This paper is best viewed in APA format. Please view it on Google Drive.
Previous research has observed that when people view their task-at-hand as requiring multitasking, they tend to perform better than people who do not. A potential explanation for this is that multitasking tasks seem more challenging, increasing engagement levels that boost performance. We attempted to determine whether multitasking perception and perceiving the task as simple (simplicity perception) affect performance on a simultaneous transcription and comprehension task, and whether or not there is an interaction between those two factors. In a 2x2 between-subjects factorial design, we assigned participants to transcribe a video while watching it. We gave them explicit directions that correspond with each condition in order to affect their perception. Afterwards, they completed a questionnaire regarding the contents of the video, which we used to determine their performance. We found no statistically significant main effect of multitasking perception, main effect of simplicity perception, or interaction between the two. We have failed to find sufficient evidence to 1) reject that multitasking perception and simplicity perception do not have an effect on performance and 2) conclude that the effects of one factor depend on the levels of the other.
Multitasking is a skill many practice in everyday life, from scrolling through social media while watching television to taking notes while listening in on an important business call. What people do not always realize is that doing multiple things totally simultaneously is not possible: we merely switch rapidly between tasks (Srna et al., 2018). Despite this finding, however, it may be possible that the perception of multitasking itself boosts performance, potentially due to the fact that top-down processing can impact behavior, especially from feeling challenged. Contrary to popular belief, feeling stressed from challenge can actually facilitate performance (Kofman et al., 2006). Thus, knowing how multitasking operates in conjunction with perception and challenge can give us more insight as to how to be an efficient worker.
To clarify how the act of multitasking affects people, past studies have shown that humans often fail to process multiple stimuli and perform a task simultaneously. Junco et al. (2012) conducted research about how multitasking during class affects the performance of students. They wanted to answer if technology use is correlated to GPA. They hypothesized that students who use ICTs (social media) during class will have lower final GPAs because the mind won’t be able to process many things simultaneously. In that case, the students who used ICTs frequently will have relatively lower GPAs. The alternative is that ICTs usage wouldn’t negatively impact GPA because it can add educational value. Then, ICTs use factor should not be negatively correlated to GPA. The independent variable is the frequency of multitasking, measured by a survey question about time spent on ICTs. The dependent variable is GPA, taken from transcripts. Researchers surveyed students from Northeastern universities, asking how frequently the students used ICTs. Results show that when students used Facebook or texted, GPAs were significantly lower. In a broader sense, this adds further evidence to the answer that attempting to divide attention reduces performance across active tasks. This study is relevant because it demonstrates that multitasking itself does not improve performance. This leaves room for researchers to examine if the perception of multitasking may improve performance without concern that multitasking itself would be a confound to the result. One such study is below.
Srna et al.’s (2018) study wanted to answer this broad question of whether perceived multitasking improves performance on tasks. They first sought to answer a more specific question: if a video transcribing task was framed to be a combination of multiple tasks, would participants perform better? The researchers hypothesized that when participants view themselves as multitasking, their performance would be enhanced because challenge elevates engagement. The alternative exists that the perception of multitasking results in no difference in performance because top-down processing may not influence engagement. If the perception of multitasking enhances performance, then participants under the multitasking-framed condition should perform better. If it doesn’t, there should be no significant difference in performance. Performance is the dependent variable, measured by words transcribed, transcription accuracy, and quiz score. The independent variable is the perception of the assigned task as a multitasking task or not determined by researchers’ instructions. The participants were randomly assigned to either the perceived multitasking condition or the non-multitasking condition. Participants transcribed the video as it played, and they took a quiz on the material. Participants in the perceived multitasking condition scored statistically significantly higher on words transcribed, transcription accuracy, and the comprehension pop-quiz. From this, we can conclude that if a video transcribing task was framed to be a combination of multiple tasks, participants do indeed both comprehend more information and transcribe more accurately. This study is relevant because we are able to see that the perception of multitasking can indeed influence performance.
Studies of challenge and stress on performance provide insight that may explain this phenomena. Kofman et al.’s (2006) study wanted to answer if completing a reaction time based task (classic Stroop test and task-switching test) would be quicker if participants were stressed. They hypothesized that stress would reduce switching cost in the task-switching test and reduce Stroop interference in the Stroop test. This is because the physiological effects of stress would boost their performance. If this were true, then we’d expect reaction times in both tests to decrease. The alternative exists that stress does not actually boost performance because stress is largely an emotional state and cannot have any effect on skill. If this were true, then we wouldn’t expect any difference in performance. In a between-subjects crossover design, the independent variables are stress levels and test type. Participants took one test stressed and the other test unstressed. The dependent variable is the correct response time measured. In session one, half of the participants did task-switching and half did the Stroop test, assigned through counterbalancing. In the second session, right before major exams, the participants completed the test that they did not do in the first session. In this session, researchers purposely gave an “anxiety-inducing” questionnaire. In both sessions, researchers measured their score on those tests as well as their levels of stress. Results indicate that under stress, response times have significantly decreased for both test types, answering that stress does indeed lower response time on response based tests. This study is insightful because if stress can facilitate performance on executive function tasks, a.k.a. functions that are crucial to learning, then the perception of multitasking can very well facilitate performance because perceiving a task to be a multi-tasking task can make it seem more difficult and thus induce stress.
A second example of how perception affects multitasking is provided by Levy et al. (2001). They wanted to study if equalizing task priorities (focusing equally on each task) can eliminate dual-task interference (poorer performance from dividing attention). To be more specific, they wanted to know whether giving equal priority to two tasks in dual-tasking as a way of eliminating interference is instruction dependent. Participants were asked to do two tasks. The first asked the participants to indicate the pitch of a tone by pressing keys in response to a low, medium, or high tone, respectively. The second task asked participants to indicate the position of a disk. The experiment was a within-subject design with three levels of the independent variables: 1) single-stimulus condition in which only one type of stimulus (disk or tone) was presented, 2) OR condition in which either a tone or a disk was presented, and 3) AND condition in which participants will respond to both stimuli presented simultaneously. The dependent variable is response time: the time it takes to respond correctly. They hypothesized that the dual-task interference will decrease when equal priority is given to the tasks because more effort will be put into each task. The alternative would be that equalizing priority doesn’t eliminate dual-tasking interference even with instruction because top-down processing may not affect performance. If the former were true, then the response time for the OR and AND condition should be significantly less from the single stimulus condition because a slower response time indicates a greater interference. If the latter was, then response times wouldn’t be faster. Results indicate that reaction times for the single-stimulus condition were significantly faster than the AND condition or the OR condition, and were significantly slower in the AND condition than in the OR condition. This suggests that performance on multitasking depends on the type of task presented, and equalizing task importance cannot eliminate the dual-task interference. The findings of this study points out a very important feature that the type of task might play a role in multitasking performance. If task type affects multitasking performance, it could be suggested that the positive effect of multitasking perception could also be task-dependent and the effect might not be generalized to all kinds of tasks.
Individual differences might also be a determining factor on multitasking performance. Morgan et al. (2013) studied the question of whether different people multitask differently. They proposed the question of how cognitive factors predict multitasking ability and adaptability, hypothesizing that different people with different cognitive strength will perform differently on multitasking because strength helps performance. If individual differences in cognitive ability can affect multitasking ability, then the scores on cognitive tests will correlate with the multitasking performance. The alternative would be that individual difference in cognitive strength does not affect ability on multitasking because tasks do not rely on strength. If this were true, strength scores will not correlate with performance. In the study, 32 participants were first asked to complete a battery of cognitive tests, which measure ability in different cognitive aspects, such as scholastic aptitude. Then the participants were asked to complete a series of experiment trials which stimulate a multitasking condition in which participants need to constantly focus on different indicators on the screen. Then the difficulty of the flight simulator gradually increased. Results of data analysis stated that there is a significant correlation between scholastic aptitude and general multitasking ability as well as between working memory and general multitasking ability. Thus, there is a significant correlation between spatial manipulation and adaptability. In conclusion, scholastic aptitude and working memory can predict multitasking abilities. The findings of this study suggests that the ability to multitask can vary between individuals with different cognitive strengths. Therefore, it could lead to future research about the extent of the positive effect of multitasking perception on different groups of people with different learning abilities.
These studies supply us with good ground knowledge about multitasking and cognition. Multitasking itself is not necessarily beneficial, but compared to not perceiving a task as multitasking, the perception of multitasking will indeed lead to relatively improved performance. The physiological effects of stress or challenge improves performance, and believing they need to multitask makes participants feel more challenged. People with higher cognitive strength multitask better than those with lower cognitive strength. Different tasks also could potentially lend better to multitasking performance. This is important to consider when designing an experiment because 1) different people may perceive challenges differently and 2) not every type of task may be improved with the perception of multitasking or challenge.
A gap that exists in this research is what if the perception of challenge was manipulated along with multitasking perception to test how they would affect performance and whether the two have an interaction? There could be one or more main effects, or none at all. If there is any main effect, then that means that corresponding factor does influence performance. If there are none, then none of the factors influence performance. The two factors could also have an interaction, or they may not. If the former was true, then that means that the effects of multitasking perception would depend on levels of simplicity perception and vice versa. If the latter were true, then it would indicate no dependence. We hypothesize that there are main effects of both multitasking and simplicity perception, and that there is an interaction between the two terms.
A total of 24 participants completed the study in full. They were sampled from convenience sampling, including peers in an undergraduate class at UCSD. Other demographic info was not collected.
There were two independent variables in our study. The first was multitasking perception, with two levels: present and unmentioned. The second was simplicity perception, also with two levels: present and unmentioned. Both variables were operationally defined through explicit written instructions and manipulated between-subjects, making this a 2x2 between-subjects factorial design. Participants were randomly assigned to each condition. We measured performance, operationally defined as the participants’ score on a ten multiple choice questionnaire on their comprehension of a video.
Participants interacted with an internet-browser-based survey hosted by Qualtrics. Overall, the survey appearance involved dark legible text on a light background. There were four sets of written directions, each corresponding with a condition. They either directed the participant to multitask or left that unmentioned, and explained that the study was easy or also left that unmentioned. They informed the participants of the video transcription task and the questionnaire that would follow. Instructions were short, restricted to about five to six concise sentences. On the same page as instructions, there was a ninety-second embedded video on a simple but uncommonly taught subject, such as daylight saving time. This is to ensure that anyone could understand the contents, but it would be unlikely for them to already have previous knowledge of the subject. The video could not be rewinded or fast-forwarded. Captions and speed adjustments were also turned off to prevent cheating or skewing results. On the same page as the embedded video, there was a text box where participants could type out their transcription, placed under the video in a location where participants could see both the video and their transcription simultaneously.
They could then press an arrow at the bottom of the page to advance to the questionnaire page, which was ten specific questions with answers that were mentioned explicitly in the video. They were presented in a list form, and questions ranged from having three to four choices with only one correct answer for each question. Participants could also only select one choice per question by clicking on their preferred answer. To turn in the quiz, participants had to press another arrow at the bottom of the page, which led them to a completion screen. We graded the quiz on how many participants answered correctly. We decided to keep the quiz short and as the only measure of performance because it’s the easiest to code and run analyses with. There was no time limit on any section of the survey.
Participants were given a link to access the study. Clicking that link opens our survey page. There is a brief thank-you message before participants advance to seeing the aforementioned written instructions by clicking on a forward arrow. Aside from conditional differences, all instructions told participants to transcribe a video while watching it simultaneously and then proceed to a questionnaire. To move on after the instructions, participants clicked on another forward arrow, leading to the video page. They can then press play on the video and also type in the transcription box on the same page. There is no time limit, and participants can take as much time as they feel is necessary. Another forward arrow needs to be clicked to move on to the quiz page. Participants read each question and selected an answer for each question. After a last forward arrow, participants reach a completion confirmation page which marks the end of their participation.
Due to present circumstances, participants completed the survey wherever they preferred, on any device that had access to an internet browser. Whether or not they were alone is impossible to determine. We did not control what time of day they had to complete the survey because it was deemed irrelevant to our research question.
To answer if any main effects or interactions are present, we analyzed the factors of multitasking perception and simplicity perception with a factorial ANOVA test. Results show that there is no significant main effect of simplicity perception, F(1,16)=0.06, p=.801. On average, no simplicity perception (M=5.98, SD=0.51) scored 0.19 lower than the presence of simplicity perception (M=6.17, SD=0.55). There is no significant main effect of multitasking perception, F(1,16)=0.04, p=.850. On average, no multitasking perception (M=6.00, SD=0.53) scored 0.14 lower than the presence of multitasking perception (M=6.14, SD=0.53). This can be observed in Figure 1.
Lastly, the effect of simplicity perception is different depending on the presence of multitasking perception and vice versa, but not significantly. There is no significant interaction of simplicity perception and multitasking perception, F(1,16)=0.41, p=0.531. After running a pairwise Tukey test, there are no significant pairwise interactions.
Perception's Effect on Performance
Note. Error bars represent standard error.
The factorial ANOVA test found that there were no main effects of either multitasking or simplicity perception and that there was no interaction between the two. The results of our study do not support the findings made by Srna et al. (2018) and Kofman et al. (2006). This suggests that perhaps multitasking perception and simplicity perception do not amplify performance and that the effects of one factor do not depend on the levels of the other. From these findings, our original hypotheses were not supported.
Theoretically, perhaps the notable positive effects that feelings of challenge provide can not be able to be applied to every situation. From past research, stress in general promotes response times (Kofman et al., 2006), but our study was not a test of response time. Furthermore, participants in Kofman’s study were not multitasking. In terms of comprehension performance while multitasking, the before observed effects of stress could be dampened or irrelevant altogether. Our results on multitasking perception also suggest more research needs to be done on the subject since it directly contrasts the findings of Srna et al.’s (2018) original study.
In relation to real life applications, our findings suggest that it’s not more or less practical to work under stress, since we found no significant benefits of being challenged. This also applies to multitasking perception. Cognitively, whether an individual frames their work as requiring multitasking or not doesn’t improve their performance. Therefore, challenged or not, perceptually multitasking or not, it doesn’t matter. There are no significant differences.
That being said, there were quite a few limitations to our study to consider when interpreting the results. For starters, our sample size was incredibly small. On average, each condition only had six participants. This is likely to be highly unrepresentative of the population, which makes us cautious to generalize our findings. A small sample size also decreases our statistical power. Next, participants did not complete the study in a controlled setting. Because they were allowed to take it anywhere, at any time, on any device, it’s possible that many were rushing and were distracted during it. Therefore, their answers may not be accurate to their full ability or genuine. We had no attention checks within our study to ensure that participants were paying attention. Lastly, I’m afraid our instructions were not as clear as they could have been. A couple of incomplete responses had not written a transcription, or asked what a transcription was in the response box. We can’t trust that our study is generalizable if it’s possible that some participants didn’t even understand the instructions. Because of these limitations, we are hesitant to compare it with equal standing to the studies we’ve cited.
In future studies, it would be imperative to have a larger sample size to address issues of representation and statistical power. The design and materials of the study itself would have to be modified to have clearer instructions and procedures. It would be ideal to hold the study in person in a professional setting, such as a lab. This way, participants can ask questions if needed and as a researcher, we can ensure that participants are on the same page as us and completing the study undistracted. A new question raised by our results is that could the effects of stress on performance vary based on task type? For example, a comprehension task versus a response time task? This would need to be addressed with a design that studies the effects of stress during different tasks.
Junco, R., Cotten, S. (2012). No A 4 U: The Relationship Between Multitasking and Academic Performance. Computers & Education, 59(2), 505-514. https://doi.org/10.1016/j.compedu.2011.12.023.
Kofman, Ora & Meiran, Nachshon & Greenberg, Efrat & Balas, Meirav & Cohen, Hagit. (2006). Enhanced Performance on Executive Functions Associated with Examination Stress: Evidence from Task-switching and Stroop Paradigms. Cognition & Emotion - Cognition Emotion, 20, 577-595. https://doi.org/10.1080/02699930500270913.
Levy, J., & Pashler, H. (2001). Is Dual-task Slowing Instruction Dependent? Journal of Experimental Psychology: Human Perception and Performance, 27(4), 862-869. http://dx.doi.org/10.1037/0096-1523.27.4.862.
Morgan, B., D'Mello, S., Abbott, R., Radvansky, G., Haass, M., & Tamplin, A. (2013). Individual Differences in Multitasking Ability and Adaptability. Human Factors, 55(4), 776-788. http://dx.doi.org/10.1177/0018720812470842.
Srna, S., Schrift, R. Y., & Zauberman, G. (2018). The Illusion of Multitasking and its Positive Effect on Performance. Psychological Science, 29(12), 1942–1955. https://doi.org/10.1177/0956797618801013.