| Term | Estimate | Std_Error | Statistic | P_Value | CI_Lower | CI_Upper |
|---|---|---|---|---|---|---|
| (Intercept) | -2.869 | 0.375 | -7.648 | 0.000 | -3.611 | -2.139 |
| Daily_Usage_Hours | 1.111 | 0.048 | 23.081 | 0.000 | 1.019 | 1.208 |
| Sleep_Hours | -0.580 | 0.043 | -13.356 | 0.000 | -0.666 | -0.496 |
| Time_on_Social_Media | 1.235 | 0.072 | 17.110 | 0.000 | 1.096 | 1.379 |
| Anxiety_Level | 0.003 | 0.020 | 0.159 | 0.874 | -0.037 | 0.043 |
| Depression_Level | 0.046 | 0.020 | 2.268 | 0.023 | 0.006 | 0.086 |
Model
Logistic Regression Model
This section explains the logistic regression model used to predict teen smartphone addiction based on behavioral and psychological factors.
General Logistic Regression Equation
\[ \log\left( \frac{P(\text{Addicted} = 1)}{1 - P(\text{Addicted} = 1)} \right) = \beta_0 + \beta_1 \cdot \text{DailyUsage} + \beta_2 \cdot \text{SleepHours} + \beta_3 \cdot \text{SocialMediaTime} + \beta_4 \cdot \text{AnxietyLevel} + \beta_5 \cdot \text{DepressionLevel} \]
Where:
- DailyUsage: Daily smartphone usage, in hours
- SleepHours: Average hours of sleep per night
- SocialMediaTime: Hours spent on social media per day
- AnxietyLevel: Self-reported anxiety score
- DepressionLevel: Self-reported depression score
Estimated Logistic Regression Equation
The fitted model from your data yields the following equation:
\[ \log\left( \frac{P(\text{Addicted} = 1)}{1 - P(\text{Addicted} = 1)} \right) = -2.869 + 1.111 \cdot \text{DailyUsage} - 0.580 \cdot \text{SleepHours} + 1.235 \cdot \text{SocialMediaTime} + 0.003 \cdot \text{AnxietyLevel} + 0.046 \cdot \text{DepressionLevel} \]
This equation lets us estimate a teen’s probability of addiction based on their habits and psychological state.
Model Code and Output Table
Conclusion
The logistic regression table summarizes how each predictor influences the likelihood of teen smartphone addiction:
- Intercept (-2.869): Represents the baseline log-odds of being addicted when all predictors are zero.
- Daily_Usage_Hours (1.111): A strong positive coefficient indicates that each additional hour of daily phone use significantly increases the odds of addiction.
- Sleep_Hours (-0.580): The negative sign suggests that more sleep is associated with lower odds of addiction. For each extra hour of sleep, the log-odds of addiction decrease.
- Time_on_Social_Media (1.235): Like daily phone use, more time on social media is a strong predictor of addiction.
- Anxiety_Level (0.003): This coefficient is very small and not statistically significant (p = 0.874), meaning anxiety level doesn’t show a clear relationship with addiction in this model.
- Depression_Level (0.046): Though the effect is smaller, it is statistically significant (p = 0.023), suggesting that higher depression scores slightly increase the likelihood of addiction.
Key Takeaways:
- Daily usage and social media time are the most impactful predictors.
- Sleep plays a protective role against addiction.
- Depression matters, but anxiety does not appear significant in this sample.
- The p-values confirm that most predictors (except anxiety) are statistically meaningful at conventional thresholds (p < 0.05).
These insights can inform strategies for reducing phone addiction in teens by targeting usage habits, promoting better sleep, and supporting mental health—especially depression.