Statistical Analyses used to be associated with the business industries back then. They have, however, become very important in the way humans live their life. Statistics are commonly used by the government to formulate policy bothering on their citizens. Without it, some sectors of the world will be nonfunctional.
The rise in globalized operations, as well as advancement in technology, has given businesses different insights into solving the extreme uncertainties that plague the market. Statistical analyses help to promote better decision-making based on facts and not assumptions or speculations.
Can Statistical Analyses Be Trusted?
To guarantee the validity and decision of analysis, the statistics must be reliable. Although numbers don’t lie, there are many instances where statistics cannot be trusted due to the way they are used to mislead the public.
An investigative survey conducted by a prominent scientist in 2009 at the University of Edinburgh disclosed that about 33.7 percent of respondents admitted having used questionable research methods. These methods include the modifications of results to improve outcomes, data interpretations that are subjective, holding back analytical details, and dropping observations due to their gut feelings.
Even the most reliable data gatekeepers are not protected against the carelessness and bias that is associated with data interpretation. The most common example of how stats are misused, thereby becoming unreliable, is course correlation vs. causation.
Let’s look at an example. The consumption of tea increases the risk of getting diabetes by 50 percent, and alopecia (baldness) increases the risk of cardiovascular disease by 70 percent. How can we trust this stat when it didn’t make any mention of the amount of sugar present in the tea?
When the Lie Occur
There are several situations in which statistical analyses lie. Some of these situations include the following:
The answers generated by a poll or survey is dependent on how the questions in that survey are structured. That will mean that the result obtained from such polls will be false. Some certain types of wordings are usually used to persuade respondents to give answers predictably.
Another way in which stats can lie is a situation where questions are asked using a conditional statement. The proper way to look at this is not to take stats too seriously. This is because the intended results of some of them have already been predetermined.
The issue with flawed correlation is that some of the variables will eventually correlate if you test and measure enough variables. Statistical analyses can be manipulated using sufficient data to prove a nonexistent correlation or one that is significantly not enough to prove a correlation.
This is a conscious attempt to manipulate data without pretending to be accountable. The results obtained in this way cannot in any way be true. Data omission and adjustments are the two common ways in which selective bias in statistical analyses is carried out.
The results and outcomes from the various statistical analyses flying around cannot be trusted wholly. Some of them were conducted to achieve some narcissistic ambitions.