report writing代写、代做R编程语言、代写data留学生、R设计代做

日期：2019-02-08 10:22

Guidance for report writing

Presentation and communication

Writing:

1. Use past tense.

2. Use short but complete sentences.

3. To keep paragraphs short, aim for one main idea in each paragraph and

check that the same idea isn’t repeated elsewhere in the report.

4. Any acronyms, abbreviations or uncommon terms should be defined first

time they appear in the report (consider including a glossary).

Tables and figures:

1. Each table/figure should be numbered so it can be referenced in the text.

2. Each table/figure should have a clear caption to state its purpose.

3. Graphs should have labelled axes and, where appropriate, a legend.

4. Tables should be nicely formatted (not copied directly from R or Excel)

and should use sensible numbers of decimal places.

Abstract

1. State the primary research question and explain briefly why it is of interest.

2. Very briefly describe the data used.

3. State the main statistical method used.

4. State the main results of the analysis, along with relevant statistics.

5. Using simple language, state your conclusion and its practical significance.

Do not include references here. Do not use undefined abbreviations. Do not

mention any results that are not included in the report.

Introduction

The introduction should, as far as possible, be written in simple language that

would be understandable to a non-statistician.

1. State your specific objectives or hypotheses very clearly.

2. Explain why your objectives/hypotheses are of interest and give some

details of the background and context.

3. Describe the data, explaining when/why/how it was originally collected.

4. State which statistical methods you used to address these objectives/hypotheses.

5. In the final paragraph, outline the structure of the report.

Do not give too much detail of the methods or the data set (this should be done

in the Methods section). Do not describe any data that was not used.

1

Methods

The data:

1. Define the variables used, including any units and abbreviations.

2. For categorical data, state the possible categories.

3. Explain how and why you derived any new variables.

4. Explain how outliers or missing data were handled (e.g. imputation).

5. Say how any training/test data sets were obtained and state their sizes.

Do not list every variable in the data set if not all of them were used/relevant.

Do not paste R summaries of the data into the report. Do not present statistics

calculated from the data here, unless necessary to justify the methods.

The methods:

1. Explain the purpose of each statistical method and briefly justify its use.

2. State any assumptions made (e.g. normality of residuals) and how they

were checked (e.g. quantile plots).

3. State the significance level used for hypothesis testing and confidence intervals,

or any other relevant thresholds.

4. State which variables were used in each analysis (e.g. the response and

covariates in a regression analysis).

5. Cite any R packages used and say what they were used for.

Do not state results here. Do not describe methods that you did not use. Do not

state your methods in an ambiguous way that could not reasonably be reproduced

by the reader (e.g. for a mixed effect model, make clear which parameters

were fixed or random). Do not paste R code unless absolutely necessary.

Results

Data:

1. State both the number of observations in the original data set and the

number of observations used in the final analysis.

2. Present simple descriptive statistics (e.g. means and standard deviations)

to summarise relevant variables, preferably in a single table.

3. If appropriate, present graphs or numerical summaries describing relationships

between variables (e.g. correlations or scatter plots).

4. Summarise the amount of missing data for each relevant variable (preferably

within the table) and give reasons (if known).

5. Say how many observations required imputation before the analyses.

2

Analyses:

1. Present all relevant estimates, preferably in a table.

2. Where possible, estimates should always be accompanied by a measure of

their precision (e.g. confidence intervals),

3. Discuss any checks of your model’s assumptions, then state whether any

assumptions were violated and how this was dealt with.

4. Discuss and explain any model comparison, selection or validation.

5. For any important hypothesis tests, clearly state the result of the test

(accept or reject the null) and, if possible, present a p-value.

Do not repeat details that are already clear from the Methods section. Do not

give an in-depth interpretation of the results or limitations (this should be done

in the Conclusions section). Do not give a table or graph without referring to

it in the text. Do not copy and paste R output. Do not repeat statistics from

the tables in the text.

Conclusions

In simple language, understandable to a non-statistician, summarise and cautiously

interpret your results, with reference the objectives outlined in the introduction.

Make clear whether/how your results have answered your original

research question. Try to answer each of the following:

1. What are the practical implications of these results, and are they consistent

with existing beliefs (e.g. in previous literature)?

2. Are there any known limitations of your analysis (e.g. potential biases or

imprecision) and how might they have affected your results (e.g. underestimation

or overestimation of parameters)?

3. Are your conclusions generalisable and likely to be valid outside the context

of your current data set?

4. If you could do the analysis again or if you could collect new data to

perform a new analysis, what might you do to address the limitations or

improve generalisability?

5. Does your analysis raise any questions for future research?

Do not just repeat the statistical results from the previous section. Do not

make any claims unsupported by the results of your analysis. Do not state your

conclusions too strongly (e.g. ‘this analysis has proved...’).

3

Bibliography

1. List references in the order that you have cited them in the report.

2. Use a consistent referencing style throughout (e.g. Harvard).

3. Cite any notable R packages used, for example by using the citation()

function.

4. Only include papers/packages that you actually refer to in the report.

Appendix

1. Include your final R code, with explanatory comments.

2. State any seeds used for random number generation.

3. If necessary, include results and explanations of any additional analyses

that you could not fit into the Results section.