Single-Tasking vs. Multitasking: A Data Breakdown of What the Research Actually Shows in 2026

TL;DR. “Multitasking” is three different behaviours, and the evidence is different for each. Switching between tasks costs roughly 0.6 to 1.1 seconds per switch in the classic laboratory study (Rubinstein, Meyer and Evans, 2001), more when the rules are complex. Doing two demanding things at the same moment is something only 2.5 percent of 200 tested students managed without a performance drop (Watson and Strayer, 2010), although simple, heavily practised task pairs can be combined at a cost of 4 to 11 milliseconds (Schumacher and colleagues, 2001). Habitual media multitasking is only weakly linked to poorer cognitive control (z = .138 across 118 assessments), and on objective tests the link is no longer significant after bias correction (Parry and le Roux, 2021). The largest verified losses appear where someone else sets the pace: lecture comprehension was 11 percentage points lower for laptop multitaskers and 17 points lower for students who merely sat in view of them (Sana, Weston and Cepeda, 2013). At the desk, interrupted work was finished slightly faster, in 20.31 to 20.60 minutes against 22.77, but with clearly higher stress (Mark, Gudith and Klocke, 2008). A CEOtudent calculation shows why the famous “40 percent” figure cannot come from switch costs alone: at laboratory rates, about 275 switches a day add up to 3 to 5 minutes. The expensive part is the detour, not the switch. The CEO move is to decide which task pairs are allowed to run in parallel. The student move is to measure your own cost for two weeks.

This article is the comparison companion to attention residue in the AI era. That piece explains what a switch leaves behind in your head. This one answers a different question: when you line up the studies side by side, where does single-tasking clearly win, where is the difference small, and where is multitasking harmless?

Why does this comparison matter in 2026?

Because the default working day is now built out of interruptions. Microsoft’s Work Trend Index special report of June 2025, based on aggregated Microsoft 365 activity up to 15 February 2025 and a survey of 31,000 knowledge workers in 31 markets, reports that employees are interrupted every 2 minutes during core hours by a meeting, an email or a notification. The same report counts 117 emails received per day and 153 Teams messages per weekday for the average worker, and 48 percent of employees (52 percent of leaders) say their work feels chaotic and fragmented.

Two details in the methodology matter before anyone repeats those numbers. The headline total of 275 interruptions a day is counted over a 24-hour day, and it describes the top 20 percent of users by volume of incoming pings, not the typical employee. It also counts incoming signals, not observed shifts of attention. Microsoft’s annual report of 5 May 2026 contains no newer interruption figure, so the 2025 telemetry remains the latest available from that source.

A leader allocating capital would not accept “everything, all the time” as a strategy. Attention is the scarcer asset, and the allocation question is the same: which activities may share a time slot, and which must have it to themselves? The rest of this article answers that with data. For the wider allocation framework, see the attention portfolio.

What exactly counts as multitasking?

Researchers use one word for three different things, and popular advice usually mixes them.

1. Task switching. You do one thing, then another, then return. Only one task is active at a time, and the cost is paid at each change.
2. Concurrent dual-tasking. Two tasks overlap in the same seconds, such as driving while holding a conversation.
3. Media multitasking as a habit. A trait measured by questionnaire: how often a person uses several media streams at once. Studies then ask whether heavy and light media multitaskers differ on cognitive tests.

A finding about one of these does not transfer automatically to the others. The sections below take them in turn.

How large is the switching cost in the laboratory?

The most cited study is Rubinstein, Meyer and Evans (2001), published in the Journal of Experimental Psychology: Human Perception and Performance. In four experiments with 12, 36, 36 and 24 undergraduates, participants either repeated one task or alternated between two, classifying geometric patterns or solving arithmetic problems. The mean switching-time costs were 975, 653, 614 and 1,096 milliseconds. Complex rules added a further 823 to 972 milliseconds compared with simple ones, a cue announcing the next task reduced the cost, and switching to a less familiar task took longer than switching to a familiar one.

So the measured cost is real, it scales with complexity, and it is on the order of a second per switch for these tasks. What the paper does not contain is any estimate of lost productive time. The widely repeated claim that switching costs up to 40 percent of productive time traces to a public summary page of the American Psychological Association dated 20 March 2006, which reports it as something David Meyer has said. The same page describes individual switch costs as small, sometimes a few tenths of a second.

Why is there a cost at all? A 2006 brain-imaging study in Neuron (Dux, Ivanoff, Asplund and Marois, 14 participants) found that when two decisions had to be made close together, the second was postponed while the first was selected, with accuracy essentially unchanged (94.6 percent against 95.2 percent). The authors describe a network of frontal areas acting as a central bottleneck. The decision stage queues; it does not run in parallel.

Can practice remove the cost?

For simple tasks, largely yes. Schumacher and colleagues (2001, Psychological Science) gave 8 participants two basic reaction tasks that used different senses and different responses: saying a word in response to a tone, and pressing a key in response to a visual position. By the fifth session, after 2,064 trials on each task, the dual-task cost had fallen to 4 milliseconds on one task and 11 on the other. The authors limit the claim to basic choice-reaction tasks and to at least some participants.

For demanding tasks, almost nobody escapes. Watson and Strayer (2010, Psychonomic Bulletin and Review) tested 200 undergraduates in a driving simulator, alone and while doing a hard memory-and-arithmetic task over a hands-free phone. Five of them, 2.5 percent, showed no decrement on any measure. A simulation put the rate expected by chance at 0.16 percent, so the authors treat these “supertaskers” as a real but rare group, and describe the remaining 97 percent and more as unable to combine the two without substantial costs.

There is also evidence that everyday practice matters. In a sample of 1,511 science-centre visitors aged 7 to 70, Matthews, Mattingley and Dux (2022, Scientific Reports) found that people who reported more everyday technology multitasking performed better, not worse, on a dual-task test. The relationship was significant only from age 7 up to 29.25. The study is cross-sectional, so it cannot say which way the influence runs, but it contradicts the idea that heavy media multitaskers are always worse at juggling.

Where is multitasking most expensive?

Where the pace is set by someone or something else. Three bodies of evidence agree.

Lectures. Sana, Weston and Cepeda (2013, Computers and Education) ran a simulated classroom with a 45-minute lecture and a comprehension test. Students assigned to do online side tasks on a laptop scored 0.55 against 0.66 for those who only took notes (20 per group). In a second experiment, students who did not multitask but sat in view of a multitasking peer scored 0.56 against 0.73 (19 per group). The authors report these gaps as 11 percent and 17 percent; they are differences in the share of correct answers, so 11 and 17 percentage points is the precise wording. The second number is the striking one: in these two experiments the gap for bystanders was larger than the gap for the multitaskers themselves.

Reading. A meta-analysis of 22 randomized studies (Clinton-Lisell, 2021, Journal of Research in Reading) found that multitasking lowered reading comprehension overall (g = -0.28). The harm was concentrated in studies where the experimenter fixed the reading pace (g = -0.54). When readers controlled their own pace, there was no reliable effect on comprehension (g = -0.14, p = .10), but reading took longer (g = 0.52). In other words, with control over the pace people protect comprehension by paying in time.

Driving. Strayer, Drews and Crouch (2006, Human Factors) compared simulator drivers talking on a phone with the same drivers at a blood alcohol concentration of 0.08 percent. According to the paper’s abstract, conversation on either a handheld or a hands-free phone delayed braking and increased accidents, and the authors conclude that the impairment can be as profound as that of driving drunk. The University of Utah’s summary of the study adds the detail: 40 participants, braking 9 percent slower while on the phone, and three rear-end collisions, all during phone conversations and none in the alcohol condition. Those percentages come from the university’s release, not from the paper text we could read.

The common thread is control over timing. A lecture, a meeting, a road and a live conversation do not wait for you.

Are heavy media multitaskers cognitively worse off?

Much less than the first study suggested. Ophir, Nass and Wagner (2009, PNAS) surveyed 262 students and then compared small groups of heavy and light media multitaskers, 15 to 22 people per group. The heavy group had a switch cost 167 milliseconds larger and was 77 milliseconds slower when distractors were present. The paper became the basis for the claim that media multitasking damages attention.

Later work shrank the effect:

• Wiradhany and Nieuwenstein (2017) ran two direct replications containing 14 tests with an average power of 0.81. Five were significant in the predicted direction and two survived a Bayesian analysis. Their meta-analysis of 39 effect sizes gave a weak pooled effect of d = .17, which fell to between .001 and .07 and lost significance after correction for small-study bias.
• Parry and le Roux (2021) pooled 46 studies and 118 assessments. The overall association was small (z = .138). It was larger when people rated their own attention (z = .200) than on objective performance tests (z = .091), and the performance estimate was no longer significant after trim-and-fill correction (z = .032). By function, the association was absent for task management, the category that includes switching itself (z = .031). Nearly three quarters of the studies (72.88 percent) sampled students.
• The newest pooled estimate we could verify, a 2025 three-level meta-analysis in Current Psychology covering 33 studies and 36,861 participants, reports a correlation of r = 0.19 between media multitasking and poorer attention, and notes that the result depends on how attention is measured. Only the abstract was accessible to us.

The fair summary: people who multitask heavily with media report more attention problems than objective tests detect, the measured association is small, and nothing in this literature establishes cause. For the broader picture on screens, see screen time and cognitive performance.

What happens at work when you are interrupted?

Three studies led by Gloria Mark of the University of California, Irvine and one experiment from the City University of New York give the clearest picture.

Interrupted work can be faster, at a price. Mark, Gudith and Klocke (2008, CHI) had 48 participants answer a set of emails with no interruptions or with interruptions about every two minutes. With the interruption time itself subtracted, the task took 20.31 and 20.60 minutes in the two interrupted conditions against 22.77 without interruptions. Error counts did not differ. People wrote longer emails when uninterrupted, and reported more stress when interrupted (9.46 and 9.13 against 6.92 on a 20-point scale), more frustration (6.63 and 6.48 against 4.73), more time pressure and more effort. People compensate for interruptions by working faster and writing less, and they pay in strain after only 20 minutes.

A moderate amount of switching can beat none. Adler and Benbunan-Fich (2012) let 103 students switch freely among six problem-solving tasks while 102 worked through them in a fixed sequence. On average the two conditions did not differ. Within the switching group, however, productivity followed an inverted U: medium switchers scored 71.20, low switchers 59.17 and high switchers 56.66, against 60.83 for the no-switching control. Accuracy told a different story at the top end: high switchers scored 30.89 against 41.90 for the control. Switching level was self-chosen, so this is not a causal estimate, but it warns against treating zero switching as the goal.

The delay after an interruption is long. In a field study of 24 information workers observed for more than 700 hours (Mark, Gonzalez and Harris, 2005, CHI), 57.1 percent of work segments were interrupted. Of the interrupted work, 77.2 percent was resumed the same day, on average 25 minutes and 26 seconds later and after work in 2.26 other projects. The spread is very wide (a standard deviation of 54 minutes 48 seconds). Work that people interrupted themselves took longer to resume (29 minutes 1 second) than work interrupted by others (22 minutes 37 seconds).

Screen focus is short. Logging 40 employees of one technology company for about 12 working days, Mark and colleagues (2016, CHI) found a mean of 47.0 seconds (median 40.2) before a person changed the window in focus, and an average of 272.7 switches between applications per day. Shorter focus went with lower self-rated productivity at the end of the day. The measure is window switching, not a biological attention span.

Which famous numbers survive a check against the primary source?

The table below is a CEOtudent audit. Each popular claim was traced to the document that first printed it.

Popular claim	What the primary source says	Verdict (CEOtudent audit)
Multitasking costs 40 percent of productive time	The figure does not appear in Rubinstein, Meyer and Evans (2001). It is a remark attributed to Meyer on a 2006 APA summary page.	An estimate in conversation, not a measured result
It takes 23 minutes and 15 seconds to refocus	No paper prints it. It comes from a 2006 Gallup interview with Mark about 36 observed workers, and describes the average delay before interrupted work was resumed, with about two other tasks in between. The published figure is 25 minutes 26 seconds (24 workers).	Real observation, wrong label: resumption delay, not refocus time
Attention span is now 47 seconds	Mean time before changing windows was 47.0 seconds in 40 employees of one company (Mark and colleagues, 2016).	Accurate as a window-switching average, not a general attention span
Multitasking lowers IQ by 10 points, more than marijuana	Never published. The researcher’s own 2010 note describes eight employees whose test scores fell from 143.38 to 132.75 under distraction, says the study was misrepresented and that the drug comparison was not his.	Unsupported
Only 2 percent of people can multitask	The paper says 2.5 percent, five of 200 students, on one driving-plus-memory task pair (Watson and Strayer, 2010).	Close, but specific to one hard task pair
The brain cannot multitask	Imaging shows a decision bottleneck that postpones the second of two overlapping decisions (Dux and colleagues, 2006). Simple practised tasks were combined at a cost of 4 to 11 ms (Schumacher and colleagues, 2001).	Overstated: true for two demanding decisions, not for every pairing

What do the switch costs actually add up to?

If the 40 percent figure were a consequence of switch costs, the arithmetic should get somewhere near it. It does not. The table is a CEOtudent calculation: it multiplies two independent counts of daily switches by the lowest and highest mean switch costs in the 2001 experiments.

Daily count (source)	At 614 ms per switch	At 1,096 ms per switch
272.7 application switches (Mark and colleagues, 2016; mean of 40 employees)	2.8 minutes	5.0 minutes
275 incoming pings (Microsoft, 2025; top 20 percent of users, 24-hour day)	2.8 minutes	5.0 minutes

Forty percent of an eight-hour day is 192 minutes. Five minutes is about 1 percent of that day. Even if every one of roughly 275 daily events triggered a full laboratory-sized switch cost, the total would be about one fortieth of the famous figure.

This is a rough illustration, and its limits are worth stating. The laboratory tasks were simple classifications and arithmetic, not knowledge work. A ping is not always a switch. The two daily counts measure different things and their closeness is a coincidence. But the conclusion does not depend on precision: the per-switch reaction-time cost is too small, by more than an order of magnitude, to explain large productivity losses.

The loss has to come from somewhere else, and the field data point to where. An interruption is rarely a one-second event. It is a detour: 25 minutes 26 seconds on average before the original work was picked up again, with two other projects in between, and 22.8 percent of interrupted work not resumed that day at all (100 minus the 77.2 percent that was). The switch is cheap. The detour is expensive. And the experiment in which interruptions did not slow people down still found them more stressed after 20 minutes.

What does the verified data say? Key studies at a glance

Study	What was compared	Sample	Reported result
Rubinstein, Meyer and Evans, 2001	Alternating two tasks against repeating one	4 experiments, 12 to 36 students each	Mean switch costs of 614 to 1,096 ms; 823 to 972 ms more for complex rules
Schumacher and colleagues, 2001	Two simple reaction tasks together after practice	8 participants, 2,064 trials per task	Dual-task cost of 4 ms and 11 ms
Watson and Strayer, 2010	Simulated driving alone against driving plus a memory task	200 students	2.5 percent showed no dual-task cost; chance rate 0.16 percent
Matthews, Mattingley and Dux, 2022	Everyday media multitasking against dual-task performance	1,511 people aged 7 to 70	More media multitasking went with better performance up to age 29.25
Sana, Weston and Cepeda, 2013	Laptop multitasking in a 45-minute lecture	20 per group; 19 per group	Scores 0.55 against 0.66; bystanders 0.56 against 0.73
Clinton-Lisell, 2021	Reading with and without a second task	22 randomized studies	Comprehension g = -0.28; -0.54 at fixed pace; -0.14 self-paced (not reliable); time g = 0.52
Parry and le Roux, 2021	Media multitasking against cognitive control	46 studies, 118 assessments	z = .138 overall; .200 self-report; .091 performance tests, .032 after correction
Mark, Gudith and Klocke, 2008	Email task with and without interruptions	48 participants	20.31 and 20.60 minutes against 22.77; stress 9.46 and 9.13 against 6.92
Adler and Benbunan-Fich, 2012	Free switching against a fixed sequence	205 students	Productivity 71.20 (medium) against 60.83 (none); accuracy 30.89 (high) against 41.90 (none)
Mark, Gonzalez and Harris, 2005	Observed interruptions in office work	24 workers, more than 700 hours	57.1 percent of segments interrupted; 77.2 percent resumed the same day after 25 minutes 26 seconds

The Task-Pairing Matrix: what may run in parallel?

The matrix below is a CEOtudent editorial framework. It sorts task pairs by the two variables that separate large costs from small ones in the studies above: whether both tasks need decisions, and who controls the pace. It has not been tested as a package; each row names the evidence it draws on.

Pairing	Example	What the evidence shows	Rule
Two decision streams, pace set by someone else	A call while driving; messaging during a lecture or meeting	Delayed braking and more simulator crashes; lecture scores down 11 and 17 points; comprehension g = -0.54 at fixed pace	Single-task
Two decision streams, pace set by you	Writing a report while answering chat	Comprehension effect small when self-paced (g = -0.14), but time rises (g = 0.52) and so does stress (9.13 to 9.46 against 6.92)	Alternate in planned blocks and cap the switches
One decision stream plus a practised or low-demand activity	Walking during a call; shading shapes while listening to a dull message	Dual-task cost of 4 to 11 ms after practice on simple tasks; in one small study doodlers recalled 29 percent more	Allowed
Any pairing where an error cannot be undone	Sending money, publishing, operating machinery	More than 97 percent of tested people incurred substantial costs on a demanding pair	Single-task and add a check

The doodling result comes from a study of 40 participants (Andrade, 2010) in which half shaded printed shapes while monitoring a monotonous recorded message. It is small, and only its abstract was available to us, but it marks the opposite end of the range: a low-demand side activity can help when the main task is boring.

How should you run your day on this evidence?

The CEO move: allocate, then enforce.

1. Classify your recurring pairs. List the five combinations you do most often (meeting plus inbox, writing plus chat, commute plus calls) and place each in a row of the matrix.
2. Protect every fixed-pace input. Meetings, lectures, live calls and driving get the device closed. This is also a courtesy: in the lecture study, people sitting in view of a multitasker scored lower too. The cost side of meetings is covered in the true cost of meetings.
3. Budget switches, not only minutes. The medium switchers were the most productive group in the one experiment that measured it, and the high switchers the least accurate. Aim for a few planned changes per work block, not zero and not constant.
4. Shrink the detour. Since the expensive part is the 25-minute return path, write one line about where you are and what comes next before leaving a task. Self-interruptions took longer to recover from than external ones, so the most valuable interruptions to remove are your own.
5. Reduce the incoming stream. Fewer pings mean fewer candidate detours. A structured method is in the notification audit protocol, and a way to measure the result is the deep work deficit index.

The student move: test it on yourself for two weeks. The protocol below is a CEOtudent suggestion, not a validated instrument.

• Week 1, baseline. Work as usual. Each day record three things for one recurring task: minutes to finish, errors found on review, and end-of-day stress on a 1 to 20 scale.
• Week 2, single-task blocks. Do the same task in blocks with notifications off and one window open. Record the same three numbers.
• Decide with a rule set in advance. Keep the blocks if time is equal or lower and stress falls by 2 points or more, which is about the size of the gap in the 2008 experiment (6.92 against 9.13 to 9.46). If nothing moves, the task probably belongs in the “allowed” row and your effort is better spent elsewhere.

When is single-tasking not the answer?

• When the work is interrupt-driven by design. Support, operations and on-call roles cannot remove interruptions. The lever there is fast resumption (notes, checklists, clear hand-offs), not isolation.
• When the main task is monotonous. A low-demand side activity may keep you engaged, as in the doodling study.
• When the tasks are simple and well practised. The laboratory cost shrank to a few milliseconds after about two thousand trials per task.
• When moderate switching keeps momentum. Stalling on one hard problem can cost more than a planned change of task; the inverted-U result is a reminder that the optimum is not always zero.

Single-tasking is a tool for the pairings that punish divided attention. Applied to everything, it becomes its own kind of waste. Related reading: decision fatigue and day structure and the dopamine myth.

FAQ

Is multitasking a myth?
Partly. Two demanding decisions made at the same moment queue at a bottleneck, and only 2.5 percent of 200 students combined simulated driving with a hard memory task without a drop. Simple, practised tasks can be combined at almost no cost.

How much time does task switching cost?
In the classic laboratory study the mean cost was 614 to 1,096 milliseconds per switch. The 40 percent figure is not in that paper. The larger real-world cost is the delay before interrupted work is resumed, which averaged 25 minutes 26 seconds in a field study of 24 workers.

Does multitasking lower your IQ?
There is no published study showing that. The claim comes from an unpublished 2005 test of eight employees, and the researcher himself has said it was misrepresented.

Does media multitasking damage attention?
The evidence does not show damage. Across 46 studies the association with cognitive control was small (z = .138) and, on objective tests, no longer significant after bias correction. In one sample of 1,511 people, heavier media multitaskers under 30 performed better on a dual-task test.

Is it true that it takes 23 minutes to refocus?
The number comes from an interview, not a paper, and it describes how long interrupted work waited before being resumed, with other tasks in between. It is not a measurement of how long concentration takes to return.

Should I stop multitasking completely?
No. Single-task whenever someone else sets the pace or an error cannot be undone. Alternate deliberately when you control the pace. Combine freely when one of the activities is automatic.

Sources

1. Rubinstein JS, Meyer DE, Evans JE. Executive control of cognitive processes in task switching. Journal of Experimental Psychology: Human Perception and Performance. 2001;27(4):763-797.
2. American Psychological Association. Multitasking: Switching costs. Research summary page, 20 March 2006.
3. Dux PE, Ivanoff J, Asplund CL, Marois R. Isolation of a central bottleneck of information processing with time-resolved fMRI. Neuron. 2006;52(6):1109-1120.
4. Schumacher EH, Seymour TL, Glass JM, Fencsik DE, Lauber EJ, Kieras DE, Meyer DE. Virtually perfect time sharing in dual-task performance: Uncorking the central cognitive bottleneck. Psychological Science. 2001;12(2):101-108.
5. Watson JM, Strayer DL. Supertaskers: Profiles in extraordinary multitasking ability. Psychonomic Bulletin and Review. 2010;17(4):479-485.
6. Matthews N, Mattingley JB, Dux PE. Media-multitasking and cognitive control across the lifespan. Scientific Reports. 2022;12:4349.
7. Sana F, Weston T, Cepeda NJ. Laptop multitasking hinders classroom learning for both users and nearby peers. Computers and Education. 2013;62:24-31.
8. Clinton-Lisell V. Stop multitasking and just read: Meta-analyses of multitasking’s effects on reading performance and reading time. Journal of Research in Reading. 2021;44(4):787-816. Abstract.
9. Strayer DL, Drews FA, Crouch DJ. A comparison of the cell phone driver and the drunk driver. Human Factors. 2006;48(2):381-391. Abstract, with study details from the University of Utah release of 29 June 2006.
10. Ophir E, Nass C, Wagner AD. Cognitive control in media multitaskers. PNAS. 2009;106(37):15583-15587.
11. Wiradhany W, Nieuwenstein MR. Cognitive control in media multitaskers: Two replication studies and a meta-analysis. Attention, Perception, and Psychophysics. 2017;79(8):2620-2641.
12. Parry DA, le Roux DB. “Cognitive control in media multitaskers” ten years on: A meta-analysis. Cyberpsychology: Journal of Psychosocial Research on Cyberspace. 2021;15(2), article 7.
13. Chen H, Peng L, Peng J, Liu C, Yin L, Zhang Y, Cheng Y, Shi Z. The relationship between media multitasking and attention: A three-level meta-analysis. Current Psychology. 2025;44(7):6326-6347. Abstract.
14. Mark G, Gudith D, Klocke U. The cost of interrupted work: More speed and stress. Proceedings of CHI 2008:107-110.
15. Adler RF, Benbunan-Fich R. Juggling on a high wire: Multitasking effects on performance. International Journal of Human-Computer Studies. 2012;70(2):156-168.
16. Mark G, Gonzalez VM, Harris J. No task left behind? Examining the nature of fragmented work. Proceedings of CHI 2005:321-330.
17. Mark G, Iqbal ST, Czerwinski M, Johns P, Sano A. Neurotics can’t focus: An in situ study of online multitasking in the workplace. Proceedings of CHI 2016:1739-1744.
18. Microsoft. Breaking down the infinite workday. Work Trend Index Special Report, 17 June 2025; and Work Trend Index Annual Report, 5 May 2026.
19. Gallup Management Journal. Too many interruptions at work? Interview with Gloria Mark, 8 June 2006.
20. Wilson G. The “Infomania” study. Author’s explanatory note, 16 January 2010.
21. Andrade J. What does doodling do? Applied Cognitive Psychology. 2010;24(1):100-106. Abstract.

The audit table, the switch arithmetic, the Task-Pairing Matrix and the two-week protocol are CEOtudent analyses built on the sources above. All other figures are reported as printed in those sources.

---

This content was compiled with the support of AI following in-depth research, then written and prepared for publication by the CEOtudent editorial team.

This post is also available in: