Introduction

This lab report is an individual assignment. While discussions about the data and R code with classmates are allowed, forming questions and answering them about the data should be done independently. Please follow the Report Guidelines available on the course helper webpage before starting the report.

Data Retrieval

library(palmerpenguins)
library(ggplot2)
library(dplyr)
penguins <- penguins %>%  tidyr::drop_na() # drop any missing values

Data Description

Ready to dive into the fascinating world of Antarctic penguins? Our dataset, drawn from the palmerpenguins package, offers a detailed look at 333 penguins across 8 variables, capturing the diversity of these creatures across different islands. The data encompasses various measurements like bill length, flipper length, and body mass, along with categorical data on species, sex, and the year of observation. This rich dataset serves as an excellent foundation for exploratory data analysis and hypothesis testing. Let’s embark on this analytical journey and uncover the hidden stories of these penguins.

Column Name	Description
species	A categorical variable denoting penguin species (Adélie, Chinstrap, and Gentoo)
island	A categorical variable denoting island in Palmer Archipelago, Antarctica (Biscoe, Dream, or Torgersen)
bill_length_mm	A quantitative variable denoting bill length (millimeters)
bill_depth_mm	A quantitative variable denoting bill depth (millimeters)
flipper_length_mm	A quantitative variable denoting flipper length (millimeters)
body_mass_g	A quantitative variable denoting body mass (grams)
sex	A categorical variable denoting penguin sex (female, male)
year	A quantitative variable denoting the study year (2007, 2008, or 2009)

Required Analyses:

Formulate three research questions that pave the way for hypothesis testing using statistical tests such as ANOVA, z-tests, t-tests, Chi-square tests, and linear regression parameter significance tests. Accompany these tests with a confidence interval to quantify key parameters, providing insights into the population based on your findings. Whether differences are statistically significant or not, your analysis will illuminate various aspects of the penguin population. For example, if you find two groups have a statistically significant difference then your CI should quantify how different they are in the population. If you don’t find a statistically significant difference, then you should get a CI for your overall population parameter (not separated by groups if you didn’t find a difference).

Expectations

Remember, EDA (Exploratory Data Analysis) is as crucial as finding a p-value. Incorporate basic descriptive statistics and graphs to support all your analyses.
Highlight any statistically significant findings with detailed EDA, using specific statistics and/or confidence intervals to articulate the relationships between variables.
Examine the assumptions necessary for your chosen inference methods.
Be vigilant for outliers in your quantitative data during EDA. Decide whether to include or exclude them based on their influence on your analysis.
In your introduction, go beyond merely stating the source of your data. Delve into the background information provided with the dataset to enrich your report’s context.

Submission Details

Submit your report as a single PDF file to Gradescope. Ensure that your R-code is included only in the appendix and not in the main body of your report. While including plots and numerical summaries in the main content is acceptable, the R code that generates them should be hidden.

To hide the R code in your report:

Use the chunk option echo = FALSE to prevent the code from displaying but still execute it. This is ideal for hiding the code but showing the output (e.g., plots or summaries).
Alternatively, you can use include = FALSE to both hide the code and prevent its output from being shown in the document. This is useful if you want to execute some preparatory code without showing any results in the report.

In RStudio’s R Markdown, insert a new R chunk with Ctrl + Alt + I or Cmd + Option + I. Choose the appropriate chunk option based on your specific needs for each section of your report. For example, to hide code but execute it, use the chunk header {r echo=FALSE}.

Report Structure

In statistics, experts often share their work through articles in magazines (called journals) or books. The main type of article is called a ‘primary paper’. It’s like a story of what the researcher did, how they did it, and what they found out. Since working with data (or ‘data analysis’) is a big part of statistics, it’s important for students, even in beginner classes, to learn how to write in this style.

Abstract

This short section (about 5% of your paper’s length) should quickly summarize your work. It should cover the reason for your analysis, the main question you’re addressing, the statistical methods you used, your findings, and what you believe these findings mean. Essentially, by reading just the abstract, someone should grasp the essence of your entire paper.

Introduction

Start by setting the scene for your analysis: provide some background and mention previous related studies, if available. Clearly state the main focus or objective of your data analysis. Explain why this focus is important or interesting. Briefly outline the methods or approaches you took, highlighting if you conducted an experimental or observational study.

Methods

Detail the techniques you used both to gather and analyze your data. The description should be clear enough that someone else could replicate your analysis. Specify how the data was collected, the statistical tests performed, and any relevant setups. While you can reference the manual for in-depth statistical test details, avoid including complex mathematical formulas. Write in the past tense, describing the entire analysis approach as though you conducted it independently. The aim is to provide a concise overview of your analysis method, not a step-by-step breakdown.

Results

This section should present the main findings from your data analysis. Describe the key patterns, trends, and notable differences you observed. While visuals like charts and graphs will illustrate your results, it’s essential to explain them verbally too. State the primary data trends and support them with statistical outcomes in brackets. Direct the reader to any relevant figures or tables. Remember, this section is for presenting findings only; interpretations and broader implications should be reserved for the following Discussion section.

Discussion and Conclusion

Start by recapping your main findings and linking them to your initial questions and hypotheses. Examine your results in the context of existing knowledge and research. Delve into broader implications, and consider how your results fit into a larger landscape. Address potential issues in your analysis like violation of assumptions, outliers, bias, and other errors that could impact the validity of your conclusions. Reflect on any ethical constraints tied to your study and their potential influence.

Grading Rubric

Aspect	Excellent: 10	Satisfactory: 7	Needs work: 5
Introduction	Introduction captivates with original insights and creative angle, setting a compelling stage for the analysis.	Introduction shows some original elements and creativity but lacks a captivating edge.	Introduction is generic, with no original insights or creative elements.
—	—	—	—
EDA & Data Description	Comprehensive exploration and succinct description of data. Insightful initial analyses that enhance understanding of the dataset.	Basic exploration and description of data. Some relevant analyses but lacks depth.	Minimal or incorrect data exploration. Descriptions are vague or missing.
—	—	—	—
Properly Labeled and Captioned Figures	Figures are well-crafted with clear, informative labels and captions. Visuals enhance the narrative and data understanding.	Figures are adequately labeled and captioned, but visuals do not significantly aid comprehension.	Figures are poorly labeled or not captioned, providing little to no informational value.
—	—	—	—
Numerical Summaries Table	Tables are well-organized, offering clear, comprehensive statistical summaries that enhance the analysis.	Tables present essential data but lack clarity or detail, making them less effective.	Tables are poorly organized, lacking critical data or are confusing.
—	—	—	—
Statistical Methodology	Methodology is robust, well-justified, and appropriately complex, showing deep understanding of statistical tools.	Methodology is sound but lacks sophistication or full justification, showing only moderate understanding.	Methodology is flawed or overly simplistic, indicating a poor grasp of statistical principles.
—	—	—	—
Results Section	Results are detailed, clearly expressed, and supported by the data and statistical analysis, demonstrating thorough analysis.	Results are reported with basic details and support but lack depth or clarity in places.	Results are incomplete, unclear, or not supported by the data or analysis.
—	—	—	—
Conclusion	Conclusion draws powerful, well-supported inferences from the results, offering meaningful implications and future directions.	Conclusion summarizes the results adequately but with limited implications or future insight.	Conclusion is cursory or fails to adequately tie results to broader implications or suggestions.
—	—	—	—
Statistical Jargon & Application	Employs a rich vocabulary of statistical jargon correctly, enhancing the academic rigor of the analysis.	Adequate use of statistical terms, though sometimes imprecise or sparingly used.	Limited or incorrect use of statistical jargon, detracting from analytical depth.
—	—	—	—
Submission Promptness	Both the proposal and final report are submitted on time or ahead of schedule, demonstrating good time management and commitment.	The proposal or final report is slightly late, indicating room for improvement in time management, but does not necessitate resubmission.	The proposal or final report is late, suggesting significant issues with time management and may require resubmission.
—	—	—	—
R-code Organization (in Appendix)	R-code is meticulously organized and selectively commented, demonstrating efficiency and clarity.	R-code is organized with functional comments but lacks the precision and clarity of top submissions.	R-code is disorganized or missing critical elements, indicating a need for further refinement.