You may be asked to write information that involves calculations or statistics. Check and recheck your math.
Consider this slogan on a bottle of grape soda:
Two liters, 50% More than One Liter!
Although it’s easy to shake our heads and lament the prevalence of such “marketing math” in our society, these errors are easy to make.
Providing recipes for solutions. When you give the volumes for components in a solution, make sure that they add up to the final volume you provide. Ensure that a person who is trying to replicate your experiment can make the solutions simply by following your provided recipe. For instance, when you provide a pH, is it clear whether or not it is for the entire solution or just one component of the solution? If you have performed a serial dilution, will a reader know what concentration you started with, what concentration you ended with and how many individual dilutions you made?
Check and recheck units. If you have not heard about the tragedy of the Mars Climate Orbiter, one team used imperial units and another metric units, the end result was that the orbiter burned up in the Mars atmosphere. Although most of us will not be writing a design protocol or navigation maneuvers for a Mars orbiter, if you are writing patient education material on a subject like nutrition, the imperial v. metric unit issue is important.
It’s all relative. Many data, such as fluorescence and luminescence readings, are reported as “relative light units”. These data for experimental values need to be normalized to background, negative and/or positive controls to have any meaning. If you are comparing two separate experiments, you may need to talk about percent increase or decrease above or below a control used in both experiments.
- Are your data normally distributed? If so, you can talk about mean and standard deviation. If not, mean and standard deviation from the mean are not appropriate calculations to use.
- When you decide to represent data graphically, make sure that you know whether your data are continuous or categorical. Do not plot categorical data as a line. If you decide to break continuous data into categories, make sure you have a rationale for the categories, and make sure you include your rationale in your figure legend or results section.
- When plotting a linear regression analysis, be sure that the line is generated from the equation and is not plotted point-to-point.
- Ask yourself whether a graphical depiction of the data is really necessary. Does your graph point out a trend in the data? If there is no trend or pattern to illustrate, would your data be just as meaningfully displayed in a table or talked about in the text?
- When you create a figure, can the reader get the take-home message by reading the figure legend and looking at the figure? Or, does the reader need to read your discussion section three times over to understand your figure? Don’t make your readers work to hard to understand your data, especially editors, reviewers and dissertation committee members. Also, the more convoluted your presentation of the data is, the more likely you are to make a mistake. Lang and Secic have produced a great handbook for people who must report statistical analysis as part of their work, How to Report Statistics in Medicine. The book is an easy read and a great tool for writers and editors.
Finally, be consistent. If you give a ratio as 1:10, use the same nomenclature throughout the document. Do not change from 1:10 to “1 in 10” to 1/10, to 0.1, etc. You will only confuse your reader.
These are only a few tips for reporting data and numbers in your science writing, but they are a start clear presentation of data for your audience.
Latest posts by Michele Arduengo (see all)
- Questions of Genome Privacy and Protection - June 15, 2018
- Catalyzing Solutions with Synthetic Biology - May 17, 2018
- Orchestrating the Genome: Final Thoughts for #HumanGenomeMonth - April 30, 2018