Discuss how you have seen statistics used to mislead in your work or life. How do you (or would you) now critically analyze statistical evidence rather than accepting it as true and representative?
Statistics are always used to mislead. They make people feel safe. We trust numbers. What people don't realize is that those numbers are manipulated by their very nature. Our trust in them is highly misplaced. They are used to convince people when in reality, they don’t provide complete proof.
Apparently, the manipulation of statistics is nothing new:
"Figures often beguile me, particularly when I have the arranging of them myself; in which case the remark attributed to Disraeli would often apply with justice and force: 'There are three kinds of lies: lies, damned lies and statistics.' " -Mark Twain
It is very easy to manipulate statistics to suit purpose. For instance, in the U.S., it may be stated on the news that the number of people reported as unemployed has dropped from a previous month. While this may be true, it is misleading to listeners because all the facts have not been included, facts such as the number of people who have just given up on looking for jobs has increased and those who are yet entitled to draw unemployment benefits has decreased because they have already drawn for their allotted number of quarters. So, in truth, those who are without jobs are greater in number from last month, not less. Thus, the economic situation of the population is worse, not better.
The problem with statistic is that even if you have the numbers, it is very difficult to establish a correlation. In a world as complex as ours, it is very difficult to consider all the variables. The jobs statistics that was mentioned above is a very good example. Yes, we got more jobs, but most of these jobs were seasonal. With the holidays now gone, these people will be laid off. Another example on this topic was when the government was so happy to have more jobs during the census, but these census jobs were completely temporary.
Every day if we read the papers and listen to politicians we are treated to a number of ways in which figures are used to willfully mislead and deceive. Clearly examples would be the way that politicans from opposing parties each have their own set of statistics that they used to suggest their own reality of events.
I remember one grading period in which most of my middle school students decided to take the term off. Sadly, I gave many "F" grades that term--all of them deserved and averaged accurately. I hoped that it would shock the kids back into reality, and it did. However, my principal was highly upset, since my number of failing grades was well above the number of other teachers in the school. I was criticized for poor teaching, failing to motivate the students, and for grandiose expectations of my students.
Statistics are often cited to indict public education, or, at times, to trumpet gains made. In my state, schools that make "expected growth" toward state standards are often praised in the media, while moderately high-performing schools that miles ahead of these schools in achievement often receive no recognition. Unemployment figures, as #3 says, have been grossly distorted by both sides in recent political debates, as have predictions about the federal debt.
I have often seen statistics used to mislead in political arguments. For example, politicians often claim to have "created" a certain number of jobs, while challengers say that the claim is false. It's difficult for a member of the public to know who is telling the truth. Thanks to the internet, however, it is easier to explore such issues than it ever used to be. There are also fact-checking sites that now make such exploration easier than before, although some claim that they are biased!
This depends on the statistic. When it comes to survey data, my typical way to critically analyze it is to wonder exactly how (and find out if I can) the questions were worded. This can make a huge difference in the outcome of surveys.
Outside of that, I tend to want to see good definitions of the variables being studied. I also want to know about who has funded the survey to see if I should worry that the statistics will be skewed in some particular way.