Data Analyses:Writing Reports About Your Results

Marta ShoreStat 4893W

Fall 2013

You've finished analyzing your data in R:

Preliminary data analysis Inference choice Assumptions Hypothesis tests Additional statistics Stuff you didn't really need to do

Now you can just write it up, right?

Not if you want your results to be understood by others.

Before you start writing, you need to: determine your audience,

organize your thoughts, prepare your results, and

write so that your audience can easily understand your work.

Steps from analysis to presentation

Determine audience Create a Technical Report Prepare any graphs or tables Write your report

Step 1: decide who the audience is for your report

What do they need to know? (Sims, pp. 423-6) What is their interest?

Other statisticians: process and result Customer/boss: result

Are they more interested in significance or magnitude of result?

GIVE BOTH Emphasize what is wanted

Step 1: decide who the audience is for your report

How much statistics do they know?

How technical can you be? (Montgomery pp 38-39)

Maintain same technical level throughout

How much will you need to explain about the procedures used?

Depends on knowledge level

Even if don't know stats, be respectful and assume intelligence (Weisberg, Slide 22)

Step 2: Create a Technical Report

(The Technical Report concept outlined on the next four slides is from Foster and Wild, 2010, pp.95-100)

A technical report is a report you write for yourself to summarize the results of your analysis. There are three parts:

Exploratory Analysis

Checking Assumptions

Statistical Inference

Step 2:Technical ReportExploratory Analysis

Using short sentences, review: Each graph for overall trends Any notable outliers or deviations Numerical summaries and what they tell

you.

Remember, this is for your benefit, so use whatever terminology and language helps you!

Step 2: Technical ReportChecking Assumptions

Look at the questions of interest and determine:

What hypothesis testing you want to do What are the conditions for performing the

test? Does your data meet those conditions?

Are there additional steps you'll need to take?

Step 2: Technical ReportStatistical Inference

Look at the hypothesis test you did. What was the result? Is there significance? What is the magnitude of the significance

and where does the significance lie? Confidence Intervals

R2

Residuals

Interaction graph

Step 3: Deciding on Graphs

After the technical report,figure out what graphs and what data were essential to your analysis.

Go back to R and make the graphs pretty. See document in moodle for how

Make tables accessible (Weisberg, slide17) Round numbers to 2-3 significant figures

Use borders sparingly and strategically

Step 3: Deciding on Graphs

You don't have to present everything as a figure. Use figures strategically (Zeleznik, pp 103-104)

Use those figures that easily illustrate a point.

Present only one type of information in each figure

Refer to and describe the figure in the document.

Use the same words and terms in the document that are used in the figure.

Once you've summarized your data with a technical report and chosen graphs, you can

begin creating your report.

But when you start, the focus shifts from

understanding your datato

writing for your audience

Step 4: Writing a Report

Your main goal is to make a document that is easy to read.

The reader can, at a glance, see the overall structure of the report.

The document is consistent in formatting and style.

Step 4: Writing a Report

It is good to create an overall look, or format, for your document before you start writing.

How are you going to divide up the report?

How are you going to format your text?

What titles are you going to give to each section?

Where do you want to place any graphs or tables?

Step 4: Writing a Report

Good reports differ from essays in that the writing should be broken into smaller, easily identified sections. (Sims, pp. 221)

Why? Smaller sections (with titles) help the

reader quickly see the overall structure.

Readers with limited time can find the section that is most important to them.

The writer can stay concise and on point!

Step 4: Writing a Report

Plan how you will format your text: Use one serif font for all text in the report.

(Oehlert, 2007)

Serif font has embellishments on the ends of the letters.

The emphasis is on the sentences and concepts.

Some common serif fonts:

Times New Roman, Georgia

Step 4: Writing a Report

Formatting your text (continued): Keep the font size consistent and between 10

and 12 points. (Sims, p. 243)

Write complete paragraphs and use single or 1.5 spacing.

Use italics or boldface for emphasis, but do so sparingly. (Zeleznik, p. 120)

The paragraphs can be aligned to the left, or indented up to an inch.

Step 4: Writing a Report

Every section (and subsection) should have a title. Titles should be:

In the same sans serif font (Oehlert, 2007).

Sans serif fonts do not have embellishments

The emphasis is on the letters and words

The most common sans serif fonts:

Arial, Helvetica, Veranda Choose fonts that don't distract:

Comic sans

Step 4: Writing a Report

Title guidelines: continued (Sims pp. 235-6) Make the font size larger than the text and

consistent. (20% bigger)

Separate the titles from the text.

At least two blank lines above

Only one blank line below Use boldface, underlining, size, or indentation to

delineate different levels of titles.

Align main section titles to the left.

Step 4: Writing a Report

After planning the format of the texts and the titles, I recommend making an outline:

Think about each section's content Write each section's title

And then, finally, you can write!

Step 4: Writing a Report

Rules for writing: Decide on the verb tense you want to use, and be

consistent throughout the document.

Decide on the technical level you want to use, and be consistent throughout the document.

Write simply.

Separate mathematical formulas from the text.

Explain what your conclusions mean in the context of the research!

Writing Your Report

For your data analysis report: Your audience knows basic statistical terms

but wants to know the results in words, not numbers.

Treat each paragraph as a section Give it a title

You must discuss any graphs or tables used.

Graphs and tables need to be at the end.

Summary

Do your analysis in R Figure out who your audience is Write a technical report about your results Prepare any graphs or tables Write your report:

Format the report so the structure is apparent

Be consistent in tone and verb tense

Write for your audience

References

1.Forster, M and Wild, C.J. (2010).Writing about Findings: Integrating Teaching and Assessment, in P. Bidgood, N. Hunt and F Jolliffe (Eds). Assessment Methods in Statistical Education:An International Perspective (pp. 87-102). Chichester, UK: John Wiley and Sons, Ltd.2.Montgomery, S. L. (2003). The Chicago Guide to Communicating Science. Chicago, IL: The University of Chicago Press.3.Oehlert, G. (2007). Lecture notes from Statistics 8801: Statistical Consulting4.Sims, B. K. (2003). Technical Communication for Readers and Writers, 2nd ed. Boston, MA: Houghton Mifflin Co.5.Zeleznik, J. M., Benson,P.J. & Burnett, R.E. (1999).Technical Writing: What It Is & How To Do It. New York, NY: Learning Express.6.Weisberg, S. (2013, September). How to talk about statistics... Lecture conducted for Stat 4893W at University of Minnesota Twin Cities, Minneapolis, MN

Slide 1Slide 2Slide 3Slide 4Slide 5Slide 6Slide 7Slide 8Slide 9Slide 10Slide 11Slide 12Slide 13Slide 14Slide 15Slide 16Slide 17Slide 18Slide 19Slide 20Slide 21Slide 22Slide 23Slide 24Slide 25