Assignment 2: Interpreting Quantitative Findings
2661657L
17/12/2023
# Install packages if missing
list.of.packages <- c("tidyverse", "moderndive", "jtools", "devtools", "vtable", "devtools")
new.packages <- list.of.packages[!(list.of.packages %in% installed.packages()[,"Package"])]
if(length(new.packages)) install.packages(new.packages)
# # Word count
# devtools::install_github("benmarwick/wordcountaddin", type = "source", dependencies = TRUE)
# Load packages
library(moderndive)
library(tidyverse)
library(haven)
### Download and read data ----------------------------------------------------
# create folder
dir.create('data')
#download data sets
download.file('https://www.ark.ac.uk/teaching/NILT2012GR.sav', 'data/nilt2012.sav', method='curl')
download.file('https://www.ark.ac.uk/teaching/NILT2012LGBT.sav', 'data/nilt2012_2.sav', method = "curl")
# read datasets
nilt_1 <- read_sav("data/nilt2012.sav")
nilt_2 <- read_sav("data/nilt2012_2.sav")
### Join datasets ------------------------------------------------------------
# drop duplicated columns in nilt_2
nilt_2 <- select(nilt_2, -househld: -healthyr, -wtfactor)
# Join data sets in 1
nilt <- left_join(nilt_1, nilt_2, by = "serial")
# Coerce variables ---------------------------------------------------------
# Identify numeric variables
unique_levels <- apply(nilt,2,FUN=function(x)length(na.omit(unique(x))) >= 8)
numeric_vars <- names(unique_levels)[unique_levels == TRUE]
# exceptions as numeric
exc_num <- c('highqual', 'tea', 'ansseca')
numeric_vars <- numeric_vars[!numeric_vars %in% exc_num]
numeric_vars <- c(numeric_vars, 'wtfactor')
# Coerce to their type
nilt <- nilt %>% mutate(across(all_of(numeric_vars), as.numeric))
nilt <- nilt %>% mutate(across(!all_of(numeric_vars), as_factor))
# drop unused levels
nilt <- droplevels(nilt)
# Save data ------------------------------------------------------------------
# save as rds
saveRDS(nilt, "data/fullnilt_2012.rds")
# clean global environment
rm(list=ls())Word count: 3498
Introduction
The examination of social attitudes within diverse and multicultural societies has been a prominent focus in academic research. This study endeavours to investigate the interplay between sexual orientation and annual income to uncover potential disparities. Drawing inspiration from prior research on this subject a research question will be answered, and hypotheses investigated. We are told “research shows LGBTQ+ workers encounter a startling earnings gap in employment”. This indicates that this is an area requiring further research to ascertain the true extent of this issue. This inquiry delves into how sexual orientation intersects with economic outcomes. Furthermore, this report will present findings related to the number of weekly hours worked by individuals from different social groups to identify what demographics typically occupy the most well paid, high-ranking jobs. This will provide further insight into a potential pay disparity and explore the existence of discrimination in terms of who has access to these jobs.
In the exploration of attitudes towards minority populations, traditional studies have presented two opposing theoretical orientations. Freeman (1948) suggests that sizable minority populations may elicit hostility, particularly when perceived as a threat or in situations of competition for limited resources. Conversely, contact hypothesis (Allport, 1954) proposes that positive interactions arise from increased exposure and contact with diversity across different contexts, this is supported by numerous scholars with research to boot (Christ et al., 2014). These theoretical frameworks serve as a foundation for understanding the shaping of attitudes towards minority groups, including individuals of different sexual orientations. Furthermore, this begs the question, in the modern world in which we live, does discrimination still manifest itself by way of pay differences and if so, what is the true severity of this issue?
Research Question:
This study, utilizing data from a specified timeframe, aims to address the following research question:
“To what extent is there a statistically significant association between sexual orientation (homosexual, bisexual, and other) and annual income among individuals in Northern Ireland?”
Hypotheses:
Null Hypothesis (H1): No statistically significant association exists between sexual orientation and annual income among individuals in Northern Ireland.
Alternative Hypothesis (H2): There is a statistically significant association between sexual orientation and annual income among individuals in Northern Ireland.
Ultimate Hypothesis:
Expanding on the research question and null hypothesis, the ultimate hypothesis aligns with the investigation into potential associations between sexual orientation and annual income:
H: The annual income of individuals in Northern Ireland demonstrates statistically significant variations based on sexual orientation indicating that sexual orientation is a pertinent factor in predicting economic outcomes.
This exploration of hypotheses aims to provide valuable insights into potential economic disparities associated with sexual orientation, offering illumination on the dynamics within the specified population. The subsequent sections will provide a detailed examination of the dependent variable, annual income, ensuring a thorough analysis of the research question.
Data and Methods
The dataset employed in this study originates from the Northern Ireland Life and Times (NILT) survey of 2012, accessible through ARK Northern Ireland, a collaborative initiative between them and Queen’s University Belfast. This survey encompasses responses from 1,204 participants who consented to a face-to-face interview. Therefore, this offers a comprehensive view of attitudes and demographics within Northern Ireland during the specified period.
The Dependent Variable:
The dependent variable for this investigation is annual income, measured in British Pounds (£). Income stands as a pivotal metric for evaluating economic outcomes and disparities in the context of sexual orientations (gay, bisexual, and other). Participants were requested to disclose their annual income, both before and after tax, and this variable will be examined in relation to their sexual orientation in order to investigate a possible correlation between those variables. The figures used will represent the income of individuals after tax.
nilt <- read_sav("data/nilt2012.sav")
nilt <- nilt %>% mutate(orient = as_factor(orient))
nilt_subset <- select(nilt, orient, persinc2)
install.packages('vtable')
library(vtable)
hist(nilt_subset$persinc2)
Tail Orientation: The right tail of the histogram is longer than the left tail. The bulk of the data points are concentrated on the left side of the histogram, while the right side extends further, containing a few high values. This tells us that there are outliers towards the right side of the histogram indicating more affluent individuals responded to the survey. This increases the reliability of our research as it ensures a diverse demographic. Furthermore, this tells us that the vast majority of respondents belong to the working or middle class which we can see due to the distribution of values and the fact they are saturated on the left
Mode, Median, and Mean Relationship:
-The mode (the most frequently occurring value) is less than the median.
-The median (the middle value when the data is arranged in ascending or descending order) is less than the mean.
-The mean (average) is influenced by the presence of higher values in the right tail, pulling it in that direction.
Data Distribution: The majority of observations cluster towards the lower end of the scale, with fewer instances of higher values. This suggests that the data is concentrated in the lower range, but a few extreme values pull the overall distribution to the right. The majority of respondents fall into the 0-25000 range.
Interpretation: In practical terms, a right-skewed histogram indicates that there are outliers or extreme values on the higher end of the scale. For example, in this dataset of household incomes, the majority may earn moderate incomes, but a few extremely high-income households can cause the distribution to be right skewed.
Independent Variables:
In addition to the dependent variable, independent variables are incorporated into the analysis:
Sexual Orientation:
This categorical variable classifies participants into straight, gay, bisexual, or other groups based on their self-identified sexual orientation.
Participants were asked the question, “Can you tell me which of these best describes you? Sexual orientation” To answer, they were then given the choice of the following options; -1- I am heterosexual or straight, -2- I am gay or lesbian (homosexual), -3- I am bi-sexual, -4- Other, -999- Not answered/refused
The following summary table tells us about the results from this question from the 1191 individuals who answered this question
library(vtable)
sumtable(nilt_subset, vars = c('orient'))| Variable | N | Percent |
|---|---|---|
| orient | 1191 | |
| … Not answered/refused | 0 | 0% |
| … I am heterosexual or straight | 1173 | 98% |
| … I am gay or lesbian (homosexual) | 14 | 1% |
| … I am bi-sexual | 2 | 0% |
| … Other | 2 | 0% |
We are told that 98% of people identify as heterosexual, this is expected and is in line with averages worldwide. The ONS tells us that “43.4 million people (89.4% of the population aged 16 years and over) identified as straight or heterosexual. 748,000 (1.5%), described themselves as gay or lesbian.” Therefore, the fact that those who identify as homosexual, or bisexual are such a small minority is not a surprise and exemplifies the necessity for any disparities in pay to be researched.
1% of individuals identify as homosexual and 0.16% of individuals identify as bisexual. This gives us an insight into the numerical details of the demographic that is to be investigated as to our research question.
Hours Worked
This variable will give an insight into whether any pay disparity is due to the impact of sexual orientation or infact, due to the number of hours worked per week
Participants were asked to disclose their average working hours per week, answering the question “How many hours per week do you normally work in your job?”. They then answered along a numeric scale from 0-100
This research project will be centered around a regression analysis in order to discern what link there is, if any, between sexual orientation and personal income. Regression analysis allows one to model and quantify the relationship between a dependent variable and one or more independent variables. This is particularly useful when considering our aim is to understand how changes in one variable are associated with changes in another. Furthermore, regression analysis allows you to control for confounding variables by including them as covariates in the model. This helps in isolating the effect of the independent variable(s) on the dependent variable. Regression analysis helps in identifying which independent variables are statistically significant predictors of the dependent variable, providing insights into the relative importance of different factors. Also, we are told by Hair et al (2019) about the utility of regression in testing hypotheses, controlling for confounding variables, and identifying essential predictors. This indicates that this choice of analysis is strong for answering the research question and investigating all hypotheses. # Results and discussion
library(tidyverse)
library(vtable)
library(haven)
nilt <- read_sav("data/nilt2012.sav")
nilt <- nilt %>% mutate(orient = as_factor(orient))
nilt_subset <- select(nilt, orient, persinc2)
ggplot(nilt, aes(x = persinc2, y = orient)) +
geom_point(position = "jitter") +
labs(title = "Annual Personal Income after tax vs Sexual Orientation",
x = "Income", y = "Sexual Orientation" )
This scatterplot graph elucidates the relationship between individuals’ income and their sexual orientation. A conspicuous pattern emerges, revealing that a significant majority, post-taxation, fall within the income bracket of under £20,000 per annum. As expected, this aligns with the contemporaneous national average, as articulated in the SPI report (2012/13), citing annual median income figures of £21,000 before tax and £18,700 after tax. Notably, individuals identifying with a minority sexual orientation—comprising gay, lesbian, bisexual, or other—predominantly populate this income stratum, indicative of their association with the working and low-middle classes. However, a more nuanced analysis exposes that merely 33% of those identifying with a minority sexual orientation surpass the national average income, underscoring that the preponderance of individuals within the homosexual, bisexual, or “other” categories earns salaries commensurate with or below the national mean.
Furthermore, upon scrutinizing higher income brackets, a salient revelation materializes; none of the individuals identifying with a minority sexual orientation report an annual income surpassing £40,000 post-tax. Intriguingly, every individual earning beyond £60,000 post-tax identifies as heterosexual. This nuanced examination underscores a palpable disparity between those earning at the higher echelons of income and those receiving an average salary, notably linked to sexual orientation. While a significant proportion of heterosexual individuals earn an income in line with the national average, the upper-income stratum is exclusively populated by heterosexual individuals. This stark demarcation suggests a correlation between sexual orientation and income, delineating a noteworthy socio-economic divide within the studied population.
library(tidyverse)
library(vtable)
library(haven)
nilt <- read_sav("data/nilt2012.sav")
nilt <- nilt %>% mutate(orient = as_factor(orient))
nilt_subset <- select(nilt, orient, persinc2)
ggplot(nilt_subset, aes(y = orient, x= persinc2)) +
geom_boxplot() +
ggtitle('Annual Income vs Sexual Orientation')
The boxplot visually encapsulates the distribution of income among individuals based on their sexual orientation, providing a succinct representation of the dataset’s central tendencies and variability. Notably, a substantial concentration of incomes, predominantly falling below the median, is observed across all sexual orientations. This pattern mirrors broader socio-economic trends, underscoring the prevalence of salaries within the lower to mid-income brackets. Within the cohort identifying with minority sexual orientations—encompassing individuals who identify as gay, lesbian, bisexual, or other—a discernible clustering around the lower interquartile range is evident, elucidating their affiliation with the income strata characteristic of working and low-middle-class demographics. Despite this congruence with national income averages, the boxplot highlights a notable disparity between sexual orientation groups when examining higher income percentiles. Heterosexual individuals exhibit a more dispersed distribution of incomes, with a discernible presence in the upper interquartile range. In stark contrast, individuals with minority sexual orientations show limited representation in these higher percentiles, accentuating a socio-economic divide wherein those earning beyond the median predominantly identify as heterosexual. This visual representation augments our understanding of the income dynamics across sexual orientation groups, portraying a nuanced socio-economic landscape within the studied population. Therefore, this evidence supports the hypotheses that there is a significant correlation between sexual orientation and personal income. Additionally, we are told “On average, homosexual and bisexual men in Europe, Australia and North America earn around seven percent less than their heterosexual counterparts”. This contributes to our understanding of this issue on a societal level and shows this issue is widespread. Pay disparity is clearly correlated to sexual orientation.
library(tidyverse)
library(vtable)
library(haven)
nilt <- read_sav("data/nilt2012.sav")
nilt <- nilt %>% mutate(orient = as_factor(orient))
nilt_subset <- select(nilt, orient, persinc2)
ggplot(nilt, aes(x = rhourswk, y = persinc2, color = orient)) +
geom_point(position = "jitter") +
labs(title = "Annual Income vs Hours worked vs Orientation",
x = "Hours Worked per week", y = "Annual Income" )
This scatterplot graph offers a comprehensive view of the intricate interplay between individuals’ income, sexual orientation, and their weekly working hours. A discernible trend emerges, revealing that a majority of individuals, regardless of sexual orientation, dedicate a standard workweek of around 35-45 hours. Notably, the spectrum of working hours for those with a minority sexual orientation, including gay, lesbian, bisexual, or other identifications, is notably concentrated within this range. However, a more intricate analysis unravels a distinctive pattern among heterosexual individuals who not only dominate the higher income brackets but also exclusively engage in extended work hours, surpassing 50 hours per week. This convergence implies a potential correlation between the commitment to longer working hours and the attainment of higher income, a trend exclusively observed among heterosexual individuals. Further scrutiny of the upper-income stratum, where individuals earn above £60,000 annually, unveils a consistent pattern: this cohort predominantly engages in 25-55 hours of work per week, affirming the notion that higher income brackets are achieved by a dedicated yet balanced commitment to work hours. This intriguing dynamic emphasizes the dual influence of both sexual orientation and work hours on income disparities within the examined population.
| Annual Personal Income (GBP) | |
| Sexual Orientation: Homosexual (ref.: Heterosexual) | -2,199.928 |
| (4,687.426) | |
| Sexual Orientation: bi-sexual | 627.237 |
| (9,848.683) | |
| Sexual Orientation: Other | 465.931*** |
| (52.692) | |
| Hours Worked | 5,132.971** |
| (1,973.586) | |
| Observations | 456 |
| R2 | .148 |
| Adjusted R2 | .142 |
| Residual Std. Error | 13,894.770 (df = 452) |
| F Statistic | 26.069*** (df = 3; 452) |
| Notes: | *P < .05 |
| **P < .01 | |
| ***P < .001 | |
The presented regression analysis aims to discern the intricate relationship between annual personal income (measured in GBP) and various factors, with a particular focus on sexual orientation and to further investigate this issue, the number of hours worked. The coefficients associated with each level of sexual orientation shed light on their respective impact on income. It’s crucial to interpret these results with statistical rigor:
Homosexual Orientation: The coefficient of -2,199.928 for individuals with a homosexual orientation, in comparison to the reference category (heterosexual), suggests a negative effect on annual income. This tells us there is a clear and serious pay disparity between straight and gay individuals. Given the wide confidence interval (4,687.426), caution is warranted in attributing significant importance to this effect but this is negated due to the severity of the disparity in income. This supports the hypothesis that annual income and sexual orientation are inexplicably linked, with negative repercussions being incurred for those of a minority sexual orientation.
Bi-sexual Orientation: The coefficient of 627.237 for individuals with a bisexual orientation indicates a positive effect on annual income. Again, considering the broad confidence interval (9,848.683), the precision of this estimate may be limited. This is further exacerbated by the limited sample size, only two respondants identified as bisexual which limits the reliability of this data. It is also clear from the graphs seen above that these individuals were in the lower-income bracket and as such, it is clear they receive less pay due to their sexual orientation when compared with the pay of straight individuals.
Other Sexual Orientation: The coefficient of 465.931 for individuals with an “other” sexual orientation, in contrast to the reference category, is statistically significant at the 0.001 level. This suggests a positive impact on annual income for those with an “other” sexual orientation, and the relatively narrow confidence interval (52.692) enhances the precision of this estimate.
Hours Worked: The coefficient of 5,132.971 for hours worked implies a positive association with annual income. The statistically significant p-value (< 0.01) underscores the robustness of this relationship. This indicates that, on average, for each additional hour worked, there is an increase in annual income by approximately £5000.
The overall model’s goodness-of-fit is assessed through the R-squared value (0.148), indicating that approximately 14.8% of the variability in annual income is explained by the included variables. The adjusted R-squared (0.142) accounts for the number of predictors and provides a slightly more conservative estimate.
The residual standard error of 13,894.770 represents the typical deviation of observed values from the regression line.
The F-statistic (26.069) is significant at the 0.001 level, suggesting that the model as a whole is statistically significant. This also supports the hypothesis that income and sexual orientation are directly correlated.
In summary, the regression model provides insights into the relationship between sexual orientation, hours worked, and annual income, it is essential to consider the wide confidence intervals for certain coefficients. Further nuanced analysis and caution are advised, especially when interpreting the coefficients associated with sexual orientation categories with less precision. However, in general it is clear that income and sexual orientation are linked and those from a minority sexual orientation demographic suffer negative consequences due to this.
The observed trend indicating that individuals with a minority sexual orientation tend to receive lower annual incomes finds resonance with several social science theories that explore the complexities of socio-economic disparities. The phenomenon aligns with elements of social stratification theories, which posit that societal structures often lead to the unequal distribution of resources and opportunities. The minority stress theory, a framework within LGBTQ+ studies, may further elucidate this pattern by highlighting the additional stressors faced by individuals with non-heteronormative sexual orientations. Discrimination, stigma, and biases prevalent in the workplace can impede career progression and income attainment. Additionally, insights from symbolic interactionism could be relevant, suggesting that societal norms and stereotypes might influence employers’ perceptions and evaluations, contributing to the observed income disparities. This is supported by Pereira (2009) who describes the extent of social prejudice against minority sexual orientations and how this manifests itself in reality. The intersectionality framework also plays a role, emphasizing how various social identities, including sexual orientation, intersect to shape individuals’ experiences and opportunities within the socio-economic landscape. The observed income disparities among sexual orientation groups, therefore, resonate with a multifaceted web of social science theories that scrutinize the intricate dynamics of societal structures and individual experiences.
Conclusion
In conclusion, this comprehensive quantitative investigation has meticulously explored the intricate relationship between sexual orientation and annual income among individuals in Northern Ireland. Guided by the research question “To what extent is there a statistically significant association between sexual orientation (homosexual, bisexual, and other) and annual income among individuals in Northern Ireland?” and the subsequent hypotheses, the study has yielded robust evidence supporting the ultimate hypothesis that sexual orientation significantly influences economic outcomes, resulting in a notable and unfavorable income disparity for individuals with minority sexual orientations.
Hypothesis 1, asserting the absence of a statistically significant association between sexual orientation and annual income, has been convincingly contradicted by the empirical evidence. The regression analysis has brought to light a clear and substantial negative impact on annual income for individuals with a homosexual orientation, debunking the notion of an equal economic playing field among different sexual orientation groups.
Building on this, Hypothesis 2 posited a statistically significant association between sexual orientation and annual income. The results of the regression analysis unequivocally support this assertion. Individuals with a bisexual orientation exhibited a positive impact on annual income, albeit with limited precision due to the small sample size. Meanwhile, those with an “other” sexual orientation experienced a statistically significant positive impact on annual income, highlighting the economic disparities across different sexual orientation categories.
The ultimate hypothesis, which posited that the annual income of individuals in Northern Ireland demonstrates statistically significant variations based on sexual orientation, indicating that sexual orientation is a pertinent factor in predicting economic outcomes, has been robustly substantiated. The negative repercussions incurred by individuals with a minority sexual orientation, particularly those identifying as homosexual, underscore the profound impact of sexual orientation on economic disparities.
Addressing the research question, this study not only confirms the existence of a pay disparity associated with sexual orientation but emphasizes its severity and significance. The observed income disparities resonate with established social science theories such as social stratification, minority stress theory, and symbolic interactionism. These frameworks collectively contribute to a comprehensive understanding of the multifaceted dynamics shaping economic outcomes in the context of diverse sexual orientations.
While the findings of this study provide compelling evidence, questions and concerns arise that necessitate further research. The limited sample size of individuals with a bisexual orientation warrants additional exploration with more extensive datasets to enhance the precision of estimates. Additionally, future research should delve into the nuanced mechanisms and contextual factors that intensify the negative impact of minority sexual orientation on income.
In summary, this study stands as a pivotal contribution to unraveling the complexities of economic disparities linked to sexual orientation in Northern Ireland. The evidence supports the argument that sexual orientation is a crucial factor influencing annual income, with individuals from minority sexual orientation groups experiencing a significant and detrimental economic outcome. As we navigate the complexities of societal structures and individual experiences, this study paves the way for future research endeavors aimed at addressing and rectifying these disparities, fostering a more equitable socio-economic landscape for individuals across diverse sexual orientation groups.
Word count - 3281
Bibliography
Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2019). Multivariate Data Analysis. Cengage Learning. (Chapter 8: Correlation Analysis)
https://www.bbc.com/worklife/article/20220603-the-big-lgbtq-wage-gap-problem, The big LGBTQ+ wage gap problem, By Megan Carnegie9th June 2022
Sexual orientation, England and Wales: Census 2021, Office of National Statistics
“Personal income by tax year”. HM Revenue and Customs. 31 January 2014. Retrieved 15 September 2015.
https://www.independent.co.uk/news/uk/home-news/income-gap-gay-straight-men-b1894125.html, Sexuality pay gap: Gay men ’earn £1,500 less than straight men’Ella Glover Friday 30 July 2021
Pereira A, Monteiro MB, Camino L. Social Norms and Prejudice against Homosexuals. The Spanish journal of psychology. 2009;12(2):576-584. doi:10.1017/S1138741600001943
Freeman, Kathleen & Diels, Hermann (1948). Ancilla to the Pre-Scratic Philosophers a Complete Translation of the Fragment in Diels Fragmente der Vorsokratiker. (discusses the impact of minority populations on a foundationally sound society)Harvard Univ. Press.
Allport, G. W. (1954). The nature of prejudice. Addison-Wesley.
Contextual effect of positive intergroup contact on outgroup prejudice. Oliver Christ, Katharina Schmid, Simon Lolliot and Miles Hewstone February 3, 2014