Statistical Backgrounds…beneficial?

26 09 2011

Now I’m not entiirely sure what was meant by a statistical background, but in this blog I’m going to assume it means statistics from previous studies.

Straight away, we can see that having a strong statistical background will put us in a strong position. It would make sense to carry out a study expecting to find a certain result, if psychologists before us have found the same thing in their own studies. For example, last year we had to do a study testing if there was such a thing as a congeniality effect. We had quite a strong statistical background, as we found a study by Levine and Murphy (http://psycnet.apa.org/?&fa=main.doiLanding&doi=10.1037/h0062586) showing that there was indeed a congeniality effect.

Although it would indeed seem very strange to conduct a study trying to find opposite results to a previous similar study, there is a chance that the previous studys results aren’t valid. No one’s perfect, not even psychologists, and it has been said it is almost impossible to prove anything. In fact, some studies are carried out to DISPROVE other studies, such as Samuel and Bryant’s study on conservation in children (http://onlinelibrary.wiley.com/doi/10.1111/j.1469-7610.1984.tb00152.x/abstract) which tried to disprove Piaget’s theory about the age at which children can conserve.

In conclusion, I think it does help to have a strong statistical background for a study, but it doesn’t mean everything. Even when we conducted our study on congeniality, we found a previous meta-analysis by Eagly (http://cdp.sagepub.com/content/10/1/5) that showed no such thing as a congeniality effect.

Advertisements

Actions

Information

6 responses

7 10 2011
psucfa

In terms of previous studies that you mentioned, I think that a strong statistical background (I’m going to use this term to describe a good knowledge of statistics) would be important at looking into the raw data collected from said study and seeing if you come up with the same data or something different.
Of course, a strong statistical background can be useful in other aspects of life aswell can it not? For instance in opening new job possibilities in one’s career?

7 10 2011
sinesofmadness

I see that you have taken the term statistical background differently to me. You make a very valid point. The very fact that previous studies have been disproven shows how important a statistical background really is. Without a basic knowledge of statistic (never mind a strong one) you would be very susceptible to dishonest researchers using statistics incorrectly.

6 10 2011
psuc6f

I would have to agree with psud77 in that the appealing nature of manipulating variables when conducting a similar study would prove to be tempting in some cases, and statistics can easily be manipulated to show what the researcher set out to prove / disprove. However, having a strong statistical background would allow other researchers to examine the statistics and understand what they truly imply.

6 10 2011
Psucc0

Interesting ideas, Its great how even wrong statistics can inspire psychologists to conduct further research to further hone the theory, which i guess is similar to what we did with the congeniality project. It would seem kinda boring if everyone just took scientists word for it, imagine what a mess the world would be!
Also you made a valid comment on my blog that just to get a good head start in University we needed to at least have a basic understanding of stats which does seem to bring the importance home to us!

6 10 2011
prpdh

I can understand where your coming from when you say that it doesn’t seem worth conducting a study when it already has a statistical background, and that you did raise the issue that some studies have been found to have different results when repeated. But it is worth pointing out that perhaps in some investigations the results may have been conducted under very different circumstances to.
A good example of this is in the investigation we all did into the dreaded Congeniality Effect was primarily tested on university students using a topic (global warming) that we are all well versed it, which could be why the overall result was rather low. Which does agree with Eaglys meta-analysis but there are a large number of investigations which were conducted a while ago that would be difficult to replicate or just very impractical.

5 10 2011
psud77

You mention in your post the advantage of having previous data from other psychologists when carrying out an investigation, I agree that this is very useful in terms of research methods and designing an investigation. However I feel that this is not the same for statistical testing and analysis. It is of course necessary to build on and test previous investigations and theories but the temptation to bend statistics to meet our own ideas, or even those of someone else, is too great. The temptation would be to make the stats we see fit our theory whereas we should be adapting a theory to better suit the statistical findings we uncover.

Stats should feed the theory, not theory feed the facts.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s




%d bloggers like this: