You are on page 1of 19

De La Salle University – Manila

Gokongwei College of Engineering

Electronics and Communications Engineering Department

ENGDATA
Engineering Data Analysis

Project Written Report

Caffeine Consumption and its Effects to the Engineering Students of De La Salle University

by

Alonzo, Annika Dominique

Cadeliña, Amgad Isaiah

Delgado, Mark Denzel

Medrano, Mark Brendon

Torre, Joshua Emmanuel

ENGDATA – EQ1
I. Statement of the Problem

Over the years, studies have shown that there is an upward trend in the caffeine
consumption, especially coffee, of Filipinos which is 3.4 million of 60 kilogram bags of
coffee alone — not including tea, soda, and other caffeinated drinks — in the year 2019
(Sanchez, 2020). During the pandemic, there are student reports saying that academic
workloads have significantly increased as compared to face-to-face classes, thus,
resulting in the exacerbation of stress levels of the students.

With this, the researchers address the general problem of this study that is:
1. Is there an increase in the caffeine consumption of the engineering
students of De La Salle University before and during the pandemic?
Specifically, it seeks answers to the questions:
1. What is the effect of the caffeine consumption of the students with regard
to their overall productivity?
2. What is the effect of the caffeine consumption of the students with regard
to their academic performance?
3. What is the correlation between the caffeine consumption and length of
sleep of the students?
4. What is the correlation between the caffeine consumption and the quality
of sleep of the students?

II. Data Collection

The researchers utilized Google Forms to create a survey that was sent out to a
random sample of DLSU engineering students, wherein the respondents were limited to
the choices already laid out by the researchers. In the survey, there are two questions that
relate to the personal information of the student (ID number and course program) and
nine questions that relate to the main topic of this research. Out of the nine questions,
seven use the Likert scale for the respondents to choose from, the other two questions are
more of a multiple choice type question.

Onto the results of the survey, as indicated in the survey, the emails of the
respondents will be protected as part of the Data Privacy Act of 2012. However, the ID
numbers and course programs of the respondents will be used to show the demographics
of the survey respondents, a total of 50 respondents were recorded.
Graph 1. ID Numbers of the Respondents

Graph 2. Course Programs of the Respondents

As seen in Graph 1, the majority of the respondents are ID 119 students with 43
respondents out of 50 (86%), it is followed by ID 118 students with only 6 respondents
(12%), and lastly, there is only one respondent (2%) that comes from ID 120. In Graph 2,
it can be seen that the majority of the respondents come from the Computer Engineering
(CpE) course with 30 of the 50 respondents (60%). There are then 11 respondents (22%)
from the Electronics and Communications Engineering (ECE) course, followed by 4
respondents (8%) from the Industrial Engineering (IE) course, 3 respondents (6%) from
the Chemical Engineering (ChE) course and 1 respondent (2%) each from the Mechanical
Engineering (ME) and Civil Engineering (CE) course.

Graph 3. Caffeinated Drinks Consumed by the Respondents

The respondents were asked about the caffeinated drinks they consume, in this
question they are able to pick more than one drink. Graph 3 shows the responses, 38 of
the 50 respondents (76%) answered that they drink coffee, while 27 respondents (54%)
answered that they drink soda, this is followed by 17 respondents (34%) answered that
they drink tea. Then only 5 respondents (10%) answered that they drink energy drinks
and one respondent (2%) answered that they do not consume caffeinated drinks.
Graph 4. Caffeinated Drinks Consumed Before the Pandemic

Graph 5. Caffeinated Drinks Consumed During the Pandemic

The respondents were then asked about how often they consumed caffeinated
drinks before and during the pandemic, this is the highlight of the research, the
researchers used a Likert scale when asking these questions to the respondents. On the
Likert scale, the range is from 1 to 5, wherein 1 is the lowest (not at all) and 5 is the
highest (always). Graph 4 shows the responses of the respondents’ consumption of
caffeinated drinks before the pandemic. There were 19 respondents that answered 4 on
the Likert scale which would mean that they often consumed caffeinated drinks, this is
followed by 13 respondents that answered 3, 9 respondents that answered 5, 7
respondents that answered 2, and 2 respondents that answered 1 on the Likert scale. With
this, it can be said that the majority of the respondents often consumed caffeinated drinks
but not always. Graph 5 shows the responses of the respondents’ consumption of
caffeinated drinks now, during the pandemic. There were 16 respondents that answered 5
on the Likert scale which would mean that they always consume caffeinated drinks now,
this is followed by 11 respondents that answered 4 and 3, 9 respondents that answered 2,
and 3 respondents that answered 1 on the Likert scale. It can be assumed that the majority
of the respondents always consume caffeinated drinks now more than before the
pandemic.

Graph 6. Comparison of Consumption Before and During the Pandemic

The respondents are then asked if they agree that they consumed more caffeinated
drinks now than before the pandemic, a Likert scale was used in this question with a
range of 1 to 5, wherein 1 is the lowest (strongly disagree) and 5 is the highest (strongly
agree). Graph 6 shows that majority of the respondents, 16 respondents, answered 1 in
the Likert scale which would mean that they do not consume more caffeinated drinks
now than before the pandemic, this is followed by 13 respondents that answered 5, 10
respondents that answered 2, 6 respondents that answered 4, and 5 respondents that
answered 3 on the Likert scale.

Graph 7. Individual Productivity of the Respondents

The respondents are then asked about their individual productivity, a Likert scale
was used with a range of 1 to 5, wherein 1 is the lowest (not productive) and 5 is highest
(highly productive). Graph 7 shows that half of the respondents, 25 respondents,
answered 3 on the Likert scale, which would mean that their productivity is just average,
this is followed by 13 respondents that answered 4, 7 respondents that answered 2, 4
respondents that answered 5, and 1 respondent that answered 1 on the Likert scale.
Graph 8. Overall Academic Performance of the Respondents

The respondents are then asked about their overall academic performance, a
Likert scale was used with a range of 1 to 5, wherein 1 is the lowest (needs improvement)
and 5 is the highest (outstanding). Graph 8 shows that 19 respondents answered 3 on the
Likert scale which would mean that they have an average academic performance, this is
followed by 17 respondents that answered 4, 10 respondents that answered 2, 3
respondents that answered 5, and 1 that answered 1 on the Likert scale.
Graph 9. Average Hours of Sleep of the Respondents

The respondents were then asked about their average hours of sleep, here they
chose an answer among the choices laid out by the researchers, Graph 9 shows that the
majority of respondents, 13 respondents, answered that they have an average of 7 hours
of sleep per night. This is followed by 9 respondents each that answered they have an
average of 5 and 6 hours of sleep, then 7 respondents each that answered they have an
average of 4 and 8 hours of sleep, 3 respondents answered they have an average of 3
hours of sleep, and lastly, 1 respondent each answered that they have less 3 hours and
more than 10 hours of sleep.
Graph 10. Sleep Quality of the Respondents

Then the respondents were asked about their sleep quality, a Likert scale was used
with a range of 1 to 5, wherein 1 is the lowest (poorly: disrupted sleep) and 5 is the
highest (restful: sound sleep). Graph 10 shows that majority of the respondents, 17
respondents, answered 4 on the Likert scale which would mean they have a good quality
of sleep, this is followed by 16 respondents that answered 3, 9 respondents that answered
2, 7 respondents that answered 5, and 1 respondent that answered 1 on the Likert scale.
Graph 11. Motivation of the Respondents When Drinking Caffeinated Drinks

The last question on the survey or more so, a statement, asked the respondents if
they felt more motivated to study and do academics if they consume a caffeinated drink, a
Likert scale was used with a range of 1 to 5, wherein 1 is the lowest (strongly disagree)
and 5 is the highest (strongly agree). Graph 11 shows that the majority of respondents, 15
respondents, answered 3 on Likert scale which would mean that they are neutral on this
question, this is followed by 11 respondents that answered 5, 10 respondents that
answered 2, 9 respondents that answered 4, and 5 respondents that answered 1 on the
Likert scale.

III. Analysis

To properly treat the data that the researchers collected, they commonly used
hypothesis testing in order to answer the research questions stated. Through all the testing
that has been conducted, the researchers set the confidence interval to 99% or a
significance level (α) equal to 0.01, this value is set to a very low probability in order to
lessen the possibility of a Type 1 error to occur.

In order to answer the general problem of this study, the researchers used two
different types of hypothesis testing. The first one is the paired t-test, this type of
hypothesis testing is appropriate for the research question since the researchers are
concerned about the difference of caffeine consumption before and during the pandemic.
For the second testing, the researchers also used a two-sample Z-test in order to compare
the means of the two samples. In the testing process, the first sample is set to be the
caffeine consumption during the pandemic. On the other hand, the second sample is set to
be the caffeine consumption before the pandemic.

For the specific research questions, the researchers used correlation. Correlation is
an appropriate statistical testing in order to show the effects of caffeine to various
variables like the following: productivity, academic performance, length and quality of
sleep of students. The researchers also tested the population correlation coefficient (ρ)
using the computed sample correlation coefficient (r). This testing is important in order to
prove that the correlation in the sample data is strong enough to represent the correlation
in the population. In the testing process the researcher also set the confidence interval to
99% in order to lessen the probability of Type 1 error to occur.

IV. Results

Comparison of Caffeine Consumption Before and During the Pandemic

For the first sample, caffeine consumption during the pandemic, a mean sample
(x̅1) of 3.56 and a sample standard deviation (s1) of 1.2803, is computed . On the other
sample, caffeine consumption before the pandemic, a mean sample (x̅2) of 0.04 and a
sample standard deviation (s2) of 1.0736 is gathered.
Using the information gathered above, a two-sample Z-test is used to show the
difference between the two sample means. To do this a significance level (α) of 0.01 is
considered. The testing procedure is shown below:

α=0.01
Ho: μ1-μ2=0 → No significant difference between two population means
Ha: μ1-μ2>0 → First population mean is greater
z = 0.169 → Computed test statistic
z > 2.326 → Computed critical region using 𝑧αwith df = infinity
Fail to Reject Ho → The test statistic is not included in the critical region

In the testing procedure above, the result leads to failure of rejecting the null
hypothesis. Meaning the data gathered does not provide enough evidence to prove that
the population mean of the first sample is greater than the second sample.

Other than the two-sample Z-test, a paired T-test is also conducted to find out if
there is a significant difference between the two samples. For this testing, each sample
from the second sample is subtracted from the first sample. Since the difference between
the two samples are considered, a sample mean difference (μd) of 3.52 and a sample
standard deviation difference (sd) of 1.0736 are also computed. Since the first sample is
the caffeine consumption during the pandemic, the researchers expect that there is a
positive difference between them (Ha). Furthermore, a significance level of (α) of 0.01 is
also considered for uniformity. The testing procedure is shown below:

α=0.01
Ho: μd=0 → Sample mean difference is 0
Ha: μd>0 → Sample mean difference is positive
t = 0.2303 → Computed test statistic
t > 2.326 → Critical region using 𝑡α with df = infinity
Fail to Reject Ho → The test statistic is not included in the critical region

In the paired t-test conducted above, the decision led to failure of rejecting the
null hypothesis. This means that the population mean difference is approximately equal
or close to 0, meaning there is no significant difference between the two samples. The
two hypothesis tests conducted for the main research question both failed to reject the
null hypothesis. In other words, the caffeine consumption before and during the pandemic
of students has no difference.

Effects of Caffeine Consumption to Individual Productivity


Graph 12. Scatter plot of caffeine consumption vs individual productivity

In this part, the independent variable (x) is equated to the caffeine consumption of
students during the pandemic, on the other hand the dependent variable (y) is equated to
the individual productivity of the students as seen in graph 12. To show, or at least find
out the effects of caffeine to productivity, correlation is used and a sample correlation
coefficient (r) of 0.1883 is computed. With this value alone, it can be concluded that there
exists a positive correlation between the two variables, however it is too weak. In order to
furtherly prove that this data can represent the correlation in the population, a hypothesis
testing for population correlation coefficient (ρ) is conducted below:

α=0.01 → Significance level


Ho: ρ=0 → Indicates that there is no significant correlation between x and y
Ha: ρ>0 → Indicates that there exists a positive correlation between x and y
t = 1.328 → Computed test statistic
t > 2.326 → Critical region using 𝑡 with df = infinity
α
Fail to Reject Ho → The test statistic is not included in the rejection region

The hypothesis testing conducted shows that there is not enough evidence to
prove that there is a positive population correlation between caffeine consumption and
individual productivity of the student. However, considering that the value of sample
correlation coefficient is still positive, there is still a positive correlation within the
sample, but since this is too low and weak, the model is not enough to represent the
population as agreed by the hypothesis testing.

Effects of Caffeine Consumption to Academic Performance

Graph 13. Scattered plot of caffeine consumption vs self-rated academic performance

Graph 13 depicts how consumption of caffeine affects its self-related academic


performance. It can be seen on the graph that the student’s intake of caffeine lies on the
x-axis, while self-rated academic performance lies on the y-axis. To substantiate that there is
a relationship happening between student’s caffeine consumption and itself-rated academic
performance, a correlation is used and has resulted in a sample correlation coefficient ( r )
which is equivalent to 0.1198. Upon getting the result it can be determined that there is a
weak positive correlation between the two variables for the reason that ( r ) is approximately
close to zero ( 0 ) but not exactly zero ( 0 ).

To give an additional proof if the sample correlation is sufficient, a hypothesis testing was
done with a significance level (α) of 0.01.

α=0.01
Ho: ρ=0 → Indicates that there is no relationship between the two variables
Ha: ρ>0 → Indicates that that there is a positive relationship between the two
t = 0.8361 → Computed test statistic
t < 2.326 → Rejection region using 𝑡α with df = infinity
Fail to Reject Ho → The test statistic is not included in the rejection region

The result of the hypothesis testing is that it fails to reject the null hypothesis since
the t-statistic is not inside the critical region. This denotes that there is insufficient data in
proving that there is a positive correlation, hence this means that caffeine consumption and
academic performance has no correlation.

Correlation Between Caffeine Consumption and Length of Sleep

Graph 14. Scatter plot showing the correlation between average sleep and caffeine consumption
of students

As shown in Graph 14, the caffeine consumption of students during the pandemic
lies on the x-axis of the graph, and the average length of sleep in hours lies on the y-axis.
To prove that caffeine consumption affects the length of sleep of students, correlation is
used and resulted in a sample correlation coefficient (r) of -0.1475. With the value of r
alone, the sample shows that there is a negative correlation between the two variables.
However, the value is very close to 0 making it very weak. To give an additional proof if
the sample correlation is enough to population, a hypothesis testing for population
correlation coefficient (ρ) is also conducted with a significance level (α) of 0.01.

α=0.01
Ho: ρ=0 → Indicates that there is no significant correlation between x and y
Ha: ρ<0 → Indicates that there is a negative correlation between x and y
t = -1.0330 → Computed test statistic
t < -2.326 → Rejection region using -𝑡 with df = infinity
α
Fail to Reject Ho → The test statistic is not included in the rejection region

The hypothesis testing about the population correlation coefficient, failed to reject the
null hypothesis. This means that the sample is not enough to represent the population to show
a negative correlation between the length of sleep and caffeine consumption of students. In
the sample alone with a very low negative value for correlation coefficient, can be a sign that
the correlation is not significant as supported by the null hypothesis.

Correlation Between Caffeine Consumption and Quality of Sleep

Graph 15. Scatter plot that shows the correlation between caffeine consumption and
sleep quality of students
As shown in graph 15, caffeine consumption is treated as the independent variable
(x) and sleep quality is treated as the dependent variable (y). To show that these two
variables have a correlation, a sample correlation coefficient (r) of -0.0032 is computed.
This value shows that there is a very low negative correlation between the two variables.
Since this claim is not yet enough for a concrete conclusion, a hypothesis testing for
population correlation coefficient (ρ) is conducted using a significance level (α) of 0.01.

α=0.01
Ho: ρ=0 → Indicates that there is no significant correlation between x and y
Ha: ρ<0 → Indicates that there is a negative correlation between x and y
t = -0.0219 → Computed test statistic
t < -2.326 → Rejection region using -𝑡 with df = infinity
α
Fail to Reject Ho → The test statistic is not included in the rejection region

As explained earlier, the sample correlation coefficient is negative and at the same
time very close to 0 making it very weak. This claim is also supported by the accepted
null hypothesis which states that there is no significant correlation between the caffeine
consumption and sleep quality of students. Furthermore, it also says that the sample’s
correlation is not enough to represent the whole population since it is very low.

V. Conclusion

It is always observed that college students have grown caffeine-dependent in


order to perform at their best in numerous aspects of their life, juggling online classes and
other daily responsibilities. Caffeine is the world's most widely used drug. Caffeine, like
other stimulants, has a significant impact on brain chemistry, making it both effective and
pleasurable. This study was conducted to observe the effects of caffeine consumption on
the engineering students of De La Salle University - Manila

In conclusion, the data gathered by the researchers from the respondents have led
them to deduce that caffeine consumption has not affected the latter on any aspect that the
researchers have prompted them to rate in a survey. The survey formulated by the
researchers and made in Google forms were efficient in gathering these data from
randomly sampled students. The demographics of the respondents is as follows: Majority
of the respondents’ ID Number were from 119 and 60% took the course of Computer
Engineering. The participants were then asked how often they drank caffeinated
beverages prior to and during the pandemic. It's safe to conclude that the majority of
respondents now consume more caffeinated beverages than they did before the outbreak
of COVID-19. Caffeine use before and during the pandemic was compared. As a
conclusion, the null hypothesis cannot be rejected. From the data found, It is evident that
the two samples did not have a significant difference among them.For productivity, The
data treatment leads the researchers to find that there is also no difference between
caffeine and a student’s individual productivity. This is seen as well in the variables such
as Academic Performance, Length of Sleep, Quality of Sleep, and Motivation. The data
treatment all led to them that there is no correlation between the variables and caffeine
consumption

VI. Discussion

Based on the conclusions discussed, the implications of the results for the sample
population for the general problem is that the pandemic has no significant effect on the
caffeine consumption of the students. Furthermore, the implications for the specific
questions state that: (a) caffeine consumption has no significant effect to the productivity
of the students which could mean that other external factors are contributing to the
productivity level in such cases that there is a correlation between caffeine consumption
and student’s productivity, (b) caffeine consumption has no significant effect to the
academic performance of the students which could also mean that other factors may
influence the result in such cases that there is a correlation between caffeine consumption
and student’s academic performance, (c) caffeine consumption has no significant effect to
the length of sleep of the students, and (d) caffeine consumption has no significant effect
to the sleep quality of the students.

The study was conducted using random sampling as the sampling method,
wherein engineering students of De La Salle University is the population. For future
research, the researchers recommend that stratified sampling method is to be used to
equally account for different batches of the college — which means proportional number
of students from ID 117, 118, 119, and 120, for instance. This is to ensure that each
subgroup is equally represented in the sample for a more precise conclusion.

You might also like