matched pairs design help

Perhaps you want Independent group design?

Matched Pairs is incredibly time consuming and can riddle your experiment with confounding variables. I don't really see how this design would be appropriate in this level of experiment. Perhaps not, but if you could clarify upon this, I could try and help out some more.

Just to clarify:

Matched Pairs: You form your experimental and control group by pairing your participants by a chosen matching attribute. I.e. intelligence. You pair two participants of 110 IQ, one goes into the experimental and the other into control. Repeat. This is pretty time consuming and has to be done very correctly for it not to interfere with your results.

Independent Group: In this design, the participant only experiences one condition, either with music or no music in this case. The simplest method would to just randomize which group the participant enters. However, you also introduce the opportunity for participant variables to occur and it's not a very economic use of participants as you effectively half your sample size.

thanks,

well I was originally going to do independent group but my teacher told me that matched pairs would be better for this experiment since it involves anagrams and she said that it would be testing their skill to solve anagrams rather then their memory and said by doing a matched pairs design it can rule this out slightly...

I don't mind if it's time consuming to be honest just need to know how to do it exactly as we never really went over it in lesson

so would I give the participants a list of words, anagrams, etc? and then match them with people who got around the same result??? (there are going to be about 15-20 in each group)

sorry about all this.

Oh, well.. Okay. I see what your teacher means.

Yep, your idea seems good. Before grouping the participants, give them a little anagram test. Say you mark it out of 10 and you should get a bell curve providing you have enough participants and no ceiling/floor effects.

Then yeah, as you said, try and pair them up. It probably won't be perfect, but it may ease some of the participant variables. So you match the 8/10's, split them up, the 7/10s, split them up, etc etc.

Might as well keep your teacher happy.

Thanks

right so I'll have all 30-40 participants in a room give them a list of about 15 anagrams give them 5-10 minutes and then see how many each of them got right and split them up that way. If there are some people who score really low and don't match with anyone would I still use them and split them into a different group and repeat the experiment with their group after testing the others if that makes sense?

so basically I'd do something like

1. test the participants (give them 15 anagrams to solve)
2. split them into groups (below average, average, above average based everyones score)
3. test the below average group with pop/rock music and then without music and then do the same with the average and above average group.

It's tricky indeed.

I would split the potential scores into percentiles as you said.

Quartile One - 0/15-3/15
Quartile Two - 4/15-7/15
Quartile Three - 8/15-11/15
Quartile Four - 12/15-15/15

Then say you end up with participants being distributed around these catagories at the end of your experiment, as I presume you won't be gathering them all and testing them at the same time.

So say when you give your first participant his/her initial test and they fall under Q3, you randomise them into a group (experimental/control). Then when you have another participant fall under Q3, you place them in the opposing group as the first mentioned participant. Continue until you have a big enough sample and exclude those participants who fall within quartiles that make the groupings unfair by putting in an extra q1/q2/q3/q4 participant in one of the groups.

It's gunna take ages though, lol.

Edit: OH and make sure you don't get ceiling/floor effects on your anagram test, or else you'll end up with all your participants in either quartile one or quartile four. It would be worth doing a pilot study to make sure you haven't made it too hard/easy.

No worries. Good luck.