Disadvantage of indepedent groups

[The Detective]

Okay this is a fairly simple question but in the exam if it asks for a disadvantage I would put more participants required....
but practice exam answers are usually about the difficulty controlling participant extraneous variables

I know both are correct but would examiners more likely give the mark for the latter response?
Thanks

[sillysmile]

yep, when compared to repeated measures, that is one disadvantage..
but the differences in participant characteristics  between the groups is a more obvious disadvantage..
it's  probably a better disadvantage to mention.

[jinny1]

how small do the sample need to be for it to be deemed not repreesentative

[sillysmile]

how small do the sample need to be for it to be deemed not repreesentative

I have heard that 30 is considered the smallest size that can be representative.
but I'm not completely sure.

[Russ]

It's worked out statistically, but I don't think you need to know it. Have they changed Psych at all in the last few years?

[sillysmile]

It's worked out statistically, but I don't think you need to know it. Have they changed Psych at all in the last few years?

not since 2005 I think.

[iNerd]

It's worked out statistically, but I don't think you need to know it. Have they changed Psych at all in the last few years?

isn't it n = rootN where n is your sample size and N is your population.
So if you have 100 kids doing Psychology you'd need to experiment on root100 = 10 ...to be a proper "sample". Not sure lol; someone slam this if its totally wrong

[minilunchbox]

What's another advantage for independet groups apart from no order effect?

[sillysmile]

What's another advantage for independet groups apart from no order effect?

it is the easist and most time effective design.

[masonnnn]

i'd say individual participant variables is the main one as 'more participants required' would also apply to matched-participant so it's not really a disadvantage specific to just independent groups design.

and

how small do the sample need to be for it to be deemed not repreesentative

we learnt that 10% is usually required.
yet if you get asked for a sample size just say like, 100. that large a number should always do.

[Visionz]

What's another advantage for independet groups apart from no order effect?

IMO thats a weakkkkkk advantage anyway. Its the disadvantage of another design rather than a direct advantage of IG design.

[sillysmile]

i'd say individual participant variables is the main one as 'more participants required' would also apply to matched-participant so it's not really a disadvantage specific to just independent groups design.

and

how small do the sample need to be for it to be deemed not repreesentative

we learnt that 10% is usually required.
yet if you get asked for a sample size just say like, 100. that large a number should always do.

yeah, but there is much less difference in ppt characteristics in matched pairs.

[sillysmile]

What's another advantage for independet groups apart from no order effect?

IMO thats a weakkkkkk advantage anyway. Its the disadvantage of another design rather than a direct advantage of IG design.

but they do have advantages and disadvantages relative to each other.

[Glockmeister]

The rule of thumb of these things is usually around about 30, but it does tend to depend on the sort of study that you're trying to do. There's also statistical techniques that can show how much people you need before you have a chance of finding an effect, but I haven't studied them yet.

The problem with saying things like 10% of the population of is that say you're trying to generalise to the entire world. With the world population now standing at 6,697,254,041 that would mean you need 669,725,404.1 people or 81836.75 people to get a meaningful sample. Not feasible, one would think.

[jinny1]

The problem with saying things like 10% of the population of is that say you're trying to generalise to the entire world. With the world population now standing at 6,697,254,041 that would mean you need 669,725,404.1 people or 81836.75 people to get a meaningful sample. Not feasible, one would think.

geeeez GLockmeister even knows the exact number of humans on Earth... is he God or somethin?