1
General Social Survey 2018 Dataset Blog Post
Students Name
Institution Affiliation
Course Title
Due Date
2
General Social Survey 2018 Dataset Blog Post
This analysis was based on 2018 General Social Survey dataset. Income inequality
remains a pressing social issue. By examining how age and sex affect income, we can gain
insights into economic disparities in society. The General Social Survey (GSS) provides a
valuable dataset for exploring these patterns, making it an ideal choice for this analysis.
Specifically, I asked:
Research Question
Does the respondent's sex and age impact the respondent's income in constant dollars?
Hypothesis
H0: There is no statistically significant effect of the respondent’s sex and age towards
the respondent's income in constant dollars.
H1: There is a statistically significant effect of the respondent’s sex and age of
respondent towards the respondent's income in constant dollars.
Variable 1: Respondents Age
Variable 2: The Yearly Income of the Respondent.
Descriptive Statistics Summary
Histogram with Density Curve
The histogram with a density overlay shows that most respondents earn below
$50,000, while a long right tail extends towards higher incomes. This indicates income
inequality, where only a small fraction earns substantially more than the majority.
3
Boxplot
The boxplot confirms this skew. While the majority of incomes cluster tightly around
the median, several high-income outliers stretch the boxplot upward. These outliers are
realistic given that some respondents occupy top-paying jobs.
Table 1: Descriptive Table
Descriptive table
Std. Change Statistics
Adjusted Error of R
R R the Square F Sig. F
Model R Square Square Estimate Change Change df1 df2 Change
1 .281a .079 .074 31020.590 .079 16.632 2 387 .000
a. Predictors: (Constant), AGE OF RESPONDENT, MALE
b. Dependent Variable: RESPONDENT INCOME IN CONSTANT DOLLARS
Table 1 indicates R square equals .0.079. This indicates that the model can explain
7.9% of the variance in the respondent income in constant dollars.
4
The Histogram Boxplot
Figure 1: Test for significant outliers in independent variable age of respondent
According to figure 1, for variable 1, our independent variable age of respondent has
no significant outliers.
5
Figure 2: Test for significant outliers in dependent variable respondent income in constant
dollars.
According to figure 2, for variable 2, our dependent variable respondent income in
constant dollars has no significant outliers since it is normal to have extremely high-paid
individuals.
Measurement Validity and Reliability
Validity: Income is self-reported, raising the possibility of bias—respondents may
overestimate or underreport. Age and sex are straightforward demographic variables and thus
more valid.
Reliability: Income measures can vary depending on how the question was
interpreted (gross vs. net income, monthly vs. yearly reporting). Repeated surveys with the
same respondent might not yield identical results. However, demographic variables like age
and sex are highly reliable.
6
Future Research
We concluded that our data were randomly and independently sampled because we
had no access to the research design, and hence the assumption of independence of
observation was not violated. We should have more than two independent variables if no
assumptions were violated, including the assumption that our dependent variable should be
measured on a continuous scale. Multiple regression was performed to see if the gender and
age of the respondents affect their income in constant dollars.
7
Reference
Wang, T., Li, Q., Chen, B., & Li, Z. (2018). Multiple outliers detection in sparse high-
dimensional regression. Journal of Statistical Computation and Simulation, 88(1), 89-
107.