Data Presentation and Explanation of the Process of Investigation

Question 1 (20 Marks)

The question 1 data file contains measurements on the level of growth, over a one year period, in the flower spike length of Australian grass trees in four different locations: Yanchep, Mundaring, Mandurah, and Pickering Brook.  All of the plants in each location are approximately the same age.  The measurements are recorded in cm.  The indexing of the values is such that in the data file 1 = Yanchep, 2 = Mundaring, 3 = Mandurah, and 4 = Pickering Brook.  There is also an additional column that identifies when the area was burnt.  So, although there are several data columns, you have a factor variable for location (four levels), and a factor variable for time of burn (two levels).  Depending on the choices you make in the modelling process, including the tests that you use, you can end up with quite different conclusions about the relationships present in the data set. 

For this question quite different answers can get full marks.  The key is that you convince the reader that your approach is sound.  If you reach a plausible conclusion, but the presentation quality is poor and or I find the logic you followed unconvincing. You will not get full marks for the question.  See the marking guide for further details.

Note: For any multiple comparison tests implemented as part of your answer to this question you are to use the Tukey method.  

Question 2 (20 Marks)

The response variable is the number of seeds produced from each plant.  The plants are Australian Grasstrees.  There is a factor variable: location (Yanchep and Mundaring).  The location information has been coded twice in the data file: once as a long format factor variable (Location) and once in wide format dummy variable code (Yanc and Mund).  There is also a quantitative variable (length of the plant flower spike) measured in metres that may help predict the response. 

There is no absolute correct answer for this question.  I have been able to come up with a wide variety of different models that I thought are reasonable/could be defended.  There are a number of ways of looking at the data and what is important is that your reasoning makes sense.  For further details on the way marks will be allocated for the question see the marking guide.

General Marking matters

The questions focus on writing simplified responses in the format of scientific papers, which in this instance will be: Introduction, Methods and Data, Results, and Discussion.  The Assignment is essentially a capstone type assessment where you are required to bring together your knowledge of statistics and scientific writing.  The content of each section should be as described below:

Introduction – a simple overview statement of the nature of the experiment/research.  This can largely be a rephrasing of the detail I provide to explain the question context.  You are not expected to do any primary research on the topic to write the introduction.      

Methods and Data – as you will be using data sets provided you do not need to make any statements about the method used to collect data.  Rather, you need to include statements related to the approach/steps used for data analysis. 

Results – this section will contain summary table(s) for the testing you have conducted.  You also need to include very brief summary statements that explain how the results should be interpreted. 

Discussion – this section is meant to be relatively short, however this is also the section that allows you to demonstrate a high level of understanding of what you have done.  As per the introduction, you are not expected to demonstrate specific knowledge of the science aspects here, but demonstrate an understanding of the modelling process.

Other comments

The assignment is about data presentation and explaining the process of investigation.  A visual plot is an essential part of each question answer, so make an effort to get the plot looking nice.  Think hard about the different types of plots you can generate, and the specific visual representations that would be of value, and any limitations of a specific plot, given your final conclusions.  All figures and tables must have appropriate caption labels.  Also think hard about how you might use a table format to assist in presenting information.  As a guide to the standard required for tables and figures look at the production standard in some journal articles.  Also follow the journal practice of providing plots that reproduce clearly in black and white if you are going to submit in black and white, which is what I expect.  Tables might have a structure like that shown below.  Do not report Standard Errors, and p-values, and t-values.  These three sets of measures contain the same information so just report one set of values.  Convention differs by discipline.  Below I have one example using standard errors and one using p-values.