Short Videos Viewing Behaviours Have Negative Impact on

Undergraduate Students’ Visual Short-Term Memory (VSTM)

Performance

Ruoqi Huang

Department of Psychology, University of Toronto, Toronto, Ontario, Canada

Keywords: Visual Short-Term Memory, Short Video, The Atkinson-Shiffrin’s Modal Model Of Memory, Reading, Pattern

Recognition.

Abstract: Visual Short-Term Memory (VSTM) is the ability to retain visual input for a short period of time before being

processed by the later memory systems. VSTM is thought to be related to reading ability, which is a major

activity of undergraduate student’s daily academic activities. The pandemic nurtures the popularity of short

videos, particularly in the young population, where they included undergraduate students. While users spend

an average 24.5 minutes on short videos every day, the negative impact of short videos is still not clear from

a memory perspective. This study plans to investigate how viewing short videos influences VSTM

performance by designing a VSTM test using monochrome pattern recognition. Result shows significant

difference between experimental group and control group in answering rate (p = 0.042), accuracy (p = 0.027),

and reaction time (p < 0.001). Result further suggests that undergraduate participants that attempt the VSTM

test after a 5-minute short video watching session showed lower answer rate, lower accuracy, and longer

reaction time, suggesting a worse performance comparing to those did not watch short videos before the

attempt. An explanation based on the relationship between selective attention and VSTM was proposed. This

study is aimed to provide reference for undergraduate student on how the popular short video viewing

behaviour can have negative effect on their VSTM, and possibly on their academic activities.

1 INTRODUCTION

The Atkinson-Shiffrin’s Modal Model of Memory

suggests that external sensory input is first stored in

Sensory Memory, with attended memory stored as

VisualShort-Term Memory (VSTM)(Atkinson,

Shiffrin 1968). One type of Short-Term Memory that

temporarily stores visual information is Visual Short

Term Memory (VSTM) (Luck 2007). Unlike the

sensory storage, which has a high capacity, related to

the spatial position, maskable, and brief, VSTM is

not tied to spatial position, protected against

masking, and degenerates after a few seconds

(Newhagen, Reeves 1991).

Since VSTM receives visual input from sensory

memory and stores the visual information for further

cognitive processing, it is shown to facilitate

character recognition, driving, and recognizing

passing by objects, which are all vital to one’s normal

function in daily life (Curby, Gauthier 2007,

Koyama, Stein, Stoodley, Hansen 2011, Reimer,

Mehler, Wang, Coughlin 2012). More specifically,

VSTM is involved in people ’s reading process, that

allows one to receive sensory input from the text, and

temporarily store for further processing (Koyama,

Stein, Stoodley, Hansen 2011) Interference with

VSTM can affect one’s performance in daily tasks

involving visual information processing, thus

impaired one’s normal functioning.

Prior research on VSTM has shown that

memories are prone to interference from a second

irrelevant task (Rademaker, Bloem, De Weerd, Sack

2015). The factor that distracts VSTM is known as a

distractor (Rademaker, Bloem, De Weerd, Sack

2015). While older research suggests that VSTM is

only vulnerable to distractors containing conflicting

information along stimulus of the relevant dimension

(Reimer, Mehler, Wang, Coughlin 2012), more

recent research has proven that the more a distractor

differs from the memory stimuli, the more

considerable interference the distractor will cause

(Rademaker, Bloem, De Weerd, Sack 2015). Hence,

performing a highly irrelevant task before the

298

Huang, R.

Short Videos Viewing Behaviours Have Negative Impact on Undergraduate Students’ Visual Short-Term Memory (VSTM) Performance.

DOI: 10.5220/0011367900003444

In Proceedings of the 2nd Conference on Artiﬁcial Intelligence and Healthcare (CAIH 2021), pages 298-305

ISBN: 978-989-758-594-4

memory task will not only affect but strongly affect

the performance of a VSTM test. In this research,

short videos are chosen to be the highly irrelevant

distractor.

Short videos are becoming increasingly popular

as a novel means of entertainment, particularly in

Mainland China (website: http://www.bigdata-

research.cn/content/202005/1073.html 2021). With a

large population of potential customers and the

lifestyle changes due to the COVID-19 pandemic,

the market scale of short videos in Mainland China

has a like-for-like growth of 45.3% in 2020, reaching

62 billion yuan (website: http://www.bigdata-

research.cn/content/202005/1073.html 2021). The

users spend an average of 24.5 minutes per day on

these short video apps, with Douyin(The Chinese

Version of TikTok) being their most popular choice

(website: http://www.bigdata-

research.cn/content/202005/1073.html 2021). Short

videos are also popular among undergraduate

students, statistics showed that users between the age

of 20 and 29 consisted of up 22.4% of the short video

users, most of which are in the range of

undergraduate students’ ages (Statista.com website:

https://www.statista.com/statistics/1095186/tiktok-

us-users-age/ 2021).

Short videos are usually 15 seconds or 60 seconds

long, and users can scroll to the next video if they are

not interested in the current ones (Tiktok.com

website: https:// www.tiktok.com 2021). The most

easily accessible page on Douyin is called #fyp,

namely “For You Page,” where recommendation

algorithms recommend the videos according to the

users ’ viewing habits. The more the users watch, the

more customized the recommendations will be

(TikTok website: https://newsroom.tiktok.com/en-

us/how-tiktok-recommends-videos-for-you 2021).

The viewing of short videos is a consistent reward

loop, where the algorithm presents more videos, the

user likes to increase the user’s watching duration,

using positive reinforcement. Users often spend

more time than they planned watching short videos

(website: http://www.bigdata-

research.cn/content/202005/1073.html 2021).

During those times, they consistently receive visual

sensory inputs, which are considered distractors, and

may interfere with their subsequent memory

processes, including VSTM.

To our knowledge, there is currently no research

on how watching short videos can impact VSTM.

Researching how watching short videos consistently

affects VSTM can help us better understand how this

emerging way of entertainment affects our cognitive

functions. Our null hypothesis(h0) is watching short

videos would not affect VSTM.

According to the null hypothesis, we have

proposed a research hypothesis to reveal the

correlation between watching short videos and

VSTM. The research hypothesis suggests that if an

individual watches short videos for an above-average

time consistently, their performance on a VSTM test

will worsen.

We will examine the null and research hypothesis

by designing and performing a pattern recognition

test that assesses the participant’s VSTM

performance before and after the video-watching

behaviour. The pattern recognition test first shows

the participant a decorative pattern. After a delay, the

participant will be asked to recall the seen pattern by

dragging the spectrum to find the decorative pattern

they think is most like the seen pattern. The

participant will perform the test with and without

watching short videos for an above-average amount

of time, with a meditation between the sessions. If

the research hypothesis is valid, we should see a

decrease in VSTM test performance after the video-

watching task, comparing to the performance

without the video-watching task.

The stimuli are collected and processed from

open access patterns available online. The

experiment will be programmed through PsychoPy.

Data collection will be through PsychoPy. Data

analysis will be done using Microsoft Excel and R.

2 METHOD

2.1 Participants

As shown in Table 1, 10 participants (2 males, 8

females; age mean = 19.6 yrs, SD = 0.8) participated

in the study. None of the participants reported any

neurological or psychiatric disorder; all of them were

native mandarin speakers.

Short Videos Viewing Behaviours Have Negative Impact on Undergraduate Students’ Visual Short-Term Memory (VSTM) Performance

299

Table 1: Prescreening Result for All Participants.

Participant

Age Gender Years of

Education

Video Vision Hand Medical

Condition

Brain

Trauma

1-001 19 Female 13 1h+/Weibo Y R N N

1-002 20 Female 13 1h Y R N N

1-003 20 Female 13 3h Y R N N

1-004 19 Female 13 1h+/Weibo Y R N N

1-005 19 Female 13 0 Y R N N

2-001 20 Female 13 2h+/Tiktok Y R N N

2-002 20 Male 14 1h+ Y R N N

2-003 20 Female 13 2h+/Tiktok Y R N N

2-004 21 Male 14 0 Y R N N

2-005 18 Female 12 10min Y R N N

Note. Participants are grouped according to their

single blinded sorting for their experimental group,

with 1-xxx being the control group and the 2-xxx

being the experimental group. The “Video” section

refers to the participants’ daily short video viewing

habit, where both time and types of short videos are

both recorded. The “Vision” session refers to

whether the participants have corrected to normal

vision.

2.2 Experimental Procedure

At the beginning of the experiment, subjects were

informed of the experimental procedure and they

each signed a consent form. The age, gender,

handedness, vision, level of education, self-report

short video viewing time per day of participants were

recorded during pre-screening.

5 of the participants were randomly selected to

undergo 5 sets of repeated experimental blocks,

during which 5 min of Douyin/TikTok videos were

viewed followed by 10 pattern recognition VSTM

questions (i.e., a total of 25 min video-watching and

50 VSTM questions). The remaining 5 participants

(control group) completed the same experimental

procedure but did not watch any videos. The control

group was closely monitored by the experimenter to

ensure that they did not engage in any attention-

demanding activities (e.g., reading, watching videos,

replying to messages). All videos were presented

from participants ’personal mobile phones due to

constraints in experimental condition.

2.3 Short Video Distractors

Short videos are used as distractors and are either 15

s or 60s in maximum length.

To maximally imitate participants

’daily short

video experience, we asked participants to watch

videos from their most frequently used short video

apps. Or if the participant does not have a habit of

watching short videos, we asked them to view videos

from a blank account on Douyin (the most popular

short video app in Mainland China) to ensure a

relative unbiased video selection.

2.4 VSTM Test

Following each video viewing period (experimental

group) or rest period (control group), a VSTM test is

used as the secondary task procedure.

CAIH 2021 - Conference on Artiﬁcial Intelligence and Healthcare

300

Figure 1 Scheme of The VSTM Test For Both Control And Experimental Group.

As shown in Figure 1, the monochrome pattern

was shown for 0.5s. There was a fixation for 0.5s

before the correct pattern. After the patter was shown

for 0.5s, there was another fixation for 0.5s, and later

participants tried to identify the shown pattern from

4 choices presented choices. The choices were

available for 1.5s.

Both groups response were recorded, correctness

rate and verbalized response time recorded.

Since verbalized response time required manual

recording, we conducted a single-blind experiment

where the experimenter who did the recording was

kept unaware of whether a subject belongs to the

experiment group or control group in order to prevent

biased data.

A pattern recognition STM test was performed on

each participant as a secondary task procedure. The

participants were briefly shown a monochrome

pattern for 500ms before trying to identify the given

pattern in the following multiple-choice questions.

Participants were instructed to pay close attention to

the test. Participants ’reaction time and accuracy

were recorded for data analysis as supported by prior

studies (Newhagen, Reeves 1991).

3 RESULTS

3.1 Measurements

The collected data was evaluated from three aspects:

the answering rate, the correctness rate, and the

reaction time. The answering rate refers to the

response the participants provided divided by the

total trials available (n = 50). The accuracy refers to

the correct choices made divided by the total number

of trials (n = 50). The response time was recorded by

the PsychoPy data collection function and extracted

from its .csv file.

Table 2 Results For Answering Rate, Correctness Rate, and Average. Time.

Control

rou

artici

ants

1-001 1-002 1-003 1-004 1-005 mean SD SEM

Answering rate 0.94 0.94 0.98 0.92 0.98 0.952 0.024 0.012

Correctness rate 0.8 0.84 0.9 0.86 0.98 0.876 0.061 0.031

Avg. Time 2.14 2.14 2.15 2.26 2.27 2.192 0.060 0.030

Experimental group participants

2-001 2-002 2-003 2-004 mean SD SEM

Answering rate

0.72 0.96 0.9 0.8 0.86 0.848 0.083 0.041

Correctness rate

0.68 0.84 0.8 0.78 0.7 0.760 0.061 0.030

Avg. Time

2.49 2.43 2.32 2.41 2.56 2.442 0.080 0.040

Short Videos Viewing Behaviours Have Negative Impact on Undergraduate Students’ Visual Short-Term Memory (VSTM) Performance

301

Table 3: P-Values For Answering Rate, Correctness Rate, and Average Time.

P-value

Answering rate 0.042

Correctness rate 0.027

. Time 0.001

Figure 2: Comparison of Mean Answering Rate and Mean Correctness Rate.

Figure 3: Comparison of Mean Reaction Time.

0,952

0,876

0,848

0,76

0,500

0,625

0,750

0,875

1,000

mean answering rate mean correctness rate

Accuracy

Control group participants Experimental group particippants

2,192

2,442

1,875

2,000

2,125

2,250

2,375

2,500

time/s

Control group Experimental group participants

CAIH 2021 - Conference on Artiﬁcial Intelligence and Healthcare

302

3.2 Answering Rates

The answering rate is how many trials can the

participant provide an answer in a given time interval

of 3.5s. Shown in Table 2, the control group

participants who did not watch any short videos had

an average answering rate of 0.952 (SD = 0.024,

SEM = 0.012), and the experimental group who

watched 5 minutes of short videos before each

sessions had an average answering rate of 0.848 (SD

= 0.083, SEM = 0.041). Shown in Figure 2, the mean

answering rate for control group was higher than the

experimental group. As shown in Table 3, the

calculated p-value using unpaired two tailed t-test

showed that the p-value for answering rate between

the control and the experimental group is 0.042, the

alpha value in this study was set to 0.05.

3.3 Correctness Rate

The correctness rate refers the number of questions

each participants answered correctly comparing to

the questions. As shown in Table 2, the control group

had an average correctness rate of 0.876 (SD = 0.061,

SEM = 0.031) and the experimental group showed an

average correctness of 0.760 (SD = 0.061, SEM =

0.030). Shown in Figure 2, the control group has a

higher average correctness rate comparing to the

experimental group. Using unpaired two tailed t-test,

the p-value for answering rate between the control

and the experimental group is 0.027, with an alpha

value of 0.05 for this study, as shown in Table 3.

3.4 Reaction Time

As recorded by the PsychoPy program, the reaction

time shows how fast does the participants response

to the choices. Shown in Table 2, the mean reaction

time is 2.192s for control group (SD = 0.060, SEM =

0.030), and 2.442s for experimental group (SD =

0.080, SEM = 0.040). Figure 3 shows a shorter mean

reaction time for control group comparing to

experimental group. The calculated unpaired two

tailed t-test result shown in Table 3 shows that the p-

value for reaction time is less than 0.001, and this

study applies an alpha value of 0.05.

4 DISCUSSIONS

4.1 Discussion: Short Video Viewing

Behaviour Has a Negative Impact

on VSTM Performance

Despite being a rising form of entertainment, there

has been little to non-research on how short-video

viewing behaviour affects VSTM, which is important

to reading (Koyama, Stein, Stoodley, Hansen 2011).

The participants in this round of experiment are all

undergraduate students, meaning that the discussion

can be narrow down to the population of

undergraduate students. In this population of 10

participants, there are significance shown in all three

measured aspects including answering rates,

correctness, and reaction time.

For answering rates, since the alpha value is 0.05

and p = 0.042, the p-value is smaller than the alpha

value, meaning that there is a significant difference

for answering rates between participants in

experimental group and control group. The plotted

graph also shows that the experimental group, shows

lower answering rate comparing to control group.

The significance implies that participant who watch

five minutes of short videos for each session, a total

of 25 minute of short video, will show less answering

rate when responding to VSTM tests.

For correctness, the alpha value is 0.05 and the

calculated p = 0.027, the p-value is smaller than the

alpha value. The significance means that there is a

significant difference for correctness between

experimental group and control group. With the

correctness of experimental group being lower, this

study reveals that watching short videos on a daily

average time can have a negative impact on VSTM’s

accuracy.

For reaction times, the alpha value is 0.05 and the

calculated p < 0.001. The p-value is smaller than the

alpha value. This result means that if the participant

watches short videos before each of the trials, they

will react slower to presented stimulus that test their

VSTM performance. Moreover, when performing

VSTM related tasks, they tend to perform slower

than others who have not watched short videos prior

to the tasks.

Since the experimental group, which are

participants that watched short-videos prior to the

trials, showed lower answering rate, correctness, and

longer reaction time, this study can conclude that

short-video viewing has a negative impact on

undergraduate student’s VSTM performance.

Short Videos Viewing Behaviours Have Negative Impact on Undergraduate Students’ Visual Short-Term Memory (VSTM) Performance

303

4.2 Conclusions

Undergraduate students spend an average of 6.5

hours on their daily activities, which most of the time

are spent on reading (St Clair-Thompson, Graham,

Marsham 2018). Since the reading activity is heavily

emphasized in undergraduate students’ academic

life, interruption on VSTM, the reading-related

memory can imply negative outcome of reading

activities.

Prior research has shown VSTM disruption by

selective attentional disruption (Makovsik, Jiang

2007). Makovski et al. performed an experiment to

prove that if attention is shifted from a distributed

mode to a focused mode, the VSTM performance

increased (Makovsik, Jiang 2007). This previous

research showed a relationship between VSTM and

attention. When the participants are viewing short

videos, the short videos are considered sensory

inputs, and according to the Atkinson-Shiffrin

’s

Modal Model of memory, the visual input are stored

in sensory memory and later filtered by attention to

store in the VSTM. In the case where experimental

groups are viewing short videos, if a further

experiment can be performed to verified that the

participant’s attention is in a disrupted state, then it

would be possible to verify the result by Makovski et

al.

From an application perspective, this study has

provided a result showing that the popular activity of

short video viewing has an impact on VSTM, which

is an academic-activity-related form of memory. The

result can provide reference for short video

application users, especially undergraduate students,

parents, universities, and the short video application

companies.

4.3 Limitations and Future Directions

The sample size of this study is limited and there is

an unequal distribution of gender. The experiment

was designed to run online, but due to time and

device limit, and the social gathering restrictions, the

experiment was run in an alternative form and run

online using zoom and reaction time was recorded

using verbal response from participants. Also, there

is an ambiguous requirement of short videos

viewing. Some participants from the experimental

group are observed scrolling the short videos

consistently, while others finish watching each video

to the end.

As a result, more data gathering is needed in the

subsequent stages of research, and more analysis can

be done once data size is sufficient. Since short

videos is a brand-new form of media, study how the

content and structure of short videos influences

VSTM performance is also considered a next step.

Possible neuroscientific studies like EEG and fMRI

can be combined with behavioural test to reveal more

biological evidence to support the gathered data.

REFERENCES

Atkinson, R. C., & Shiffrin, R. M. (1968). Human memory:

A proposed system and its control processes. In K. W.

Spence & J. T. Spence (Eds.), Psychology of Learning

and Motivation (Vol. 2, pp. 89–195). San Diego, CA:

Elsevier.

Curby, K. M., & Gauthier, I. (2007). A visual short-term

memory advantage for faces. Psychonomic Bulletin &

Review, 14(4), 620–628.

Douyin’s User Activities Dramatically Increased During

COVID-19. (n.d.). Retrieved May 28, 2021, from

Bigdata-research.cn website: http://www.bigdata-

research.cn/content/ 202005/1073.html

Koyama, M. S., Stein, J. F., Stoodley, C. J., & Hansen, P.

C. (2011). Functional MRI evidence for the importance

of visual short-term memory in logographic reading:

Logographic reading and visual short-term memory.

The European Journal of Neuroscience, 33(3), 539–

548.

Luck, S. (2007). Visual short term memory. Scholarpedia

Journal, 2(6), 3328.

Magnussen, S., Greenlee, M. W., Asplund, R., & Dyrnes,

S. (1991). Stimulus-specific mechanisms of visual

short-term memory. Vision Research, 31(7–8), 1213–

1219.

Makovsik, T., & Jiang, Y. V. (2007). Distributing versus

focusing attention in visual short-term memory.

Psychonomic Bulletin & Review, 14(6), 1072–1078.

Newhagen, J., & Reeves, B. (1991). Effects of compelling

visual images on attention to and memory for new

stories. Paper presented to the International

Communication Association, Chicago

Phillips, W. A. (1974). On the distinction between sensory

storage and short-term visual memory. Perception &

Psychophysics, 16(2), 283–290.

Rademaker, R. L., Bloem, I. M., De Weerd, P., & Sack, A.

T. (2015). The impact of interference on short-term

memory for visual orientation. Journal of Experimental

Psychology. Human Perception and Performance,

41(6), 1650–1665.

Reimer, B., Mehler, B., Wang, Y., & Coughlin, J. F. (2012).

A field study on the impact of variations in shortterm

memory demands on drivers’ visual attention and

driving

St Clair-Thompson, H., Graham, A., & Marsham, S.

(2018). Exploring the reading practices of

undergraduate students. Education Inquiry, 9(3), 284–

298.

TikTok - make your day. (n.d.). Retrieved May 28, 2021,

from Tiktok.com website: https:// www.tiktok.com

CAIH 2021 - Conference on Artiﬁcial Intelligence and Healthcare

304

TikTok user ratio in the U.S. 2021, by age group. (n.d.).

Retrieved June 25, 2021, from Statista.com website:

https://www.statista.com/statistics/1095186/tiktok-us-

users-age/

TikTok. (2020, June 18). How TikTok recommends videos

#ForYou. Retrieved May 28, 2021, from TikTok

website: https://newsroom.tiktok.com/en-us/how-

tiktok-recommends-videos-for-you

Short Videos Viewing Behaviours Have Negative Impact on Undergraduate Students’ Visual Short-Term Memory (VSTM) Performance

305