And also, M., as a former reporter, I was expected to fact-check every article I wrote to avoid lawsuits. We’re taught fact-checking in J-school. That many modern reporters and editors don’t do it is not a problem with journalism as a whole, it’s a problem with how sloppy journalism has become, ridiculously, acceptable.

Please don’t refer to the “Welsh, Scottish and Irish” as Brits. Particularly not the Irish, although I’m pretty sure a lot of Scottish people don’t really identify as “British” either.

We do: in comment 17 I provided a link to, and quotes from, an Amnesty International report about attitudes in Northern Ireland among university students.

Do you fell like this tiny slip has negatively affected the impact of these findings? I don’t understand your agenda here, why you are so eager to make this point, or what you think it will achieve?

Quite right. But I was reiterating that this was probably something related that may have made depresso act in the way they did. I’m not defending being rude, just trying to add some context… ‘UK’ is synonymous with England for many people, and that’s what’s annoying to a lot of people.

But saying “journalists don’t fact-check anymore” is inaccurate and offensive to all the ethical journalists out there.

Let me amend: the vast majority of journalists do not fact check. Especially the quality quintet, as they are called. Since this article came from The Times iirc, my point still stands.

That doesn’t mean I didn’t have a point. And the second survey was only given to college students in Northern Ireland, so those results can only be generalized to students in Northern Ireland from the college or colleges wherever the survey was conducted.

And there are always, always, always, always distinctions between and within populations. Always.

Actually, the results of this survey can never apply to people from Scotland or the US or Brunei or any other country besides England and Wales. Results of any study can only be generalized to the population from which the sample came and only if that sample is representative. Otherwise, you’re making assumptions based on nothing.

As if there is something about Scottish (or Northern Irish) culture that implies the results would be different if those areas were part of the survey.

There very well may be. There are limitations to every study. Until the group is studied, you have to assume that there are myriad differences between the group being studied and the group that was not studied. That’s the point of having controlled studies. That’s the point of having a large representative sample.

When a study about how abortion causes breast cancer comes out, we’re not pissed because they didn’t interview us, we’re pissed because it’s bad science. It’s not bad science to survey England and Wales but not Scotland or Northern Ireland – it’s just what you’ve decided is the scope of your survey.

And it’s bad science to make generalizations about all people based on the study of a few people of a particular group. You don’t have to survey Scotland and Northern Ireland. But you have to survey Scotland and Northern Ireland, as well as England and Wales, if you want to support the claim that 1 out of 7 people from the UK believe women deserve rape sometimes.

The issue of misinterpreting statistics is a much deeper one in society, and not specific to journalists. As I’m finding in the mindbogglingly hard graduate stats class I’m taking now, scientists misinterpret statistics more often than you want to think about–or don’t run the right analyses to start with. The average member of the public–and most journalists fall into that category when it comes to statistical knowledge–isn’t very good at interpreting statistics, which I think is a big problem with education, since being able to understand statistics (e.g. risk assessment) is pretty important to modern life.

I think a lot of that is an artifact that comes from how we use statistics in society, especially in the social “sciences.” Statistical analysis can be really useful when you’re looking at things that can be easily translated into numbers, but the further you get from the concrete and the objective the more trouble you run into. In most of the social “sciences” we use something called “null hypothesis significance testing. Now its not been long since I’ve had a graduate level statistics course, and I’m in the midst of a graduate level research methods course right now taught by an esteemed and multiply published professor. In both of these courses I’ve been taught that good statistics can tell you the probability of your hypothesis being correct given the data you have. Thats the conventional wisdom, its how we find things out in the social sciences. It’s also wrong. Jacob Cohen wrote an excellent article entitled “The Earth Is Round (P < 0.05)” that was somehow published in the American Psychologist in 1994. You can find it for free with a google search, good stuff if you’re interested in being able to critically read stats.

The issue of misinterpreting statistics is a much deeper one in society, and not specific to journalists. As I’m finding in the mindbogglingly hard graduate stats class I’m taking now, scientists misinterpret statistics more often than you want to think about–or don’t run the right analyses to start with. The average member of the public–and most journalists fall into that category when it comes to statistical knowledge–isn’t very good at interpreting statistics, which I think is a big problem with education, since being able to understand statistics (e.g. risk assessment) is pretty important to modern life.