MATH IA

Exploring the Effect of Screen Time on

Sleep Duration
1. Introduction
In today’s time every person uses a screen on a daily basis . Phones, TV, computer, and IPads
are used for studying,playing games, reading, scrolling reels or youtube shorts . At the same
time, many people struggle to get good quality sleep. I have personally noticed when I use my
phone for long hours and watch endless reels with no track of time —I tend to sleep very poorly
and feel more tired the next morning with dark circles . This made me think whether there is a
mathematical relationship between screen time and sleep duration.

The purpose of this investigation and IA is to examine how screen time influences a person's
sleep quality. To do this, I collected my own data and applied a range of statistical methods:
mean, median, standard deviation, histograms, boxplots, correlation coefficient, linear
regression, and residual analysis. These tools allow me to explore both the strength and the
nature of the connection between the two variables.
2. Rationale
I am interested in this topic because I often feel very exhausted during the day, and I thought
that my phone and ipad usage and screen habits are partly responsible for this . My friends and
brother also experience the same issue with sleep . Since sleep affects concentration, mood,
productivity, IQ and mental health, I wanted to see whether the connection I observe in myself
actually shows up in real data.

3. Data Collection
I have collected primary data ( directly first hand information) from 12 people around me
between the ages of 15 to 25. Each person recorded the number of hours they spent on
screens per day and the number of hours they slept each night from their phone , for seven
continuous days. I calculated the weekly average (mean) for each person, which produced 12
pairs of values: average screen time and average sleep duration.
Dataset
Participant Screen Time Sleep Duration (hrs/day)
(hrs/day)

1 3.1 8.2

2 4.5 7.4

3 6.2 6.8

4 5.8 7.0

5 2.9 8.1

6 7.3 6.4

7 6.9 6.5

8 4.1 7.6

9 8.0 5.9

10 3.8 7.9

11 5.2 7.2

12 7.7 6.1

Moreover , the sample size is small and includes 12 members. The data is authentic as it is
primarily collected by me and directly reflects the habits of real people in the same age group.
 4. Mathematical Analysis

4.1 Descriptive Statistics

Mean

The mean gives the average value for each person.

x̄ = (Σ xi) / n

For this dataset:

●​ Mean screen time:​

xˉ=5.54 hours/day
●​ Mean sleep duration:​
yˉ=7.05 hours/day

Median

The median is the middle value when the data is arranged in order either ascending or
descending ​
For even number formula that i have used is this

Median = (x(n/2) + x(n/2 + 1)) / 2

Median screen time of the people : 5.5 hours

Median sleep duration of the people : 7.1 hours

As the means and medians have less variation, the data does not seem to be strongly skewed.
4.2 Standard Deviation
Standard deviation is used to measure how much data is spread around the mean

σ = sqrt( Σ (xi - x̄)² / n )

●​ SD of screen time: 1.67 hours

●​ SD of sleep duration: 0.75 hours

There is wider variation in screen time than sleep duration.

4.3 Interquartile Range (IQR)

The IQR measures the spread of the middle 50% of the data.

IQR = Q3 - Q1

Screen Time

●​ Q1=4.1
●​ Q3=7.3

IQR=7.3−4.1=3.2 hours

Sleep Duration

●​ Q1=6.5
●​ Q3=7.9

IQR=7.9−6.5=1.4 hours

Screen time varies more than twice in relation to sleep duration.

 4.4 Graphical Analysis
Histograms

●​ Screen time values spread between 3 and 8 hours.

●​ Sleep duration clusters between 6 and 8 hours.

●​
 Scatter Plot

The scatter plot represents a clear negative downward trend:

The more the screen time the poorer and lower sleep duration.

Boxplots

Boxplots gives the data in a visually summarized manner ; it gives information about spread
and central tendency for each variable.​
They represent the median, the lower and upper quartiles, and any potential outliers.
Screen Time Boxplot

●​ Q1 = 4.1 hours
●​ Q3 = 7.3 hours
●​ IQR = 3.2 hours

This boxplot is fairly wide, which shows that screen time varies greatly among participants.
Sleep Duration Boxplot

●​ Q1 = 6.5 hours
●​ Q3 = 7.9 hours
●​ IQR = 1.4 hours

This boxplot is much narrower, representing that sleep duration is more stable and consistent
across the same age group. The comparison of the two boxplots concludes that screen time is
much more variable than sleep duration.

This supports the standard deviation and IQR results.

 4.5 Correlation Coefficient
Pearson’s correlation coefficient shows the linear relationship between strength and direction

r = [nΣxy - (Σx)(Σy)] / sqrt( [nΣx² - (Σx)²] [nΣy² - (Σy)²] )

Calculation :

r=−0.91

This indicates a strong negative correlation.

4.6 Linear Regression

A linear model can be used to show how sleep duration changes with screen time.

The general equation for linear regression model is:

y=mx+b

Slope:

m = [ nΣxy − (Σx)(Σy) ]

--------------------------------

[ nΣx² − (Σx)² ]

Intercept: b = ȳ − m x̄

Using the dataset:

m = −0.26 b = 8.49

Regression Equation:

General Predicted Sleep = 8.49 − 0.26(Screen Time)

Interpretation:
Every additional one hour of screen time is leading to a reduction of approximately 0.26
hours (about 15 minutes) of sleep per night of participants.
 4.7 Coefficient of Determination (R²)
The coefficient of determination represents how much of the variation in the dependent
variable is shown by the regression model.

R² = r²

Using the correlation value r = -0.91:

R² = (-0.91)²

R² = 0.82

This means that approximately 82% of the variation in sleep duration can be shown by
screen time.

4.8 Residuals

Residuals measure how far the actual sleep values are from

the predicted values calculated using the regression line.

The residual for each point is calculated as:

Residual = y actual - y predicted

A residual plot shows how well the linear model fits the data that we have taken. In a
good linear model, the residuals should be scattered randomly around zero without
forming any clear pattern. This indicates that a linear relationship is appropriate.
 5. Discussion
The results from all statistical tests support the idea that screen time strongly influences sleep
duration. The correlation coefficient represents a very high negative relationship, and the
regression equation demonstrates how sleep duration reduces as screen time is more. The high
R² value indicates that the model represents most of the variation in sleep of a person.

People who use screens more often, especially in the evening, may experience reduced
melatonin production because of blue-light exposure, which delays sleepiness . This biological
explanation perfectly aligns with the mathematical calculations.
 6. Evaluation
Although the results are strong, the investigation has several limitations:

●​ The sample size was small with only 12 people.

●​ Data was self-reported as I have collected it , which may include rounding or estimation
errors in calculation.
●​ Only the average value is taken, not the every day fluctuations.
●​ People close in age 15-25 are taken, so the results may not apply to older or children.
●​ Other factors such as stress, diet , physical activity, medical history should be taken into
account

Possible Improvements

●​ Include more than 12 people from wider age ranges, not just young people.
●​ Extend the data collection period.
●​ Include more variables like screen type, bedtime, or brightness settings of the screen.

7. Conclusion
This investigation found clear evidence that increased screen time is associated with reduced
sleep duration. The very high negative correlation of approx 0.91 and the high R² value
represents that as screen time is more, sleep quality decreases in a consistent way.

According to the regression model, each one more hour of screen time decreases sleep by
approximately 15 minutes. This explains that decreasing time screen use can clearly improve
sleep quality and duration.