So, by now I’m guessing you have all lived through the exam? I hope it went ok. Although, to be fair, you probably haven’t had time to think about it because you’re hurriedly revising for the next lot. Good luck! You can dooooo it!
So some feedback from this last set of blogs and comments. Read some super interesting blogs this week – the power of water memory (!) thanks to psuc0e, the complex scope that psychology covers and how we begin to measure abstract concepts such as mind, love etc (props to blc25). Theundergradpsychologist wrote a great blog on psychology and the entertainment industry – how come they’re allowed to do things that ethically, we, as psychologists, can’t touch?
There was wild discussion and debate about correlation and causation – loads of folks were having a piece of that pie and some really interesting questions arose. Some peeps to attribute good questions to are: secretdiaryofapsychstudent, rgjblog, psud6d.
And of course all the commenters – if I mention you all, my blog will just end up being a list. Some things that stood out were the idea that IN NO WAY, OVER MY DEAD BODY, could correlation EVERRRRR imply causation. I see where this is coming from and I think that this is drilled into you from an early statistical age and it is clear to see why – as many of these blogs mention, results gained from correlational studies are often grossly misinterpreted. However, a little food for thought – what if the correlational-ness of a study is actually about the experimental design as opposed to just the plain old statistics? If I run a study and manipulate a variable… for example, I am measuring happiness and looking at the effects of chocolate, sunshine and red wine on happiness – so I get a bunch of peeps/participants together, give them different amounts of chocolate, sunshine and red wine and then measure their overall happiness on the Thandi’s Happiness Scale (Gilder, 2012) and then correlate the amount of chocolate (preferably Toblerone!) on their happiness – if I have carefully controlled extraneous variables etc and, experimentally, I am fairly certain that the amount of chocolate is the only thing that’s changed, if there is a strong positive correlation (which of course there would be, duh?!) this might suggest causality. However, if I run a correlational study, and get peeps/participants to report how much chocolate, sunshine and red wine they get and then correlate that to their happiness score – I’ll see a relationship but, due to the design of my study, I can’t tell anything about causality because I’m not controlling for anything… Everyone with me? Any noodles particularly baked? We can talk about this a bit more in class if you like. We touched on something fairly similar last semester when we looked at regression – we used independent variables as predictors and looked at whether they had an effect on the dependent variable (this was the BART balloons analysis we did if your mind is drawing blanks!) – for shits and giggles, we also split the IVs into high, medium and low and ran ANOVAs on the data to see what it showed us – not much different really but when you split your IVs into different groups like that, you lose a bunch of the variance… The moral of this tale, is that it is crucial to make sure you are doing the best analysis for your experimental design. Design and statistics are NOT mutually exclusive. Far from it…
While I’m still on the topic – I was asked a question about correlation that I would like to hear your opinions/thoughts on – we know that correlation doesn’t necessarily imply causality but does causality imply correlation? Answers, thoughts etc in the comments please! I think it’s a great question.
Moving swiftly along – some great blogs to look out for – psuc5d who talks about multiple comparisons and a dead fish (incidentally, also a blogger of the week! Whoop!), psucc6 who provides great insight into informed consent, psud00 writes a great blog about the placebo effect, psycho4stats discusses ‘approaching significance’, samstats talks about basic and applied research and how they are seen when it comes to funding, ksgs talks effect size, statsjamps differentiates null and alternate hypotheses, and psuc1a (other blogger of the week!) talks clinical equipoise which is a fascinating topic – particularly for those of you interested in pursuing clinical psychology. It’s a blog what makes you think and those are always good!
So team, another good session of blogs. Well done. Some tips – if, and I wouldn’t presume that anyone would EVER do this – you are taking information from Wikipedia, please try and put it in your own words… It is a dead giveaway when, in your text, you include the hyperlinks…
I’m still seeing blogs that are far too descriptive – I am looking for your thought and argument – as I said before, it’s great to use this as a tool to revise stuff you’re not sure of but you really need to add a critical element to your discussion.
My main comment about comments (see what I did there?!) is that they’re getting too long. They must be taking you guys aaaaaaaaages and while i’m a strong supporter of hard work and elbow grease, some comments are longer than the blogs themselves. There is not a strict word count for either the blogs or comments but, this is and should be good practice at writing to a word count (I know, I’m not one to talk, am babbling!) – you don’t have to write mini-essays for your comments – try to keep them as short as possible while still maintaining the good level you’ve got them at (i was well impressed with comments this week folks!).
See you next week! Good luck with the rest of them exams!
P.S. those of you not attending class, please do! Lectures and small groups – it’s where the magic happens! When I say magic, I mean you learn stuff. It’s radical!