multiple) based on the nature of the task, there may be an explanation for the lowered performance in the mathematical task. Beilock and Carr (2005) implied that people who have the best abilities, are mostly affected by time-pressure. This might be an explanation for the difference between the effect on the mathematical task and the verbal task. If the mathematical task was easier than the verbal task, people were performing better and feeling more confident while doing the task. After the introduction of the deadlines, there was more to affect than in the verbal task. The participants

(20)

in the verbal task could not improve themselves due to the difficulty of the task, there was little ability to affect.

Hypothesis 2: Implementing deadlines does not increase the performance on a task.

To determine what the effect of deadlines is, the performance in the treatment with deadlines (part B) are compared to the performance in the baseline (part A). The Wilcoxon sign-rank test is used. This test is different from the Wilcoxon rank-sum test, since the sign-rank test is used on matched data (Stata, 2017). Since the experiment has a within-subjects design for task A and task B, the data of task A and task B are matched. Therefore, the sign-rank test is appropriate. The sign-rank test assumes in its null hypothesis that the difference between the median of the two sets of observations is equal to zero. The test compares the number of correct answers in the baseline to the number of correct answers in part B of the experiment, the part where the deadlines were introduced.

Table 5: Comparing the baseline to the treatment

Treatment Signed-rank test

z-value p-value

Mathematical task

Baseline versus single deadline 2.757 0.0058**

Baseline versus multiple deadlines 3.432 0.0006**

Verbal task

Baseline versus single deadline 0.736 0.4619

Baseline versus multiple deadlines 1.959 0.0501+

Notes: The Wilcoxon Signed-Rank Test produces z-values for the comparison of the median of the points scored in the baseline and part B for the mathematical and the verbal task. The matching p-values are notated in the third column. The notation of significance is as follows: + if p < 0.1, * if p < 0.05, ** if p < 0.01. For the

mathematical tasks, both treatments show significant differences between the scores in the baseline and the treatment task. For the verbal task, the baseline score and the multiple deadlines treatment show a difference at a 10% level.

In Table 5, the outcomes of the Wilcoxon signed-rank test on the mathematical and verbal treatments can be found. For the mathematical treatments, the null hypothesis that the treatment with deadlines (part B) has no effect on the performance (the number of correct answers of the participants) can be rejected at a level of 0.05. This means that there is some evidence that supports the statement that the median of part B differs from the median in the

(21)

baseline. There is an indication that the median in performance in the baseline is higher than in part B. This is not line with the hypothesis stated that deadlines do not affect performance. For the verbal multiple deadlines treatment, the difference in the median is significant at a 10% level (p-value of 0.0501). The null hypothesis of the Wilcoxon signed-rank test of no differences between the median of the treatment group (part B) and the baseline group (part A) cannot completely be rejected, since only one treatment was significant.

For the verbal and the mathematical task, different results are found. For the mathematical task, the hypothesis is rejected and there is empirical evidence to assume that deadlines negatively affect the performance. This can be explained by choking under pressure. The differences in the median can also be caused by other influences. For instance, there may be a significant difference in the level of the questions in the baseline compared to the treatment questions. The difference in level is not tested before executing the experiment, which could explain the difference between the two parts, maybe even better than the influence of the deadlines introduced in part B. However, since the results in the mathematical task were significant at a 1% level and for the multiple deadlines treatment in the verbal task at 10%, there is empirical evidence to assume that implementing deadlines decreases the performance of a task.

4.3. Gender differences

In total 28 females and 38 men participated in the experiment. The total score of men is on average 9.74, while the mean of the total score of women was 9.29. This difference is not significant. Therefore, we can assume that men and women have a similar distribution and a similar median (see Table 12 in Appendix 3). More interesting is the performance of both genders in the different tasks and treatments.

Hypothesis 3: Men outperform women in the mathematical task under no deadline, a single deadline and multiple deadlines.

For all the treatments, there are no significant differences found between the performance of men and women (Table 6). Therefore, equal distributions and medians are likely. Based on these results, there is not enough empirical evidence to support the hypothesis that men outperform women in the mathematical task when there are no deadlines, a single deadline or multiple deadlines.

(22)

Table 6: Gender differences among treatments

Treatment Rank-sum test

z-value p-value Mathematical task Baseline -0.739 0.4597 Single deadline -1.579 0.1144 Multiple deadlines 0.967 0.3334 Verbal task Baseline -1.229 0.2192 Single deadline 0.372 0.7098 Multiple deadlines -0.444 0.6572

Notes: The Wilcoxon rank-sum test compares the medians of men and women in the treatments. The notation of significance is as follows: +if p < 0.1, * if p < 0.05, ** if p < 0.01. No significant differences are found.

Hypothesis 4: Women outperform men in the verbal task when there are no deadlines.

Women scored on average 3.43, while men had an average score of 4.17 in the verbal task without any deadlines. This difference in medians was not significant (Table 6). Therefore, no results are found to support the fourth hypothesis that women outperform men in the verbal task when there are no deadlines. Men even performed better in the experiment on the verbal task, even though this difference is not significant.

4.4. Effects of control variables

Even though there is no basis to assume that the data is normally distributed, some regressions are done on the data. The control variables are now also taken into account. An overview of the control variables is given in Table 9 in Appendix 3. From the regression on the total score in Table 7 can be stated that being assigned to the mathematical task is significantly (p-value is 0.000) improving the total score of the participant, by 5.3 points, which was also shown by the non-parametric test. Liking deadlines was also significantly (at a level of 5%) improving the total score of the participant. At a ten percent significance level, the variables age and studying Economics and Business Economics (compared to Business Administration) also affected the total score. These variables are easy to justify. When someone is older, they may have already gone to University and may even have a little more experience with experiments like these. For the study track it is important to note that the Bachelor in Economics is generally more mathematical, so students may self-select their study track on their performance in math, which is a skill that was needed for this experiment. Finally, there are also no gender differences found in the regression.

(23)

Table 7: Regression on total scores

Variables Total score

Coefficient P-value

Single deadline treatment 0.0477 0.954

Mathematical task 5.4070 0.000**

Age 0.4290 0.094+

Male 0.5414 0.542

Native English -1.6453 0.197

Study EBE 1.7585 0.089+

I like time pressure -0.4210 0.268

I perform better with time

pressure 0.0144 0.970

I like deadlines 0.7620 0.044*

I perform better with deadlines 0.0144 0.970

I perform better with multiple deadlines than with one deadline

-0.1358 0.611

Constant -4.3960 0.390

R-Squared 0.5637

Observations 66

Notes: The variable ‘Single deadline treatment’ is a dummy variable with a value of 1 if the participant was selected for the treatment with a single deadline and a value of 0 if the participant was selected for the multiple deadlines treatment. The variable ‘Mathematical task’ is a dummy variable with a value of 1 if the task executed was mathematical and a value of 0 if the task was verbal. The variable ‘Male’ is a dummy variable with a value of 1 if the participant was male and a value of 0 if the participant was female. The variable ‘Native English’ is a dummy variable with a value of 1 if the mother tongue of the participant is English, otherwise the value is 0. The variable ‘Study EBE’ is a dummy variable with a value of 1 if the participants studies Economics and Business Economics and has a value of 0 if the participant studies Business Administration. For the statements, a scale of 1 to 7 is used, with 1 representing strongly disagree and 7 representing strongly agree. The regressor is the total score of the participant in the entire experiment. Significance is indicated as follows: +if p < 0.1, * if p < 0.05,

(24)

5. Conclusion and discussion Conclusion

To give an answer to the question whether multiple deadlines increase the performance compared to a single deadline, an experiment is conducted to compare the performance of students in a task where there is a single deadline for a task to a situation with multiple deadlines for a task. Additionally, the effect of setting deadlines is examined by conducting a baseline task without any deadlines. Since there is empirical research that finds that men and women respond differently to time pressure and perform differently in mathematical and verbal tasks, the experiment used both a mathematical and a verbal task to check for gender differences.

It is expected that participants in the multiple deadlines treatment will experience more time pressure than in the single deadline treatment, which will lead to choking under pressure. Choking under pressure will lower the performance and the effect will be larger in the multiple deadline treatment. Moreover, with a single deadline, time can optimally be allocated, which is improving performance. Due to the assumption that choking under pressure will negatively affect the performance when deadlines are set, it is expected that implementing deadlines does not increase the performance on a task compared to a situation without deadlines. Furthermore, gender differences in the experiment are expected. More specifically, it is expected that men perform better in the mathematical task, while women will outperform men in the verbal task, solely when there are no deadlines.

The results of the experiment show that the performance in the multiple deadlines treatment was lower than in the single deadline treatment. For the mathematical treatment, this difference was significant. For the verbal treatment, no significant difference was found. When looking at the overall effect of deadlines compared to a situation without deadlines, deadlines negatively affect the performance. There was a significant lower performance for the mathematical task. Comparing the baseline to the task under multiple deadlines for the verbal task, the difference in performance is also significant. However, no significant difference was found for the single deadline treatment compared to the baseline. The effects of the deadlines were also checked for gender differences; however, these effects were barely existing and definitely not significant.

When looking at the results, at first sight it looks that it is not possible to assume that there is a difference in behaviour of participants when there are multiple deadlines for a task, instead of just one. However, looking more closely to the results, it becomes clear that the

(25)

effect depends partly on the type of task. There are multiple indicators that the results between the different type of tasks differ. For the mathematical task, the performance is influenced by setting deadlines. Moreover, the effect is larger when multiple deadlines are set. For the verbal task, the effect is somewhat present, but not significant. This is a clear indication that research in the field of deadlines and time pressure should take into account that the type of task may affect the result.

Limitations

The findings that the performance in the verbal task was not significantly affected by deadlines could possibly be explained by the construction of the experiment. The questions in the verbal task had more possible answers to choose from, which made guessing less profitable than in the mathematical task. This guessing error should be equal for both the tasks. Also, the questions in the baseline and the treatments were not tested extensively enough for potential differences in difficulty. Now the effect of deadlines is harder to measure due to this bias. Additionally, the verbal task is conducted in English, while most of the subjects weren’t native English speakers. Therefore, there may be some influence on the performance caused by language barriers. Other factors that may have played some role in the results of the experiment is the loss of control caused by conducting an online experiment instead of a lab experiment. There is some loss of control in this setup. This mainly influenced the liability of the time spent on questions, which was now unusable. Also, it was not possible to pay all subjects, which lowered the monetary incentives for the participants. Improving the setup of the experiment may give more insight into the behaviour of students and employees to the implementation of deadlines in the study and work place.

Discussion

Besides the setup of the experiment, there are some explanations for the lowered performance when multiple deadlines are set. Choking under pressure becomes more of an issue when multiple deadlines are set. Combining this with the loss of optimal time-allocation for executing the task when multiple deadlines are established, it is no surprise that multiple deadlines lower the performance, compared to a situation with no deadlines or one deadline. When comparing this result to the findings of Ackerman and Gross (2003), where students with less free time outperformed students who had more free time, it may be the case that in this experiment, the time pressure was too high to improve performances. Therefore, it is important to realize that there is a difference between time-constraints and time pressure and

(26)

that time-constraints might be performance improving, but time pressure is not. Implementing deadlines should thoroughly be considered to avoid choking under pressure. Future research could find out more about the trade-off between performance improving time-constraints (those which solve the problem of self-control when it comes to self-imposed deadlines) and time-constraints that result in time pressure, resulting in choking under pressure.

This paper did not find any gender differences, while literature has found different responses of men and women to the type of task used in experiments. Moreover, the competitive incentives used in the experiment were also not causing any gender differences. Therefore, the results found in this paper are not in line with the results that were found in current literature (Bracha & Fershtman, 2013; Niederle & Versterlund, 2007; Niederle & Vesterlund, 2011; Shurchkov, 2012), while there is no clear explanation for this observation. However, finding different results in behaviour between the mathematical and the verbal task is in line with Shurchkov (2012). This difference between the mathematical task and verbal task may be explained by the theory of Beilock and Carr (2005), who state that people who have the best abilities, are mostly affected by time-pressure. Since the scores in the mathematical task were significantly higher, this explanation is possible. However, this paper finds results that lead to the conclusion that verbal tasks should not be forgotten when doing analyses on the behaviour of students and employees when time constraints, time pressure and competitive incentives are present.

(27)

References

Ackerman, D. S., & Gross, B. L. (2003). Is time pressure all bad? Measuring the relationship between free time availability and student performance and perceptions. Marketing Education Review, 13(2), 21-32.

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

Beilock, S. L., & Carr, T. H. (2005). When high-powered people fail: Working memory and “choking under pressure” in math. Psychological Science, 16(2), 101-105.

Bisin, A., & Hyndman, K. (2009). Procrastination, self-imposed deadlines and other commitment devices.

Bracha, A., & Fershtman, C. (2013). Competitive incentives: working harder or working smarter? Management Science, 59(4), 771-781.

Burger, N., Charness, G., & Lynham, J. (2011). Field and online experiments on self-control. Journal of Economic Behavior & Organization, 77(3), 393-404.

Chuderski, A. (2016). Time pressure prevents relational learning. Learning and Individual Differences, 49, 361-365

O'Donoghue, T., & Rabin, M. (1999). Doing it now or later. American Economic Review, 103-124.

Fibonicci. (2017). Verbal Reasoning - Word Analogies Test. Retrieved from https://www.fibonicci.com/verbal-reasoning/analogies-test/

Fibonicci. (2017). Numerical Reasoning – Number Sequences Test. Retrieved from https://www.fibonicci.com/numerical-reasoning/number- sequences-test/

Gneezy, U., Niederle, M., & Rustichini, A. (2003). Performance in competitive environments: Gender differences. The Quarterly Journal of Economics, 118(3), 1049-1074.

Lau, R. R., & Redlawsk, D. P. (2001). Advantages and disadvantages of cognitive heuristics in political decision making. American Journal of Political Science, 951-971.

Maule, A. J., Hockey, G. R. J., & Bdzola, L. (2000). Effects of time-pressure on decision-making under uncertainty: changes in affective state and information processing strategy. Acta psychologica, 104(3), 283-301.

Niederle, M., & Vesterlund, L. (2007). Do women shy away from competition? Do men compete too much?. The Quarterly Journal of Economics, 122(3), 1067-1101.

Niederle, M., & Vesterlund, L. (2011). Gender and competition. Annu. Rev. Econ., 3(1), 601-630.

(28)

Ordonez, L., & Benson, L. (1997). Decisions under time pressure: How time constraint affects risky decision making. Organizational Behavior and Human Decision Processes, 71(2), 121-140.

De Paola, M., & Gioia, F. (2016). Who performs better under time pressure? Results from a field experiment. Journal of Economic Psychology, 53, 37-53.

Rieskamp, J., & Hoffrage, U. (2008). Inferences under time pressure: How opportunity costs affect strategy selection. Acta psychologica, 127(2), 258-276.

Shurchkov, O. (2012). Under pressure: gender differences in output quality and quantity under competition and time constraints. Journal of the European Economic Association, 10(5), 1189-1213.

Stata. (2017). Nonparametric methods: Wilcoxon rank-sum test. Retrieved from https://www.stata.com/manuals/rranksum.pdf

Stata. (2017). Nonparametric methods: Wilcoxon signed-rank test. Retrieved from https://www.stata.com/manuals/rsignrank.pdf

Verplanken, B. (1993). Need for cognition and external information search: Responses to time pressure during decision-making. Journal of Research in Personality, 27(3), 238-252.

(29)

Appendices

Appendix 1: Instructions

Introduction

Welcome  to  this  experiment.  Please  do  not  communicate  while  executing  the  experiment.        

The  experiment  is  divided  into  two  parts.  Each  part  consists  of  9  questions.    

Each  answer  is  worth  1  point  when  answered  correctly.  Every  wrong  answer  or  not  filled  in  answer  gives  you  zero  points.   After  the  two  parts,  you  are  asked  to  fill  in  a  small  survey.  

   

After  the  experiment,  four  participants  are  randomly  selected  for  payment.  The  amount  of  payment  is  depending  on  your   relative  performance,  based  on  the  total  number  of  points.  For  instance,  if  you  end  up  at  first  place,  you  will  receive  20  euros.   If  you've  earned  an  amount  of  points  that  is  ranking  you  at  24th  place,  you  will  be  rewarded  5  euros.    

 

 Rank    Reward  in  euros  

1   20   2   18   3   15   4  -­  10   10   11  -­  20   8   >  20   5    

If  you  want  to  be  able  to  be  selected  for  payment,  your  e-­mail  address  is  needed  to  inform  you  about  your  prize.  These   persons  will  be  selected  and  informed  about  their  reward  before  the  first  of  November.    

 

Please  fill  in  your  e-­mail  address  if  you  want  to  be  eligible  for  payment.        

 

If  you  have  any  questions  regarding  this  experiment,  you  can  contact  me  at  aletta.verberg@student.uva.nl.         Good  luck!       Kind  regards,   Aletta  Verberg    

Instructions mathematical task, part A

This  is  the  start  of  the  first  part.  

You  are  asked  to  answer  9  number  sequence  questions.  You  don't  have  a  time  limit.     You  can  use  scrap  paper.    

 

Below  is  an  example  question  shown.     Example  question:   2      4      9      11      16      ...   o   21   o   19   o   18   o   23  

If the option 18 was selected:

This  is  the  correct  answer.    

If the option 18 was not selected:

This  is  not  the  correct  answer.    

Feedback:

The  sequence  is  alternating  between  +2  and  +5.     2  +  2  =  4   4  +  5  =  9   9  +  2  =  11   11  +  5  =  16   16  +  2  =  18    

(30)

Instructions mathematical task, part B1

You  will  now  start  the  second  part  of  the  experiment.      

You  are  again  asked  to  fill  in  9  questions,  but  now  you  have  a  time  limit  of  6  minutes.  After  these  6  minutes,  you  are   automatically  forwarded  to  the  end  survey.  Questions  without  an  answer  or  incorrect  questions  are  not  rewarded  with  any   points.    

 

You  are  allowed  to  use  scrap  paper.      

Good  luck!    

Instructions mathematical task, part B2

You  will  now  start  the  second  part  of  the  experiment.      

You  are  again  asked  to  fill  in  9  questions,  but  now  you  have  a  time  limit  of  40  seconds  per  question.    

After  these  40  seconds,  you  are  automatically  forwarded  to  the  next  question.  If  you  do  not  provide  an  answer  within  these   40  seconds,  0  points  are  rewarded.  Incorrect  questions  are  also  not  rewarded  with  any  points.    

 

You  are  allowed  to  use  scrap  paper.      

Good  luck!    

Instructions verbal task, part A

This  is  the  start  of  the  first  part.  

You  are  asked  to  answer  9  analogy  questions.  You  don't  have  a  time  limit.      

Below  is  an  example  question  shown.    

You  have  to  select  two  answers,  one  from  the  first  column  and  one  from  the  second  column.    

The  first  column  of  answers  consists  of  possible  words  that  have  to  be  filled  in  at  the  first  dots.  The  second  column  displays   the  possible  words  that  fit  the  second  dots.    

Example  question:    

...  stands  to  cow  as  wool  stands  to  ...     o   Tame   o   farm   o   milk   o   horns   o   bull   o   expensive   o   soft   o   warm   o   sweater   o   sheep  

If the options [milk, sheep] were selected:

This  is  the  correct  answer.  

If the options [milk, sheep] were not selected:

This  is  not  the  correct  answer.    

Feedback

Milk  is  something  that  can  be  gained  from  a  cow,  while  wool  is  gained  from  a  sheep.     Therefore,  milk  stands  to  cow  as  wool  stands  to  sheep.  

Instructions verbal task, part B1

You  will  now  start  the  second  part  of  the  experiment.      

You  are  again  asked  to  fill  in  9  questions,  but  now  you  have  a  time  limit  of  6  minutes.  After  these  6  minutes,  you  are   automatically  forwarded  to  the  end  survey.  Questions  without  an  answer  are  not  rewarded  with  any  points.      

Good  luck!  

Instructions verbal task, part B2

You  will  now  start  the  second  part  of  the  experiment.      

You  are  again  asked  to  fill  in  9  questions,  but  now  you  have  a  time  limit  of  40  seconds  per  question.    

After  these  40  seconds,  you  are  automatically  forwarded  to  the  next  question.  If  you  do  not  provide  an  answer  within  these   40  seconds,  0  points  are  rewarded.      

 

(31)

Appendix 2: Control Questions

What  is  your  age?    

   

What  is  your  gender?   o   Male   o   Female    

What  is  your  home  country?   o   The  Netherlands   o   Other  …    

What  is  your  native  language?   o   Dutch  

o   English   o   Other  …    

What  do  you  study?  

o   Economics  and  Business  Economics   o   Business  Administration  

o   Other  …    

I  like  time  pressure   o   Strongly  disagree   o   Disagree  

o   Somewhat  disagree   o   Neither  agree  nor  disagree   o   Somewhat  agree  

o   Agree   o   Strongly  agree    

I  perform  better  with  time  pressure   o   Strongly  disagree   o   Disagree  

o   Somewhat  disagree   o   Neither  agree  nor  disagree   o   Somewhat  agree   o   Agree   o   Strongly  agree     I  like  deadlines   o   Strongly  disagree   o   Disagree   o   Somewhat  disagree   o   Neither  agree  nor  disagree   o   Somewhat  agree  

o   Agree   o   Strongly  agree    

I  perform  better  with  deadlines   o   Strongly  disagree   o   Disagree  

o   Somewhat  disagree   o   Neither  agree  nor  disagree   o   Somewhat  agree  

o   Agree   o   Strongly  agree    

I  perform  better  with  multiple  deadlines  than  with  one  deadline   o   Strongly  disagree  

o   Disagree  

o   Somewhat  disagree   o   Neither  agree  nor  disagree   o   Somewhat  agree  

o   Agree   o   Strongly  agree    

Referenties

GERELATEERDE DOCUMENTEN

1-let lijkt me niet verstandig de te bereiken vaardigheden in de brugklas zonder meer af te leiden uit de opgaven van een of ander brugklasboek: soms wordt er meer gevraagd van

I1 s'agit de fragments de meules, de molette et polissoir en grès, un éclat de taille en grès bruxellien, un couteau à dos naturel, des déchets de taille,

'N BED RAG VAN NAGENOEG R30 MILJOEN VERSKYN TANS OP DIEKAPITAAISKEDULE VAN DIE P.U. TEN OPSIGTE VAN PROJEKTE WAT PAS VOLTOOIIS, IN DIE PROSES VAN UITVOERING IS OF NOG

Second, fear responses towards the con- ditioned stimuli did not differ for the instructed acquisition group compared to the combined acquisition group, indicating that there are

Developmental letter position dyslexia in Hebrew: Reading words, numbers and diacritics.. Language policy and localization in Pakistan: Proposal for

- Grond- en slootwater in de veenregio van oktober tot april 2010 door Grontmij; - Grondwater in de kleiregio door IDDS. Eerste ronde in november/december, de tweede en

explain that although migration caused by this environmental process will initially be internal and voluntary, with inundations progressing and the remainder of