Contents Executive Summary 3 Aim 3 Relevance in Today’s Environment 3 Survey Analysis 4 Assumptions 6 Hypothesis 6 Multitasking: 6 Methodology 6 Task 1 (Sudoku Puzzle): 6 Task 2 (Word Puzzle): 7 Experiment 1: 7 Experiment 2: 7 Experiment 3: 7 Results and Findings 8 Interpretation 9 Are women better at multitasking? 10 Conclusion 10 APPENDIX 12 References 16

Executive Summary
The question that is under consideration is whether multitasking is good and how it affects performance. Initially a survey was floated for the students of Indian institute of Management Bangalore which was done to get a general perception about multitasking among the MBA graduates. We found that a majority think that multitasking improves productivity. To validate this we conducted experiments in which subjects were made to do two tasks under three different conditions. In the first experiment, subjects were made to perform the two tasks sequentially, in the second they were forced to switch between the tasks and in the third they were free to do in their own way. Results of the experiments show that the performance was better when the tasks were scheduled (Experiment -1) than during multitasking (Experiment- 2). However, the performance was no better when the subjects were free to do the tasks as per their will (Experiment-3). Also, women are perceived to be better multi-taskers than men but the results did not show any significant difference for the same.
Aim
Multitasking is at the heart of today’s workplace. To understand COLLEGE’s perception towards multitasking we conducted a survey regarding multitasking and how people perceive it. We found that most people at COLLEGE believe that multitasking improves productivity and consider they can easily juggle between tasks.
The aim of our project is to understand how seriously multitasking affects individual performance. We also aim to find if women fare better than men when it comes to multitasking1,2. The project is important considering the fact that a lot of people believe contrary to the popular understanding in research.
Relevance in Today’s Environment
As organizations are expecting more out of their employees making the work place highly competitive, it has become very important to study the impact of multitasking on the performance of the employees. Employees today are burdened with multiple tasks that they attempt to perform together thinking that it will save time and increase their productivity. This was further supported from our survey in which more than 50% participants believe that multitasking improves productivity. This study would gauge the relevance of multitasking in today’s workplace.
Survey Analysis
A survey with seven questions was floated among the students of Indian Institute of Management Bangalore to which 88 responses were obtained. Statistics from the survey are the following:
Table1 : Gender Gender | | | | Response | % | Male | | | | 62 | 70% | Female | | | | 26 | 30% | Total | | 88 | 100% |
The participants in the survey were 30% females and 70% males that closely represent the gender distribution in the campus.
Table 2: Age Age in years | | | | Response | % | < 20 | | | | 3 | 3% | 20-22 | | | | 25 | 28% | 22-25 | | | | 35 | 40% | 25-28 | | | | 23 | 26% | >28 | | | | 2 | 2% | Total | | 88 | 100% |
Around 95% of the participants are between 20 and 28 years of age.
Table 3: Number of years of Work Experience Number of years of work experience in years | | | | Response | % | <1 | | | | 36 | 41% | 1-3 | | | | 28 | 32% | >3 | | | | 23 | 26% | Total | | 87 | 100% |
Around 58% of the participants have more than 1 year of work experience. Table 4: Average time spent on a work task uninterrupted Time | | | | Response | % | <10 minutes | | | | 6 | 7% | 10-20 minutes | | | | 25 | 28% | 20-30 minutes | | | | 32 | 36% | >30 minutes | | | | 25 | 28% | Total | | 88 | 100% |

Table 5: After interruption, how long is taken to refocus Time | | | | Response | % | < 2 minutes | | | | 11 | 13% | 2 - 5 minutes | | | | 28 | 32% | 5 - 10 minutes | | | | 29 | 33% | > 10 minutes | | | | 20 | 23% | Total | | 88 | 100% |

Table 6: Number on tasks done effectively at the same time No. of tasks | | | | Response | % | 1 | | | | 16 | 18% | 2 | | | | 52 | 59% | 3 | | | | 15 | 17% | > 3 | | | | 5 | 6% | Total | | 88 | 100% |

Table 7: Do you multi tasking improves productivity? Answer | | | | Response | % | Yes | | | | 52 | 59% | No | | | | 36 | 41% | Total | | 88 | 100% |
From the survey, it can be deduced the following: 1. Majority of the participants (59%) believe that productivity improves when one multitasks. 2. 52 of the 88 participants believe that they can do 2 tasks effectively at the same time and only 16 participants agree that they can do only one task effectively at one time. Thus, 88% of the participants are of the opinion that they are good at multi tasking. 3. Participants were asked the average time they spent on a task uninterrupted. The weighted average of the time is 26.15 minutes (Exhibit 1), which means that there are two interruptions every hour. Similarly, the average time taken to refocus after interruption was taken. The weighted average for the refocus time is 7.175 minutes (Exhibit 2). Considering eight hours of work day and two interruptions every hour, there would be a total 16 interruptions. This could be interpreted as 114.8 minutes of productivity loss in the day (7.175 minutes to refocus times 16 interruptions per day). This could mean that 24% of the work day is unproductive [6].
Assumptions
1. The tasks given across the three experiments were of equal difficulty level. 2. The average competency level of participants across the three experiments was the same. 3. The effect of any participant’s familiarity with the tasks i.e. Sudoku and word puzzle was neglected.
Hypothesis
Multitasking:
Multitasking is defined as act of dealing with more than one task at the same time [4]. Multitasking can be interpreted in many ways; we tried to address the multitasking which is more relevant to the work place. The ability of an individual to switch among various tasks, that is for example working on several tabs on the computer and switching among them according to the schedule. These scenarios are common in any working environment. We chose tasks such that they are contingent, whenever the individual switches from one to another he/she approaches it with a fresh eye. To be consistent with our definition we chose the tasks which do not involve much physical activity and require more mental ability.
We conducted our experiment with the hypothesis that multitasking improves performance for individuals (since that was the belief held majorly at COLLEGE).
Methodology
To validate the hypothesis
Task 1 (Sudoku Puzzle):
Sudoku (Exhibit 3) is played over a 9 × 9 grid, divided into 3 × 3 sub-grids called “regions”. The Sudoku puzzle begins with some of the grid cells already filled with numbers. The objective of Sudoku is to fill the other empty cells with integers from 1 to 9, such that each number appears exactly once in each row, exactly once in each column, and exactly once in each region. The numbers given at the beginning ensure that the Sudoku puzzle has a unique solution. Also when solving a Sudoku puzzle, solutions often come in waves. Multitasking can be appealing when one is stuck: one can work on the other task and can see the problem from a different angle after switching back [5].
This task as mentioned above requires more mental ability and is also not dependent on the gender which makes the multitasking experiment to have an unbiased atmosphere. The evaluation is based on allotting a single point for every correct number filled in the puzzle.
Task 2 (Word Puzzle):
This game (Exhibit 3) consists of a 4 by 4 matrix of letters from the English alphabet, players form meaningful English words (minimum 3 letter word) by joining the adjacent letters in any manner also without repeating the same element or letter of the matrix again [3]. Scores are allotted according to the number of letters in the meaningful words formed by the player. Thus, this is also a basic test of vocabulary which does not involve much of physical activity and again without any gender bias.
The objective for a participant was to maximize his combined score in the game. Experiments for testing the objectives are done under three conditions:
Experiment 1:
This experiment basically evaluates the performance of an individual when the tasks are being scheduled and is asked to follow the schedule. With an underlying aim of testing the performance in sequential tasking environment we gave a total of 10 minutes where the subjects under experiment are asked to solve the Sudoku puzzle of easy level in 6 minutes and the word puzzle for the next 4 minutes.
Experiment 2:
In this the subject is forced to switch between the two tasks for every two minutes such that the Sudoku is again given a total of 6 minutes and word puzzle a total of 4 minutes. This ensures that equal time is allotted to the tasks in both the experiments. This environment is close to multitasking and is a work place activity where individuals tend to switch between tasks according to the demand.
Experiment 3:
An opportunity is given to the subject that he is free to do what he wants or to schedule his task in the stipulate 10 minutes time such that both the tasks are being attempted and performed. This gives freedom to the individual to choose the tasks according to their will and perform it.
Results and Findings
We conducted our experiment with the hypothesis that multitasking improves performance for individuals (since that was the belief held majorly at COLLEGE). We conducted our experiments on 64 participants randomly distributed amongst the three experiments. The number of people selected for each type of experiment is mentioned below. Experiment 1 | Experiment 2 | Experiment 3 | 17 | 27 | 20 |

Our results (tabulated in Exhibit 4) indicate that scheduling of task does play a significant role in the success of the task. We find that multitasking lowers individual’s performance by a significant amount. The cost of switching between the two tasks is indeed high. When given the opportunity to freely manage their own time, individuals fare only a little better than the heavy multitaskers (Experiment 2). We also did not find any evidence of women outperforming men in Experiment 2, as other studies suggest.
The average score for a participant in Experiment 1 was 107.9, while it is 66.4 for Experiment 3 and 55.6 in Experiment 2. The score for any individual can be affected by his or her analytical abilities. To counter that bias we asked our participants to share their CAT score, which would be a measure of their analytic knowledge. Doing a linear regression (Exhibits 5, 6 and 7) on our data we find get the following result:
Score= -76.37 + 44.2*Experiment1 –7.65*Experiment2 – 5.32*Gender +1.50* CAT_Score
Experiment1, Experiment2, and Gender are categorical variables. Experiment3 is the base category. Regression Statistics | R Square | 0.331 | Adjusted R Square | 0.268 | Standard Error | 32.264 |

| Coefficients | Standard Error | t Stat | P-value | Lower 95% | Upper 95% | Intercept | -76.37 | 86.31 | -0.88 | 0.38 | -249.06 | 96.33 | Experiment 1 | 44.19 | 10.78 | 4.10 | 0.00 | 22.62 | 65.75 | Experiment 2 | -7.65 | 9.53 | -0.80 | 0.43 | -26.71 | 11.41 | Experiment 3 | 0.00 | 0.00 | 65535.00 | 0.00 | 0.00 | 0.00 | Gender | -5.32 | 9.38 | -0.57 | 0.00 | -24.08 | 13.44 | CAT score | 1.50 | 0.88 | 1.70 | 0.09 | -0.26 | 3.27 |
The p-statistics for the variables Experiment1 and Experiment3 is nearly equal to 0 – implying that both of these variables are significant in the regression. An important observation is that the coefficient of Experiment 2 (multitask) is only slightly more negative than that of Experiment 3 (free will). Does this imply: people are not good at estimating the time required for each task or do people have a tendency to multitask?
Interpretation
Our experiment has shown that the hypothesis that multitasking makes individuals more efficient can be rejected. The R2 for the linear regression is around 33% (adjusted R2 is 26.8%). The low R2 might be attributed to high variability in our data of 64 samples. Adding more variables might help or there might be inherently high variability in the data.
The coefficient for Experiment2 is -7.65, implying that individuals, when forced to switch unexpectedly will perform worse by 7.65 score units when compared to Experiment3 and by 51.85 (44.2 + 7.65) when compared to Experiment1. A high p-value for this variable is an area of concern for us. Ignoring the high p-value, this corroborates the fact that the hypothesis H0 can be rejected. The cost of switching between tasks is quite high. The cost is particularly high for Sudoku where the participant has to remember the current state of the game (as the next empty space he/she fills depends on his previous moves, when switching to word puzzle he loses this history of moves and his performance is affected).
The interesting observation comes when we compare coefficients of Experiment2 and Experiment3 (-7.65 vs.0). To understand more about this behavior we also noticed the switches each participant did when given a free choice (Experiment 3). While some participants did not even switch once – working on the game they thought they were better at – Does this mean people overestimate their capabilities? Participants switched an average 2.1 number of times. Compare this with the four interruptions they faced in Experiment2. Our survey found that people perceive multitasking improves efficiency (60% people believe said “yes” to multitasking improving productivity). Thus, it seems that people have a tendency to multitask although it negatively affects their performance. This experiments finding can also be attributed to the fact that people are unable to schedule their work optimally. And when provided with the schedule (as in Experiment1), they are able to utilize their abilities to the maximum.
Are women better at multitasking?
The average score for women in Experiment2 was 59.43 while that for men was 58.25. The average score for the two being so close implies that women are not significantly better multi-taskers than men( contrary to what popular literature has us believe). When we did a regression on the data for female participants we found the following information:
Score = 61.67 + 72.67*Experiment1 -2.24*Experiment2 | Coefficients | Standard Error | t Stat | P-value | Intercept | 61.67 | 13.87 | 4.45 | 0.00 | Experiment 1 | 72.67 | 24.02 | 3.03 | 0.01 | Experiment 2 | -2.24 | 18.90 | -0.12 | 0.91 |

This implies that when given a free hand (Experiment3) women work as badly as any average individual and their performance increase significantly when given sequential tasks (Experiment1). Although it must be kept in mind that we have given the participants analytically complex tasks, our experiment cannot answer if men and women perform equally while multitasking between other situations.
Conclusion
The experiment points to the drop in productivity when people try to handle many tasks in the same period of time, as opposed to performing tasks sequentially. The result points to the efficacy of certain practices in the academia and industry. At the same time, the results of the experiment are valid only in cases where the duration of time is held constant for tasks during multitasking and sequential task performance. It is possible that, in the workplace, such a scenario may not present itself. Future studies will need to be done to predict the effects in such cases. The performance of females and males are found to increase for sequential task execution. Factors like motivation in people affect performance and can explain the unexplained variance in the results. Further, the experiment was done on a student sample which may not effectively represent the larger population in terms of aptitude and logical skills. This doesn’t invalidate our findings since the parameter of CAT score is hardly significant to the total score. The kind of task given to the participants of the experiment can significantly affect the results. Tasks that require little focus actually show better performance in multitasking. Some of the participants who had previous experience in doing Sudoku/word puzzle showed little under-performance in multitasking scenario. Factors such as these haven’t been taken into account for the mathematical modeling and can be a point for future study.

APPENDIX

Exhibit 1 Time | Average time | Percentage | Weighted Score | <10 minutes | 5 | 7% | 0.35 | 10-20 minutes | 15 | 28% | 4.20 | 20-30 minutes | 25 | 36% | 9.00 | >30 minutes | 45 | 28% | 12.60 | Total | | 100% | 26.15 |

Exhibit 2 Time | Average time | % | Weighted Score | < 2 minutes | 1 | 13% | 0.13 | 2 - 5 minutes | 3.5 | 32% | 1.12 | 5 - 10 minutes | 7.5 | 33% | 2.475 | > 10 minutes | 15 | 23% | 3.45 | Total | | 100% | 7.175 min |

Exhibit 3
Figure 1 Sample sheet given to respondents

Word Puzzle Sudoku

Exhibit 4 Total Score = A + B | Word puzzle Score (A) | Sudoku Score (B) | Experiment 1 | Experiment 2 | Experiment 3 | Gender | CAT score | 13 | 13 | 13 | 0 | 0 | 1 | 0 | 97.7 | 14 | 0 | 14 | 0 | 0 | 1 | 1 | 97.1 | 23 | 8 | 15 | 0 | 1 | 0 | 1 | 98.97 | 23 | 15 | 8 | 0 | 1 | 0 | 0 | 96.03 | 25 | 6 | 19 | 0 | 1 | 0 | 1 | 97.1 | 30 | 4 | 26 | 0 | 0 | 1 | 1 | 99.6 | 32 | 23 | 23 | 0 | 0 | 1 | 0 | 96 | 33 | 23 | 10 | 0 | 1 | 0 | 0 | 98.67 | 34 | 29 | 5 | 0 | 0 | 1 | 1 | 88 | 35 | 23 | 12 | 0 | 1 | 0 | 1 | 96.8 | 36 | 10 | 26 | 0 | 0 | 1 | 1 | 98.74 | 38 | 10 | 28 | 0 | 1 | 0 | 0 | 89 | 39 | 25 | 14 | 0 | 1 | 0 | 1 | 98.8 | 40 | 18 | 22 | 0 | 0 | 1 | 1 | 97 | 42 | 30 | 12 | 0 | 1 | 0 | 1 | 96.6 | 47 | 32 | 15 | 0 | 1 | 0 | 1 | 99.87 | 48 | 30 | 18 | 0 | 0 | 1 | 1 | 99.87 | 49 | 39 | 10 | 0 | 1 | 0 | 1 | 98.25 | 50 | 43 | 7 | 0 | 0 | 1 | 1 | 99.28 | 51 | 25 | 26 | 0 | 1 | 0 | 1 | 96.97 | 52 | 49 | 3 | 0 | 1 | 0 | 1 | 99.28 | 56 | 36 | 20 | 0 | 1 | 0 | 1 | 99.8 | 57 | 43 | 14 | 0 | 0 | 1 | 0 | 96.03 | 58 | 24 | 34 | 0 | 1 | 0 | 1 | 95.12 | 60 | 41 | 19 | 0 | 0 | 1 | 1 | 96.6 | 60 | 44 | 16 | 0 | 1 | 0 | 1 | 96.6 | 62 | 57 | 5 | 1 | 0 | 0 | 1 | 99.2 | 64 | 35 | 29 | 0 | 0 | 1 | 1 | 96.97 | 64 | 55 | 9 | 0 | 1 | 0 | 1 | 96 | 67 | 34 | 33 | 0 | 1 | 0 | 1 | 98.86 | 69 | 45 | 24 | 1 | 0 | 0 | 1 | 96.8 | 69 | 67 | 2 | 0 | 1 | 0 | 1 | 99.2 | 69 | 37 | 32 | 0 | 1 | 0 | 0 | 97.7 | 73 | 58 | 15 | 0 | 1 | 0 | 1 | 97.8 | 75 | 61 | 14 | 1 | 0 | 0 | 1 | 65 | 76 | 45 | 31 | 0 | 1 | 0 | 0 | 96 | 78 | 46 | 32 | 1 | 0 | 0 | 1 | 99.6 | 78 | 71 | 7 | 1 | 0 | 0 | 1 | 98.25 | 78 | 59 | 19 | 1 | 0 | 0 | 1 | 97.1 | 78 | 29 | 49 | 0 | 1 | 0 | 1 | 94 | 82 | 33 | 49 | 0 | 1 | 0 | 1 | 98.8 | 83 | 80 | 3 | 1 | 0 | 0 | 1 | 95.12 | 83 | 41 | 42 | 0 | 0 | 1 | 1 | 98.86 | 84 | 81 | 3 | 0 | 0 | 1 | 0 | 97 | 85 | 36 | 49 | 0 | 0 | 1 | 0 | 98.67 | 86 | 66 | 20 | 0 | 1 | 0 | 0 | 99.7 | 87 | 70 | 17 | 0 | 1 | 0 | 1 | 99.4 | 91 | 78 | 13 | 1 | 0 | 0 | 0 | 96.97 | 91 | 46 | 45 | 0 | 1 | 0 | 0 | 97 | 94 | 61 | 33 | 0 | 0 | 1 | 1 | 99.43 | 99 | 90 | 9 | 0 | 0 | 1 | 0 | 99.7 | 101 | 65 | 36 | 1 | 0 | 0 | 1 | 99.21 | 102 | 53 | 49 | 1 | 0 | 0 | 1 | 98.8 | 108 | 78 | 30 | 0 | 1 | 0 | 1 | 98.52 | 114 | 65 | 49 | 0 | 0 | 1 | 1 | 94 | 123 | 112 | 11 | 0 | 0 | 1 | 1 | 98.52 | 125 | 109 | 16 | 1 | 0 | 0 | 1 | 98.8 | 125 | 102 | 23 | 1 | 0 | 0 | 0 | 94.89 | 130 | 95 | 35 | 1 | 0 | 0 | 1 | 96.6 | 132 | 118 | 14 | 1 | 0 | 0 | 1 | 99.8 | 154 | 136 | 18 | 1 | 0 | 0 | 1 | 97.8 | 164 | 137 | 27 | 1 | 0 | 0 | 1 | 99.4 | 168 | 119 | 49 | 0 | 0 | 1 | 1 | 99.4 | 187 | 164 | 23 | 1 | 0 | 0 | 0 | 99.7 |

Exhibit 5 Regression Statistics | Multiple R | 0.58 | R Square | 0.33 | Adjusted R Square | 0.27 | Standard Error | 32.26 | Observations | 64.00 |

Exhibit 6 ANOVA | | | | | | | df | SS | MS | F | Significance F | Regression | 5.00 | 30408.82 | 6081.76 | 7.30 | 0.00 | Residual | 59.00 | 61419.41 | 1041.01 | | | Total | 64.00 | 91828.23 | | | |

Exhibit 7 | Coefficients | Standard Error | t Stat | P-value | Lower 95% | Upper 95% | Lower 95.0% | Upper 95.0% | Intercept | -76.37 | 86.31 | -0.88 | 0.38 | -249.06 | 96.33 | -249.06 | 96.33 | Experiment 1 | 44.19 | 10.78 | 4.10 | 0.00 | 22.62 | 65.75 | 22.62 | 65.75 | Experiment 2 | -7.65 | 9.53 | -0.80 | 0.43 | -26.71 | 11.41 | -26.71 | 11.41 | Experiment 3 | 0.00 | 0.00 | 65535.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | Gender | -5.32 | 9.38 | -0.57 | 0.00 | -24.08 | 13.44 | -24.08 | 13.44 | CAT score | 1.50 | 0.88 | 1.70 | 0.09 | -0.26 | 3.27 | -0.26 | 3.27 |

References

1 The XX Factor in the Boardroom: Why women make better Directors
2 Fisher, H. (1999). The first sex: the natural talents of women and how they are changing the world
3 Molly Russ (2014). A survey of multitasking behaviours in organizations
4 http://www.oxforddictionaries.com/definition/english/multitask
5 http://puzzles.about.com/od/glossary/g/sudoku-puzzle.htm
6 http://www.wordament.com/
7 Multitasking Performance Deficits
8 Claudia Wallis. The MultiTasking Generation
9 http://www.forbes.com/sites/travisbradberry/2014/10/08/multitasking-damages-your-brain-and-career-new-studies-suggest/