Lying with Statistics

When I saw the results of the SU survey I was shocked, but when I started looking closer I realized that a lot of their infographics weren't telling the whole story or were misleading. Ask anyone who has taken more than one (or even just one) statistics course how easy it is to manipulate raw data, especially if you are breaking down your total data set by sub-groups where you have less than 100 respondents. So, with just two semesters of statistics, let's break down some of the mistakes they made.

1. Not a random sample
The survey was pushed out to student groups via social media. Given that some people are VERY passionate, it is more than likely that only people who care an extreme amount even filled out the survey. Additionally, it was pushed out to the abolition groupme, but I do not know a single person in a Greek org that found the survey through their chapter.

2. Relies on honesty of participants, ignoring motivations when taking the survey.
For all I know, any number of other people could've filled out a survey using my ID number and they wouldn't know which one was really me. So while, yes, there were 5 surveys which all had valid ID numbers, they have no idea which one was real and which were double entries.

3. Limited sample size
For the pie charts regarding opinion broken down by identity, A) there is no way to verify that someone is the identity they claim and B) we don't know how large the sample sizes are for those, and if they are less than 100 participants then the sample size is too small and it is generally agreed upon that it is not a valid sample.

Long story short: you can lie with stats just like you can lie with anything else. Be careful when looking at statistics and think critically about what the underlying data might be.