Perform and present a statistical analysis of the data set using the R methods

Introduction

Your third assignment is designed to assess your skills on model building and assessing significance of predictors using generalised linear modelling (GLM). You be working with the skink data set that was presented in lectures (see ANCOVA lecture):

Cunningham GD, While GM, Wapstra E (2017) Climate and sex ratio variation in a viviparous lizard Biol. Lett. 13: 20170218. http://doi.org/10.1098/rsbl.2017.0218

The data you will be analysing was posted online on dryad: https://datadryad.org/stash/dataset/doi:10.5061/dryad.hh67j

This manuscript presents a study investigating the date of birth, and the sex-ratio, of offspring of the Tasmanian endemic lizard (Niveoscincus ocellatus) for 16 years at two sites. For this project you will only be investigating the data relevant to the sex ratio of offspring.

The file: Climate_and_sex_ratio_Sup_Data.csv contains 2614 rows of clutch information. For each observation you are given data on:

• [Year] The year the data were collected {2000-2015}
• [Site] The location where the observations were collected: {“CP”, “OR”}. CPis the highland site (elevation:1150m)that is cooler,and OR is the lowland site (elevation: 30m) that is warmer.
• [DOB] Day of birth (1 = January 1^st)
• [Males] Number of males in the clutch
• [Females] Number of females in the clutch
• [GestPeriodTemp] Mean maximum daytime temperature (Celsius) during the first half of gestation, which corresponds to the critical sex-determining period
• [MaternalID] Unique identifier for each female mother

Instructions

Your task is to perform and present a statistical analysis of the data set using the R methods presented in lectures. Specifically, you are to address the question:

Question: Is there statistical evidence that the proportion of offspring that are female in each clutch has, on average, changed over time and/or differs between the two sites? Also, is this fraction impacted by the temperature that mothers experience during early gestation?

Hints

The only biology that you need to be aware of to complete this project is that for many reptile and fish species temperature is known to impact the propensity for offspring to develop into females (often warming induces feminisation). Also, each female Niveoscincus ocellatus may produce a single clutch each year, and clutches can have up to 9 offspring (2 offspring being most common). More cool things about the lizards and data collection can be found in the manuscript.

Note that the original data set involved multiple observations of clutches from the same lizard, which is often referred to as repeated measures. Later in the unit I will show you how best to model repeated measures using random factors. For this assignment I do not want you to use random factors; your model should be constructed and fit to the data using the glm function.

I have provided you with the complete data set provided by the authors. Some of the columns of this data set may not be relevant for your analysis. You need to decide which predictors are treated as factors or as numeric.

Marks are awarded for clarity on what you did and what you concluded from your analysis. Use text and refer to figures to help justify your model, describe effects, and clarify your answer.

You should only present figures if you provide some accompanying text that discusses the relevance of the figure. Any figure or large sections of R output that you do not use to support your conclusions will be penalised.

Note that justification of your conclusions involves showing that your final model is consistent with the data.

Perform your analysis using a R Markdown document and knit the document to a PDF and submit it to MyLO (see Week 5 Content for the link). You do not need to submit the R Markdown document just the output it produces.

Your R code should include comments (using #) to clarify what key parts are doing. Don’t go overboard with code comments so that your code is cluttered; the job of comments is to improve clarity.

KMA353 is an applied statistics unit. An important learning outcome for this unit is that you demonstrate that you can correctly interpret statistical output. To do this you need to write with clarity. I expect correct punctuation to be used. RStudio has a spell checker.

Take care when generating figures. You must place units on axes and legends where appropriate. Axes should be clearly named. All figures you generate must have a brief and informative figure caption. The caption can be provided as text under the code chunk that generated the figure. All figures should be enumerated. For example, the text under your first figure could begin something like:

Figure 1: The relation between …

Alternatively, you can add a caption within ggplot using labs(caption = "Figure 1: The relation between ..."). Embed \n within the caption text if you want to split the text over multiple lines.

Make sure your figure caption explains any potential ambiguities in your figure; for example, state what the error bars represent. Look at published scientific articles for examples of the types of information usually provided in a figure caption.

If you are having trouble with generating a PDF then you may need to install TinyTex on the computer you are working with. See the end or the workshop of Week 2 for details how to make this installation within RStudio. If you absolutely cannot create a PDF within RStudio, then knit to a Word document and in Word save it as a PDF.

Workshops are times dedicated to help you with your learning of concepts presented in lectures, and for seeking clarification with project assignments. Feel free to discuss ideas about the project with others, however the writing up of the assignment must be your own work. Very similar chunks of writing observed across multiple assignments will be assessed for plagiarism.