Why Students Hate Statistics?

Most social science students hate statistics. There is a lot of fear and anxiety attached to statistics. Many of us became social scientists, and not ‘real’ scientists, because we sucked at mathematics. We thought by choosing psychology or sociology, we were never going to have to deal with mathematics again. Then on our first day at university we learn that we have to take statistics. It’s compulsory and if we don’t pass the statistics course, we don’t graduate! Say What?!!

Students hate statistics because they think that its mathematics. Yes, on one level it is mathematics. However, learning social science statistics nowadays is not like learning mathematics. Not in the way you did it in high school. Social science statistics are more grounded in real life. It is really about learning how to figure whether your research findings are important or not. For example, is gender related to driving behaviour?

Also, nowadays computer software like SPSS makes it almost unnecessary to learn statistical formulae. What you need to learn is the conceptual basis for deciding which statistical test to run for the data you have. This simply means then that you need to learn about how to design good research studies. Learning research methodology, takes you away from maths and into the realm of learning about concepts. Learning about concepts is easy. All social science students love learning about concepts (e.g. attribution bias or cognitive dissonance). Figuring out what type of study you are about to conduct is the basis on which you figure out which statistical analysis to perform. After that you go the software and click on buttons!

This is the first thing that social science students should be told, on the first day of their statistics class. Once they relax and understand what is needed, they may realise that they don’t hate statistics after all. In fact, they love it! Okay, maybe that’s overstating it a bit…… 

Pakistani Cricket: It's Not Rocket Stats!

As the evidence keeps on mounting against the Pakistani cricket players, I am forced to look back at the week in a slightly bemused way. Before the much stronger evidence presented this weekend (05/09/2010), the scandal was everywhere in the newspapers and television. At week ago, all we knew was that the alleged match fixer had made three predictions that had turned out to be correct. This seemed to surprise people as highly unlikely. But how unlikely is it? 

I will not bore you with the numbers, but I am sure there is some person out there who has done spot-betting and won three times in a row, for purely random reasons. What we should have asked was; how many other predictions did the alleged match fixer make? And how many of those predictions where correct as well? If he had made 20 predictions and three were correct, this would have caused a totally different reaction from us. 

Imagine that you are rolling dice. How likely is for you to get three 6's in row. Very unlikely! Okay, but how likely is it for you to get three 6's in total if you roll twenty times? Aha, that changes everything! You can easily see yourself getting three 6's if you roll the dice 20 times. 

So you see why the questions that should have been asked by the reporters should have focused on how many other predictions were made and how many of those were also correct. The answer to these questions would have provided us with more insight into the potential guilt of the named parties. Instead, we spent the whole week being 'impressed' by the three predictions from an apparently edited video. It’s not really rocket stats, it just takes a moment of reflection! 

What Am I Being Tortured With Statistics?!

A lot of people hate learning about statistics. Most students often wonder why they are being tortured in this way by their professors. This is especially true for students studying for degrees that don’t seem obviously related to statistics. If you are an economics student, it seems obvious that you need to know something about analysing data and trends. But for the rest of us the question is: What has all this stats got to do with psychology, law, sociology, anthropology or even politics?

Well, let me tell you about my friends (who will remain nameless if we are to remain friends). I love hanging out with these guys, they make me laugh. But, what drives me crazy are their debates. Each person always takes up a strong position about some issue. Furthermore, each person is convinced that their position is the “truth”.

For example, the other day they got into a big debate about parliaments, and whether groups make more balanced decisions than individuals. They went back and forth about it for hours! Geeks!

The other time they had an argument about whether people with children have happier marriages than people without children. Then they had this one argument about whether more choices are better for people trying to make decisions than fewer choices. What is interesting about their debates is that they often reflect each person’s political/social biases. Other times they are just talking rubbish to while up time.

But it’s important to note that these debates often appear in serious political debates within the media. For example, there are debates around the value of juries, the sanctity of marriage or freedom of choice. I am going to assume that you are studying to learn or discover the true nature of things (in as much as that is possible). I am going to assume that your goal is not to become a charlatan. If my assumptions are correct, then I am afraid to tell you that you need to learn statistics. Why? Because statistics gives you the tools to be able to contribute to, and even answer some of the questions above with some level of scientific rigour.

Let’s look at each of these examples above one at a time:  

•  Do groups really make more balanced decisions than individuals? Well, that’s what people thought until research by Stoner (1961) on group polarization showed that groups actually make more extreme decisions than individuals!

• Are people with children happier in their marriages? Well, contrary to popular belief, marital satisfaction actually increases as children leave the nest! (Gilbert, 2006).

• Finally, are more choices better for people than fewer choices? Well, research by Schwartz (2004) on the Paradox of Choice shows that people are actually less satisfied with what they choose, if they have chosen it from a larger set of options, compared to a limited set of options!                                                      

So now you may wonder; are these researchers’ geniuses? May be they are, but that is totally irrelevant. They were just able to answer these questions and provide useful data because they learned how to use statistics. Their research questions are very fun and engaging. However, unlike my drunken friends, or your drunken friends, their contributions to the debate are more meaningful than just “having an opinion” or speaking from “your personal experience”. This is the reason why we teach statistics. It’s not to torture students, but to show them how to use an essential tool.

Besides, it’s not really rockets stats. Statistical concepts are not that difficult to learn these days. Most data analysis is easily done with computer software.  It’s important not to think about it as learning maths, but rather as learning how to use powerful tools to find answers to interesting social questions.

I am often tempted to step into the debates my friends have and say something like; “Well guys, research actually shows that….blah blah blah”. But my friends hate that! So these days, I just grit my teeth and watch. At least, I have this blog as an outlet!