Book Reflection on How to Lie With Statistics (Book Review Sample)

Book Reflection on How to Lie With Statistics
Name
Institution name
Course name and code
Instructors name
Due date
How to Lie With Statistics is a 67-year-old book that can be read in about one and half an hour or so. It contains more reasonable information that you can apply at your place of work than any other known 21st-century book. We all need to be summarized data to help us make informed decisions at our workplaces, school, and home.
On several occasions, I have seen complex mathematics concepts summarized into simple numbers. This aid the decision making: with few data present human can process at a fast rate. This is why reflections from "How to Lie with Statistics" (by Darell Huff) are still as relevant now as they were relevant about 50 years ago. Always we need to present a valid summary of the tables and graphs. When we use the generated data and summary we need to be always spot on to identify misleading or exaggerated statistics that will influence us to take any action that will benefit someone else at our expense.
The much-needed skills to scrutinize any statistics fall under the category known as "data literacy". Data literacy is defined as the ability to comprehend and be able to make an informed decision from the information presented in the statistics. Data literacy is not interesting at all but should form the core of science data education. Here are my well-learned and reflection from “How to Lie with Statistics" with the incorporation of my experience.
View Correlations with Skepticism
In most cases, we hurriedly jump to conclusion that A orb is likely caused by C or D from any given statistics. For instance, higher income is positively correlated with more years spent in college, we mistakenly conclude that more years in university will then lead to the accumulation of more wealth. However, we forget that there could be another third factor that leads to results such as parental income or willingness to work hard is behind more wealth and additional years in college. The hidden variable can often mislead us into making wrong conclusions about a particular casualty.
Sometimes we may see that two variables are correlated but in a real sense, they have nothing to do with each other. If more comparisons are drawn between data sets, more interesting relationships will be derived. To avoid falling into this trap of making conclusions hurriedly, approach each statistic with skepticism by looking for all the factors available. We always prefer a neat, casual narrative but in real perspective that is not what the data is usually telling us.
Relationships Don’t Last Forever
The relation of the data given is not permanent. This could be in a positive direction or negative direction. The relationships are limited to a particular region and any attempt to extrapolate will establish a different relationship from the previous.
Small Samples Produce Shocking Statistics
Every study conducted by the researchers has what is called samples. Samples are the subset of the population that is taken to represent the entire population during scientific research. This is only perfect when the sample size is large evenly distributed across the population. Sometimes due to lack of enough funding, behavioral, psychological, and response rates scientists are forced to use a small sample size the results produced are not accurate and no other study can produce the same results again.
To stop falling into this trap, look for the parameters in the data given. If the number of the observation is not indicated, assume that the given statistics are worthless, since the author has something to hide. Most of them tend to neglect the sample size according to the behavioral scientist.
Look at all the Numbers that Describe a Dataset
So as not to be fooled by the summarized data, you need to check the sample size. In the scenario, the sample class is not given it's prudent to check all the parameters that explain the given data. Another trick that is often used to mislead viewers is to intentionally avoid listing parameters that explain a data set such as the mean, mode, standard deviation, interquartile range, and many others. This is used to dive deep into the data which might be against the author of the data. Summary of the statistics does not tell the whole story, rather these simple small details that matter a lot in any given statistics.
Use Comparisons to a Common Baseline
When using any given statistics, the most important question is to ask how the previous value compares with the current value, rather than asking what is the value. The main perspective of this concept is to ensure that the definition does not change and not only are interested in comparing the previous values and numbers to the current values.
Look for Bias in Sample Selection
This always rises from the measurement used: taking the number of people using the I phone brand will tend to favor only the rich. It could also arise from the physical location: taking data of those who live in slums will be convenient but the results will be biased results toward more progressive views from other parts. Also checking sampling bias from the source of information. We often only like to choose the sources from materials tha