Be skeptical of statistics
By Jason Deveau
Making better sense of data.
By Jason Deveau
Growers have to interpret data all the time. Agrichemical companies tell them why their products are more effective than their competitors, seed companies tell them why their crops have better yields, and production specialists tell them why spending more time and effort will pay off in the end.
[Editor’s Note: Drainage contractors also have to interpret data all the time – or help growers make sense of their own for drainage decisions. Deveau’s tips are useful if you want to get better at making sense of data.]
To separate fact from fiction, we need to understand a little about descriptive statistics. Statistics are math tools used to describe data, find trends in data against variation, determine if a sample represents a population and to draw conclusions about data. Here are some key points to look for the next time you’re shown a graph or table.
Beware the average! “Average” can mean three different things:
- The arithmetic average is called the mean.
- The data point that occurs the most, the mode.
- The centre of a distribution, the median.
They are all averages and are different from one another.
For example, if 10 people earn $50,000 a year and one earns $500,000, the mean income is about $91,000 a year. However, the mode and median are much closer to $50,000. All three are averages, but they paint very different pictures.
For an average to have meaning, you must also know the range or variability of the data. In our example, you should know incomes span from 50 to 500 thousand. Otherwise, the average can mislead you. Don’t trust graphs that do not show you the variability, such as standard error or standard deviation. If you’re able, ask the presenter to tell you the range – if they can’t, be suspicious.
Also look out for chart junk. The best graphs are simple line or bar graphs. Anything with fancy pictures is just distraction. Pies are for eating, not for graphs.
Finally, look out for disclaimers like: “This is not a scientific poll…”; “These results may not be representative of the population”;“…based on a list of those that responded”; and “Data showed a trend but was not statistically significant.”
Like any tool, descriptive statistics can be misused (intentionally or unintentionally). Maintain a healthy skepticism and question charts, tables and conclusions where insufficient information is provided.
Jason Deveau is an application technology specialist for the Ontario Ministry of Agriculture, Food and Rural Affairs (OMAFRA).