Close Home Forum Sign up / Log in

Would this be an unbalanced design?

P

======= Date Modified 02 18 2010 09:18:25 =======
Sorry for another design related question, but its hopefully a relatively straight forward one... :$

I'm reading a paper which has 140 babies as the sample. There are two test conditions - one were face stimuli is presented to the baby and one were they are presented with a mechanical object. The researchers are just interested in differences between which stimuli the female babies prefer and which the males prefer. All straight forward enough. However there are 90 girls and 50 boys, does this mean its an unbalanced design?

I know an unbalanced design is one where there are unequal participants in each treatment group but I can't decide whether this applies to this experiment or not.

I've got two ideas:

1.) Its repeated measures, so if the treatment groups are just 'face' and 'object' then there is the same amount of participants in each treatment group as each participant takes part in both conditions. So it is a balanced design??

or

2). The treatment groups would actually be defined as 'Girls-Face, Girls-Object, Boys-Face, Boys-Object' in which case there is not the same amount of participants in each group, so it is an unbalanced design??

Argh, does that make any sense? If someone could shed some light it would be much appreciated :-)

I'd say option 2, because the researchers are actually using gender to analyse the data, rather than just ignoring it.

Although, I recently found a load of stuff saying that to be 'truely unbalanced' it would have to be more like one sample being 2.5 times larger than the other. I guess it depends how big your samples are too though as to whether it affects the analysis.

P

2 I think. It seems unbalanced to me if gender is a major factor rather than just a sort of aside to the main aim.

The difference might not be too much but depending on the test (and version of SPSS they use) analysing factors that are unbalanced like that can throw off the accuracy of the data. SPSS sometimes tries to make the numbers equal by averaging them out before it even starts anything else. Which isn't much use.

P

Thank you both, that's a relief because I had been leaning towards answer 2 so its nice to know I'm on the right track :)

C

I'd say it depends what the analysis is. If they directly compare boys and girls it seems unbalanced but if they simply do the analysis twice, comparing faces and objects for girls, then faces and objects for boys it's balanced.

16579