Inferential tests to be used in IA, Perham & Vizard

[jolthan]

Hello guys, I'm replicating Perham & Vizard's (2010) study on the effect of sound on memory.

 

The method:

PowerPoint slides would be prepared as per the materials list, with each alphabet taking up one slide. A blank slide is inserted between every slide. A total of eight powerpoint slides would make up a single condition in which afterwards, participants would have 20 seconds to recall as many letters as possible in the same order.

 

Firstly, participants had 20 seconds to recall letters shown on the screen in the exact same order. However, a popular and “liked†song will be played in the background.  This recall is repeated with 4 different sets of 8 alphabets but with different background sounds; “disliked†music being played in the background, changing-state speech playing in the background, the steady-state speech, then lastly recall was done with no background music played, mimicking the effect of studying in silence.  

 

_________

 

I would count the number of alphabets that participants got correct for each condition and compare the results from the different conditions.

It is repeated measures, but I'm having trouble figuring out the data type.

 

My teacher is not doing a very good job explaining inferential stats.. so it would be awesome if someone could tell me the test i'm supposed to do and point me to relevant information.

 

Thanks!!!

[blindpet]

If it is repeated measures and counted alphabet letters you can use an unpaired t-test. Counts are interval/ratio data. This test compares means.

 

Here is a calculator link, treat group 1 and group 2 as the different conditions

 

http://www.graphpad.com/quickcalcs/ttest1.cfm

 

You could also use the Wilcoxon t-test because your sample size is small so you can justify its use. It is the non-parametric equivalent of the unpaired t-test. You use it for ordinal data. It compares medians.

 

If you have multiple conditions (more than 2) like it seems you do then the statistics to do this are more advanced, you can consider the Friedman. It compares more than 2 conditions that are repeated measures. It also compares medians.

[jolthan]

Hello Mike! Thank you for your thought-out answer, after calculating the inferential stats, what exactly do you do with it? 

It was not explained in-dept in class, except that the P > 0.05 should be used to either prove or disprove the null hypothesis. Is that all?

 

Thanks!!

[jolthan]

bump please