Assessment 2 Instructions
Overview
For your final assessment you are required to prepare a short analysis report applying one of the statistical techniques covered in the course. The report should be presented in IMRD (Introduction/Methods/Results/Discussion) format and should be approximately 1500 words.
Topic options
You can base your report on one of the following three options:
Interrupted time series Evaluating the effect of media coverage following publication of new drug safety research on dispensing of oral contraceptives using time series data
A dataset is supplied for each option. A more detailed description of each option is provided at the end of this document.
Report structure
Your report should be presented in IMRD format and be approximately 1500 words long. The word count does not include tables, figures, footnotes, references, appendices, or the student declaration. You should include up to five tables/figures in the main body of the report, with supplementary tables/figures included in the appendix.
Below is in an example of how you might lay out your report:
1. Student declaration
2. Introduction (200-300 words approx.)
3. Methods (550-650 words approx.)
4. Results (350-450 words approx.)
5. Discussion (200-300 words approx.)
6. References (As appropriate)
7. Appendix (As appropriate)
8. Student declaration
Introduction
200-300 words approx.
The introduction section should provide some brief background information on the research area and motivate why the research question is of scientific interest. Because you have been presented with a research topic you do not need to make this section very detailed, you should simply demonstrate that you have understood the relevance of the research question, and make the connections between the research question, the available data and the analytic approach.
Methods
550-650 words approx.
In the methods section you should describe in detail the analyses that were undertaken, and why. It is important that you demonstrate that you understand the rationale for applying a particular statistical technique and/or any alternative model specifications/model diagnostics. You should demonstrate your understanding of the statistical methods with appropriate references to the literature.
Results
350-450 words approx.
In the results section you should begin with a descriptive overview of the dataset you have used, including summaries of any key variables. Follow this with a summary of your analysis results using appropriate tables and figures, as well as describing your main results in text. Appropriate model diagnostics and/or alternative model specifications should be included but can be presented in the appendix if they are lengthy.
DiscussionAssignment Project Exam Help
200-300 words approx.
Your report should include 5-10 appropriate references to the literature, with a full list of references provided in the references section. You can use your preferred referencing style. More information on academic referencing can be found on the UNSW website: https://student.unsw.edu.au/referencing
Appendix
The appendix should include any supplementary material that does not fit in the body of the report, for example additional descriptive tables or model diagnostics plots.
Student declaration
– [ ] I understand and agree I have a backup copy of the assessment.- [ ] I understand and agree
Accessing and submitting your report
1. A Microsoft Word document with the corresponding R code provided separately
2. an R Markdown file, interweaving the text of your report, statistical code and results.
Statistical codeAssignment Project Exam Help
Your assessment will be marked according to the marking rubric on the following page. Four criteria are assessed: •
Conceptual understanding (30%) • Application of statistical method (30%) • Interpretation (30%) • Presentation (10%) 6.
Option 1: Causal inference via matching
Overview
Research question
Data
The data for this assessment are drawn from the National Children’s Study Archive. The National Children’s Study (NCS) Assignment Project Exam Help
collected birth and early childhood data on more than 5,000 children and their families in the USA from 2009-2014. The NCS teaching database contains base child-level data, as well as several topic-specific modules.
The following resources will be useful to help understand the data: * Information on the National Children’s Study Archive *
Option 2: Interrupted time series analysis
Overview
Oral contraceptives (“birth control pills”) are taken by women to prevent pregnancy and to treat other conditions such as endometriosis or polycystic ovary syndrome. There are two main types of oral contraceptive pills, the most common being the combined pill, which contains a combination of the hormones oestrogen and progestogen in various formulations and doses. The second type is the progestogen-only pill (also called the “mini pill”) that contains only progestogen.
In this assessment you will use time series data to explore the impact of the media attention surrounding the publication of this study on the PBS-subsidised dispensing of combined and progestogen-only oral contraceptives. In Australia, combined contraceptive pills containing levonorgestrel/ethinylestradiol, norethisterone/ethinylestradiol and norethisterone/mestranol are subsidised through the PBS. Other combined pill formulations are not subsidised, such as those containing drosperinone, cyproterone and desogestrel and are not captured in the data. Several progestogen-only mini pills are also subsidised, including levonorgestrel and norethisterone.
Research question
What was the impact of the media attention around the increased risk of thrombosis associated with oral contraceptive use on dispensing of combined oral contraceptives?
Data Assignment Project Exam Help
there was a change in combined oral contraceptive dispensing following the media attention. * Plot the predicted counterfactual to visualise any change in dispensing following the media attention. * Use the “control” series (dispensing of progestogen-only oral contraceptives) and comment on the findings in relation to those from the combined pill series.
Option 3: Longitudinal multilevel modelling
Overview
Children’s growth is frequently monitored during their infancy and understanding growth trajectories is important for identification of chronic disorders or other health issues, reassuring parents and providing information on the health of a nation’s children.
In this assignment you will use data including repeated measurements of children’s weight to model children’s weight change over time. The data also include information on other infant, environmental and maternal factors, and you will also explore which of them could further improve the model for children’s weight gain. You should use the nlme package in R for carrying out your longitudinal multilevel modelling.
Research question
What is the optimal longitudinal multilevel model for modelling children’s weight change?
Data
The data for this assessment are drawn from the National Children’s Study Archive. The National Children’s Study (NCS) collected birth and early childhood data on more than 5,000 children and their families in the USA from 2009-2014. The NCS teaching database contains base child-level data, as well as several topic-specific modules.
You are provided with a dataset named ncs_child_repeated.csv. This file contains information on child, mother and household demographics for 4,531 kids, with 1-4 weight observations for each child (12,654 observations overall). The data are drawn from the NCS base Child dataset, and the Child health module, including children’s weight measurements at 6, 12, 18 and 24 months after their birth. Some data preparation has already taken place, in particular, the above files have been merged, only singleton births and the oldest siblings have been included (i.e. only one child per mother), and variables with high proportions of missing data have been removed.
The following resources will be useful to help understand the data: * Information on the National Children’s Study Archive *
NCS study description and guide * Codebook for the base Child data * Codebook for the Child Health module
Hints Assignment Project Exam Help
• Conduct appropriate EDA to understand the relationship between weight and age.
• There is only one child per mother, and thus you do not need to include mother as an additional level in your multilevel modelling.
• Explore alternative functions of time to improve model fit.
• Explore other covariates that help to explain variation in children’s weight.
• Some child health variables (e.g. gastro, diarrhoea, ear_infection) were not measured on every visit occasion, so if these are included you will need to think about how to handle that.
Marking rubric
HDAT9700 Statistical Modelling II, Assessment 2 Marking Rubric
Criteria Weight Developed (HD/DN) Developing (CR/PS) Developed (FL)
Conceptual 30% Demonstrates a deep Demonstrates an Demonstrates limited understanding understanding of the relevance of understanding of the statistical understanding of the the statistical technique, it’s technique, but to a limited statistical method, its appropriateness to the research level, without delving into suitability to the research question at hand, its’s benefits over appropriate additional readings design, advantages or alternative approaches, but also its signposted in the tutorial limitations. limitations. Incorporates content material.
beyond what was covered in the
tutorial, such as the pre-reading and additional references cited in the course notes, or other scholarly work.
Application of statistical
method 30% Analysis goes beyond the basics covered in the face-to-face tutorial. Statistical method is correctly applied, selection of variables is justified where appropriate, alternative model fits are explored, where appropriate; suitable diagnostic tests are performed, presented and correctly interpreted. The rationale for carrying out models/diagnostics is clearly understood and communicated. All code has a clear purpose and runs without errors.
Criteria Weight Developed (HD/DN)
Developing (CR/PS) Developed (FL)
Analysis is largely correct but Inappropriate statistical does not go beyond replication method is performed, or of what was covered in the correct method is tutorial. Limited model performed but with major specifications and diagnostics errors invalidating the are explored or are presented results; large sections of without demonstrating an code does not run or has understanding of why they are unclear purpose; performed. Some code has appropriate diagnostic unclear purpose or produces tests are not performed, an error. are misinterpreted.
Assignment Project Exam HelpCorrect output is not
presented (or all output is
Analysis output is presented
important features.
interpreted correctly and highlighting the important
Interpretation lacks nuance or
thoroughly. Implications follow statistics). Interpretation is
Interpretation 30% has minor inaccuracies or
misinterpretations.
nuanced, and demonstrate entirely; implications are
Implications, strengths and
appreciation for the strengths and absent or do not follow
limitations are discussed at a
limitations of the analysis. from the analysis.
superficial level.
Strengths and limitations are inaccurate or absent.
Report is adequately Report is poorly written presented, with only a small and presented, with typos Report is well-written and number of typographical or and grammatical errors. professionally formatted; code is grammatical errors. Tables and Tables/figures are not rigorously documented; tables and Figures are interpretable but presented or are difficult to figures are presented to a lack some details to reach interpret, with insufficient
Presentation 10%
publication-quality, with clear titles, publication standard, such as headings/ labels/ legends labelling and legends; references inadequate labelling, legends etc; code is poorly are appropriate and presented or scaling. References are commented; references consistently. present but have minor errors. are absent; or not Code commenting is present presented in a consistent but not detailed. academic format.
Note:
HD=Higher Distinction (85-100%); DN=Distinction (75-84%); CR=Credit (65-74%); PS=Pass (50-64%); FL=Fail (0-50%)
