The Art Of Statistics Book Summary
A nontechnical intro to the wide world of statistics and how we can better understand and incorporate it into our lives. How can stats help us to answer questions about our own lives and even tell better stories, we can also use stats to examine our own cognitive biases.
The Art Of Statistics – Book Summary Notes
- With a rising trend towards the use of big data and the abundance of easily accessible data to back just about any point, it has become increasingly important to understand what that data means.
- Graphics, media and various other daily influences all use data to further their points or causes. Do you know how to critically analyze if what they are presenting is actually correct?
A 5 Step Process
- The job of a statistician is to use and interpret data through a 5 step process; problem, plan, data, analysis, and conclusion.
- The first step is to figure out what is the problem that you’re seeking an answer to.
- The second, plan, is how are we going to use data to solve this? What data do we need?
- The third step is to collect the data you’ve identified previously. This can include gathering it from other sources if it is already available, or going out and obtaining it ourselves.
- The fourth stage is best served by the use of data analysis software, the data can be sorted, grouped and even displayed across graphs or other graphical representations for ease of understanding.
- The fifth and final stage is the conclusion. What can we learn from the data we collected? Did it prove or disprove what we set out to achieve? How could we do better next time?
Data and Biases
- Data and data collection are subject to the same unconscious biases as many other facets of life.
- It can start as early as the first step of data collection. When we require someone to make a decision on what the aim or objective we are looking to answer is.
- Did the definition of what you’re seeking to measure become altered partway through the data gathering? Sometimes it can happen, especially if the collection isn’t happening in a controlled environment or by the person/people who decided on the aim, to begin with.
- The design of appropriate questions for the collection of relevant data is another tricky hurdle. The language used in the questions themselves can often impact how the subject will answer. Sometimes the nature of the questions as well will lead people to tend to answer what they think you’re looking to hear or what ‘should’ be deemed as the more socially acceptable answer.
- If a study has preset answers for its questions that is another way that biases can enter the data collection. By trying to make the process easier to collate and handle at the end you possibly take away from the choices of the subjects to give you answers you were not expecting.
The Presentation of Data
- How a set of data is presented can be incredibly important. With the right set of data and the willingness to eliminate or ignore other data or ‘outliers’, it’s very easy and possible to create the illusion that the data shows exactly what you want it to.
- Even elements of the display of data can impact how it is interpreted. From color to size and font, even the language that is used in the presentation all impact on the message that the viewer will take away from it.
- Pressure on researchers can be another potential downfall. If a researcher needs to present a certain number of papers each year, it could lead to them trying to reframe data into something that it was never meant to be. Or to make slight changes or alterations to the data in order to give them something to report.
- We must all also be aware that the media tends to favor telling a certain story with the use of data, rather than focusing on what the data actually means.
- Exaggeration and extreme views tend to sell more magazines. Something to be aware of.
