Student’s In Distress: How Frequent Technology Use Hinders Social Engagement

Contents

Results

We are examining the relationships between time spent on social media and screen time in general with: - loneliness - social integration - neuroticism

We will be analyzing the relationship between these variables through a simple linear regression model. On the survey we sent out to General Psychology (PSYCH 111) students, there were a number of questions that had to do with a students loneliness, social life, and neuroticism where they were asked to rate themselves on a given scale based on how often during the last month they experienced feelings of loneliness, low social integration, and negative emotions in a number of different settings (private, social, or interpersonal). We combined every students individual scores per question to create an average loneliness score, social integration score, and neuroticism score for each participant.

Loneliness

Time on Social Media

We hypothesize that social media is correlated with loneliness; the more an individual uses social media, the higher the levels of loneliness are reported.

	Estimate	Std. Error	t value	Pr(>|t|)
(Intercept)	2.444	0.09949	24.57	4.964e-61
SocialMedia Time Con_1	0.0007471	0.0004296	1.739	0.08365

Fitting linear model: Avg_Lonely ~ SocialMedia Time Con_1

Observations	Residual Std. Error	\(R^2\)	Adjusted \(R^2\)
193	0.9781	0.01559	0.01043

SocialMedia Time Con_1
0.1248

FALSE `geom_smooth()` using formula = 'y ~ x'

Screen Time

Screen time is correlated with loneliness; the higher the screen time, the more one experiences feelings of loneliness.

mylm2 <- lm(Avg_Lonely ~ `Avg_Screen Time`, data=proj)
pander(summary(mylm2))

	Estimate	Std. Error	t value	Pr(>|t|)
(Intercept)	2.442	0.1211	20.16	3.436e-49
Avg_Screen Time	0.000607	0.0004804	1.263	0.208

Fitting linear model: Avg_Lonely ~ Avg_Screen Time

Observations	Residual Std. Error	\(R^2\)	Adjusted \(R^2\)
193	0.9817	0.008288	0.003096

pander(lm.beta(mylm2))

Avg_Screen Time
0.09104

par(mfrow=c(1,3))
plot(mylm2, which=1:2)
plot(mylm2$residuals)

ggplot(proj, aes(x = `Avg_Screen Time`, y = Avg_Lonely)) +
  geom_point(color = "dodgerblue", size = 3, alpha = 0.7) +  
  geom_smooth(method = "lm", se = TRUE, color = "purple1", fill = "gray40") +  
  labs(
    title = "Relationship Between Screen Time and Loneliness",
    x = "Average Screen Time (Minutes per Day)", 
    y = "Average Loneliness Score"
  )

`geom_smooth()` using formula = 'y ~ x'

Social Integration

Time on Social Media (on poster)

Our hypothesis was that social media and social integration are negatively correlated—the more one uses social media, the lower their social integration score.

mylm3 <- lm(Avg_SocInteg ~`SocialMedia Time Con_1`, data=proj)
pander(summary(mylm3))

	Estimate	Std. Error	t value	Pr(>|t|)
(Intercept)	2.723	0.1219	22.33	3.767e-55
SocialMedia Time Con_1	0.001262	0.0005265	2.398	0.01746

Fitting linear model: Avg_SocInteg ~ SocialMedia Time Con_1

Observations	Residual Std. Error	\(R^2\)	Adjusted \(R^2\)
193	1.199	0.02922	0.02414

pander(lm.beta(mylm3))

SocialMedia Time Con_1
0.1709

The p-value of 0.017 indicates that the relationship is statistically significant. However, the standardized β of 0.171 shows a weak positive relationship, contrary to our hypothesis. This means that, although social media use slightly increases social integration, the effect is small. The R² of 0.03 suggests that only 3% of the variation in social integration is explained by social media use, implying that other factors contribute more. The intercept estimate of 2.723 indicates that with zero social media use, the predicted social integration score is about 2.723.

ggplot(proj, aes(x = `SocialMedia Time Con_1`, y = Avg_SocInteg)) +
  geom_point(color = "dodgerblue", size = 3, alpha = 0.7) +  
  geom_smooth(method = "lm", se = TRUE, color = "purple1", fill = "gray40") +  
  labs(
    title = "Relationship Between Social Media Time and Social Integration",
    x = "Social Media Time (Minutes per Day)", 
    y = "Average Social Integration Score"
  )

FALSE `geom_smooth()` using formula = 'y ~ x'

To conclude this part of our study, while social media use is statistically significant, it explains only a small portion of the variation in social integration, suggesting that other factors have a more substantial influence on social integration.

Screen Time

Screen time is correlated with social integration; the more one engages with screens, the less their average social integration score is.

mylm4 <- lm(Avg_SocInteg ~ `Avg_Screen Time`, data=proj)
pander(summary(mylm4))

	Estimate	Std. Error	t value	Pr(>|t|)
(Intercept)	2.894	0.15	19.28	1.056e-46
Avg_Screen Time	0.000174	0.0005951	0.2924	0.7703

Fitting linear model: Avg_SocInteg ~ Avg_Screen Time

Observations	Residual Std. Error	\(R^2\)	Adjusted \(R^2\)
193	1.216	0.0004473	-0.004786

pander(lm.beta(mylm4))

Avg_Screen Time
0.02115

par(mfrow=c(1,3))
plot(mylm4, which=1:2)
plot(mylm4$residuals)

mylm <- lm(Avg_SocInteg ~ `Avg_Screen Time`, data = proj)
ggplot(proj, aes(x = `Avg_Screen Time`, y = Avg_SocInteg)) +
  geom_point(color = "dodgerblue", size = 3, alpha = 0.7) +  
  geom_smooth(method = "lm", se = TRUE, color = "purple1", fill = "gray40") +  
  labs(
    title = "Relationship Between Screen Time and Social Integration",
    x = "Average Screen Time (Minutes per Day)", 
    y = "Average Social Integration Score"
  )

`geom_smooth()` using formula = 'y ~ x'

Neuroticism

Time on Social Media

We hypothesize that social media and neuroticism are correlated; the more one uses social media, the more one has characteristics that classify under neuroticism.

mylm5 <- lm(Avg_Neuro ~ `SocialMedia Time Con_1`, data=proj)
pander(summary(mylm5))

	Estimate	Std. Error	t value	Pr(>|t|)
(Intercept)	2.123	0.05209	40.75	3.749e-96
SocialMedia Time Con_1	0.0004127	0.0002249	1.835	0.0681

Fitting linear model: Avg_Neuro ~ SocialMedia Time Con_1

Observations	Residual Std. Error	\(R^2\)	Adjusted \(R^2\)
193	0.5121	0.01732	0.01217

pander(lm.beta(mylm5))

SocialMedia Time Con_1
0.1316

par(mfrow=c(1,3))
plot(mylm5, which=1:2)
plot(mylm5$residuals)

ggplot(proj, aes(x = `SocialMedia Time Con_1`, y = Avg_Neuro)) +
  geom_point(color = "dodgerblue", size = 3, alpha = 0.7) +  
  geom_smooth(method = "lm", se = TRUE, color = "purple1", fill = "gray40") +  
  labs(
    title = "Relationship Between Social Media Time and Neuroticism",
    x = "Social Media Time (Minutes per Day)", 
    y = "Average Neuroticism Score"
  )

`geom_smooth()` using formula = 'y ~ x'

Screen Time

Screen time is correlated with neuroticism; the higher number of screen time, the more they will have neurotic tendencies/personality.

mylm6 <- lm(Avg_Neuro ~`Avg_Screen Time`, data=proj)
pander(summary(mylm6))

	Estimate	Std. Error	t value	Pr(>|t|)
(Intercept)	2.095	0.06317	33.17	3.484e-81
Avg_Screen Time	0.0004649	0.0002505	1.856	0.06506

Fitting linear model: Avg_Neuro ~ Avg_Screen Time

Observations	Residual Std. Error	\(R^2\)	Adjusted \(R^2\)
193	0.512	0.01771	0.01256

pander(lm.beta(mylm6))

Avg_Screen Time
0.1331

par(mfrow=c(1,3))
plot(mylm6, which=1:2)
plot(mylm6$residuals)

ggplot(proj, aes(x = `Avg_Screen Time`, y = Avg_Neuro)) +
  geom_point(color = "dodgerblue", size = 3, alpha = 0.7) +  
  geom_smooth(method = "lm", se = TRUE, color = "purple1", fill = "gray40") +  
  labs(
    title = "Relationship Between Screen Time and Neuroticism",
    x = "Average Screen Time (Minutes per Day)", 
    y = "Average Neuroticism Score"
  )

`geom_smooth()` using formula = 'y ~ x'

Gender Specific Studies

Time on Social Media by Gender

We hypothesize that women would have higher levels of social media use based on our understanding that women are more social and prioritize social engagement, based on past research.

pander(proj %>%
  group_by(Sex) %>%
  summarise(aveTime = mean(proj$`SocialMedia Time Con_1`),
             sdTime = sd(`SocialMedia Time Con_1`),
             sampleSize = n()))

Sex	aveTime	sdTime	sampleSize
Boy	163.6	177.2	76
Girl	163.6	156.2	117

Looking at the averages between men and women, the average amount of time spent on social media is the same (mean = 163.6) for both groups! However, the standard deviations differ, with men having a standard deviation of 177.2 and women having a standard deviation of 156.2. This difference can be attributed to the larger number of women in the study compared to men. The larger sample of women provides a more stable estimate of variability, while the smaller sample size for men is more susceptible to fluctuations and extreme values, leading to a higher standard deviation.

t-Test for Time

pander(t.test(`SocialMedia Time Con_1` ~ Sex, data=proj, mu=0, alternative="two.sided", conf.level=0.95))

Welch Two Sample t-test: SocialMedia Time Con_1 by Sex (continued below)

Test statistic	df	P value	Alternative hypothesis	mean in group Boy
0.1248	145.8	0.9008	two.sided	165.5

mean in group Girl
162.4

qqPlot(`SocialMedia Time Con_1` ~ Sex, data=proj)

ggplot(proj, aes(x=factor(Sex), y=`SocialMedia Time Con_1`, fill=Sex)) +
  geom_boxplot(color="black") +
  scale_fill_manual(values = c("Boy" = "skyblue", "Girl" = "lightpink")) +
  labs(
    title="Time Spent on Social Media",
    y="Minutes on Social Media") +
  theme(axis.title.x = element_blank())

Regression by Gender

Loneliness

We hypothesize women will have higher levels of loneliness due to the fact that we believe they will engage on social media more than men.

boys_data <- proj %>% filter(Sex == "Boy")
model <- lm(Avg_Lonely ~ `SocialMedia Time Con_1`, data = boys_data)
pander(summary(model))

	Estimate	Std. Error	t value	Pr(>|t|)
(Intercept)	2.466	0.1486	16.59	1.197e-26
SocialMedia Time Con_1	0.000576	0.0006152	0.9363	0.3522

Fitting linear model: Avg_Lonely ~ SocialMedia Time Con_1

Observations	Residual Std. Error	\(R^2\)	Adjusted \(R^2\)
76	0.9438	0.01171	-0.001647

pander(lm.beta(model))

SocialMedia Time Con_1
0.1082

FALSE `geom_smooth()` using formula = 'y ~ x'

girls_data <- proj %>% filter(Sex == "Girl")
model <- lm(Avg_Lonely ~ `SocialMedia Time Con_1`, data = girls_data)
pander(summary(model))

	Estimate	Std. Error	t value	Pr(>|t|)
(Intercept)	2.425	0.1347	18.01	2.874e-35
SocialMedia Time Con_1	0.0008901	0.0005989	1.486	0.14

Fitting linear model: Avg_Lonely ~ SocialMedia Time Con_1

Observations	Residual Std. Error	\(R^2\)	Adjusted \(R^2\)
117	1.007	0.01885	0.01031

pander(lm.beta(model))

SocialMedia Time Con_1
0.1373

FALSE `geom_smooth()` using formula = 'y ~ x'

Social Integration

We hypothesize that men who use technology more often typically report a lower rate of social interactions than compared to women.

boys_data1 <- proj %>% filter(Sex == "Boy")
modelb1 <- lm(Avg_SocInteg ~ `SocialMedia Time Con_1`, data = boys_data1)
pander(summary(modelb1))

	Estimate	Std. Error	t value	Pr(>|t|)
(Intercept)	2.906	0.2002	14.51	2.437e-23
SocialMedia Time Con_1	0.000694	0.0008289	0.8373	0.4051

Fitting linear model: Avg_SocInteg ~ SocialMedia Time Con_1

Observations	Residual Std. Error	\(R^2\)	Adjusted \(R^2\)
76	1.272	0.009385	-0.004002

pander(lm.beta(modelb1))

SocialMedia Time Con_1
0.09687

FALSE `geom_smooth()` using formula = 'y ~ x'

girls_data1 <- proj %>% filter(Sex == "Girl")
modelg1 <- lm(Avg_SocInteg ~ `SocialMedia Time Con_1`, data = girls_data1)
pander(summary(modelg1))

	Estimate	Std. Error	t value	Pr(>|t|)
(Intercept)	2.589	0.1539	16.83	8.464e-33
SocialMedia Time Con_1	0.001728	0.0006843	2.525	0.01295

Fitting linear model: Avg_SocInteg ~ SocialMedia Time Con_1

Observations	Residual Std. Error	\(R^2\)	Adjusted \(R^2\)
117	1.151	0.05251	0.04427

pander(lm.beta(modelg1))

SocialMedia Time Con_1
0.2292

FALSE `geom_smooth()` using formula = 'y ~ x'

Neuroticism (Gender)

We hypothesize that women will have a higher average neuroticism score than men. We will analyze this through an independent samples t-test!

Welch Two Sample t-test: Avg_Neuro by Sex (continued below)

Test statistic	df	P value	Alternative hypothesis	mean in group Boy
-1.488	153.6	0.9306	greater	2.121

mean in group Girl
2.235

ggplot(proj, aes(x=factor(Sex), y=Avg_Neuro, fill=Sex)) +
  geom_boxplot(color="black") +
  scale_fill_manual(values = c("Boy" = "skyblue", "Girl" = "lightpink")) +
  labs(
    title="Neuroticism: Males or Females",
    y="Average Neuroticism Score")+
 theme(axis.title.x = element_blank())