Analyzing Political Bias Perceptions| Facebook and Other Contexts

A Multi-faceted Analysis for Analyzing Perceived Bias in Different Contexts

Explore the intricate landscape of public opinion through a comprehensive analysis that navigates perceptions of political bias across various contexts. This study delves into Facebook bias among Democrats and Republicans, examines gender differences in bias perception, and offers comparative insights into experiences with traumatic events. Uncover the subtle nuances that shape our understanding of political neutrality and delve into the broader concept of the hostile mediator phenomenon. This exploration aims to shed light on the diverse factors influencing public sentiment in today's dynamic socio-political landscape.

Problem Description:

This statistical analysis assignment aims to explore the "hostile mediator phenomenon" (HMP), a broader concept related to the well-known "hostile media phenomenon" (HMP) where rival partisans perceive various public institutions as biased against their sides. The dataset focuses on Facebook bias perceptions among Democrats and Republicans, and gender differences in perceived bias about President Trump.

1. Facebook Bias Perceptions among Democrats and Republicans

Variable name in the dataset	What was measured	Possible responses	Scale of measurement
gender	Gender	0 = Male 1 = Female	Nominal
facebook_item	Facebook bias	The range of answers is between 1 to 9, with: 1 = Bias in favour of Trump 5 = Neutral 9 = Bias against Trump	Quasi-interval (Continuous)
partisan_group	Partisan groups	0 = Democrats 1 = Independents / Other 2 = Republican	Nominal

Table 1: Bias perceptions on both Democrats and Republicans

The researcher wanted to know whether Republicans and Democrats (separately) perceive a bias in favour or against President Trump on Facebook (facebook_item). To test his hypotheses, please execute the following six steps:

Solution

Hypothesis Testing Steps:

Exclude non-Democrats.
Test if Democrats' Facebook bias significantly differs from neutrality.
Cancel the filter.
Exclude non-Republicans.
Test if Republicans' Facebook bias significantly differs from neutrality.
Cancel the filter.

1. Researcher's Hypothesis: The researcher's hypothesis is two-tailed, as it posits that Democrats' Facebook bias can be either lower or higher than neutrality (5).

2. Test Conducted in b and e: One sample t-test was conducted.

3. Results of Tests for Democrats and Republicans: Democrats:

Mean: 4.86, SD: 0.085, Sample Size: 787
Result: No significant difference from 5 (p = 0.1037).

Republicans:

Mean: 6.28, SD: 0.094, Sample Size: 787
Result: Significant difference from 5 (p < 0.0001).

Summary Statistics

Mean	4.8614994
Std Dev	2.3852557
Std Err Mean	0.0850252
Upper 95%Mean	5.0284027
Lower 95%Mean	4.6945961
N	787

Test Mean

Hypothesized Value	5
Actual Estimate	4.8615
DF	786
Std Dev	2.38526

t Test

Test Statistic	-1.6289
Prob >It	0.1037
Prob >t	0.9481
Prob	0.0519

Table 4: Test for Democrats and Republicans

4. Interpretation: For Democrats, bias perception is neutral. Republicans perceive a significant bias against Trump on Facebook.

2. Gender Differences in Facebook Bias

The researcher wanted to test for gender differences in perceived bias on Facebook for or against Trump (facebook_item)

Test Required: Two independent t-tests.
Test Results: Levene’s test showed unequal variances. Independent t-test: no significant difference (p = 0.0888).
Report on Test Results: Men and women have a similar perception of Facebook bias related to Trump (Male: M = 5.68, Female: M = 5.47).
N Discrepancy Explanation: The N value is lower due to missing values in Facebook bias.

3. Comparative Analysis of Traumatic Event Experiences

The following table presents the results of comparisons between means of parents and their adolescent children in their experience of traumatic events.

All quantitative variables were tested on a scale of 1 (Low) to 7 (High).

Variable name in the dataset	What was measured	Possible responses	Scale of measurement
gender	Gender	0 = Male 1 = Female	Nominal
facebook_item	Facebook bias	The range of answers is between 1 to 9, with: 1 = Bias in favour of Trump 5 = Neutral 9 = Bias against Trump	Quasi-interval (Continuous)
partisan_group	Partisan groups	0 = Democrats 1 = Independents / Other 2 = Republican	Nominal

Table 5: A scale to compare the quantitative variables

Tests Conducted: Two independent t-tests.
Significant Results: Three significant differences in exposure, individual resilience, and self-differentiation.
Analysis for 'Individual Resilience': Significant difference (p < 0.05). Parents show slightly higher resilience (M = 3.94) compared to adolescents (M = 3.73).
Analysis for 'Family Resilience': No significant difference (p > 0.05). Parents and adolescents exhibit similar family resilience.
Interpretation of p-value (for 'Self-Differentiation'): Significant difference (p < 0.05). Adolescents (M = 4.45) self-differentiate more than parents (M = 4.23).