Project 3
Applying statistical tests and estimation
ADMINISTRATION
Weighting: Report due: Submission:
Worth 15% of final mark
5pm Monday 7th October
Submit the project to the MyLO Assignment Dropbox
MATCH BETWEEN THE INTENDED LEARNING OUTCOMES FOR THE UNIT AND CRITERIA FOR THE TASK
Objectives
Task specific criteria
Summarize and explore large data sets using appropriate numeric and graphical tools in order to communicate statistical concepts to both scientific and lay audiences.
Communicate statistical concepts to both statistically literate and lay audiences in a written report with a non- technical front-piece and a statistical appendix
Identify and apply appropriate statistical techniques to make inferences based on data.
Apply appropriate statistical techniques to make inferences about the fishing data set
Perform common statistical analyses in a statistical computing package.
Use Excel to perform hypothesis testing and to construct confidence intervals
PROJECT: YOUR ROLE
Based on the suggested analyses listed in the Task descriptions below, prepare a report that comments on the factors that are related to the measures of fishing performance. Your report should be no more than six A4 pages, it should include:
1. Frontpiece which defines the aims and presents the principal findings in a form which would be easily understood by a non-statistical readership.
2. Statistical Appendix which includes the results of the suggested analyses and describes the statistical methods employed. It can be assumed that persons reading this section of the report are statistically literate. The appendix should include all results which provide the evidence for claims made in the report. You may need to use tables to summarise the information.
Your responsibility
It is fine – encouraged in fact! – to work together with members of your tutorial group to help each other out with Excel and discuss your findings.
However, the report must be an individual effort
with you being solely responsible for the style and content and YOU must be able to explain any statements or conclusions
if called upon to do so.
NEED HELP?
You should always feel free to contact me (the unit coordinator) by email (Danijela.Ivkovic@utas.edu.au) or by posting a question to the discussion forum.
Tutors are there to help you in the project work sessions and in the computing sessions.
KMA153, 2019 S2
Project 3
FROM DESCRIPTIVE STATISTICS TO INFERENTIAL STATISTICS
In Project 1, you used methods of Descriptive Statistics to summarise data from fishing boats. The analysis produced information such as that presented
Table 1. Mean performance for fishing boats of different ages.
in Table 1. When we did Project 1 we saw differences between types of boat but we didn’t yet have the tools to establish whether the observed differences were likely to be real effects, or whether they could just be due to chance (sampling variation).
For example, is the fact that old boats took longer on fishing trips likely to be repeated in the future?
Statistical testing or estimation is required to answer these kinds of questions.
Age of Boat old new
Catch [tonnes] Value [in $1000] Effectiveness [$/h] Time [hour]
42.4 42.4
67.50
627.2
114.8 99.6
64.50 683.9
Analysis 1.
Comparison of means.
For the variables VALUE and TIME, determine whether there is evidence of a difference in mean levels between the categories of EXPERIENCE, AGE OF BOAT, GENERAL EQUIPMENT and SEARCH EQUIPMENT.
Analysis 2. Comparison of means within experience subgroups.
Performing separate analyses for the boats with experienced skippers and for the boats with inexperienced skippers, for the variables VALUE and TIME, determine whether there is evidence of differences in mean levels between the categories of AGE OF BOAT, SEARCH EQUIPMENT and GENERAL EQUIPMENT.
Analysis 3. Estimation of differences in performance parameters.
For each of the comparisons made in Analysis 1 or Analysis 2 construct a 95% confidence interval for the difference in means.
Analysis 4. Comparing probabilities of attaining a high value of catch for experienced and inexperienced skippers.
Determine the median CATCH for all boats and form a binary variable which takes the value high if the value of CATCH is above the median, and takes the value low if it is equal to or below the median. Construct a frequency table of EXPERIENCE versus the binary form of the CATCH variable. Hence, compute the proportion of high catch trips separately for experienced skippers and inexperienced skippers, and construct 95% confidence intervals for the expected probability of a high catch trip for experienced and inexperienced skippers.
Analysis 5. Relationships between categorical factors.
Test whether there is an association between EXPERIENCE, AGE OF BOAT, SEARCH EQUIPMENT and GENERAL EQUIPMENT.
Project CHECKPOINTS
Remember that as part of the performance requirements for this unit you must get tutor sign-off on at least 7/10 CHECKPOINTS. Project 3 has three CHECKPOINTS
KMA153, 2019 S2
Project 3
TIMELINE & CHECKPOINTS
Week 8: CHECKPOINT 6 Complete the analyses that ask you to perform t-tests for differences in means (analyses 1 & 2). Have the analysis where you test for a difference in mean TIME between experienced and inexperienced skippers checked by your tutor.
Week 9: CHECKPOINT 7 Complete the analyses that require you to construct confidence intervals (analyses 3 & 4). Have the analysis where you construct a CI for the difference in mean TIME between experienced and inexperienced skippers checked by your tutor.
Week 10: CHECKPOINT 8 Complete the chi-square tests for analysis 5. Have the analysis where you conduct a chi-squared test of association between the variables EXPERIENCE and AGE OF BOAT checked by your tutor.
In order to avoid panic before the due date, as you go you should consider how you will write up findings relating to the tests and what sort of information you will need to show in the statistical appendix to support your findings.
See MyLO for a Project 3 Survival Guide that contains some further tips and for the Project 3 Marking Guide.
KMA153, 2019 S2