Spearman's Rho

I'm in the middle of doing my alevel coursework on the halo effect in individual stereotyping. I got participants to firstly rate pictures of people on attractiveness and then asked them to rate how friendly, happy, intelligent, optimistic and vain each person was just by looking at the picture. I've done my experiment and now need to do spearman's rho statistical test for my results section. The thing i'm stuck on is ranking my mean scores in order because some of my mean scores are the same. Has anyone else had the same problem or does anyone know much about spearman's rho test that can help me? My coursework has to be in on wednesday and i cant see my teacher until tuesday :frown:

Any help will be really grateful.

-Katie xxx<3

Reply 1

shewantstobeme

if they're the same, it shouldn't matter :smile:

see if the example on wikipedia helps:

http://en.wikipedia.org/wiki/Spearman's_rank_correlation_coefficient

Reply 2

welshprincess

The thing i'm getting confused with is that when i'm looking at examples some numbers that are the same end up being ranked as 4.5 or 2.5 but then others that are the same get ranked as solid numbers like 5 or 6. One of my results has 2 numbers that are the same and they are both the lowest score i think its like 2.1 or something...would they get ranked as 2 and 1 or would i miss out 2 and they both become rank 1 :s i'm so confused and frustrated i could cry :frown:

Reply 3

shewantstobeme

welshprincess

i'm not really sure, because i'm not on the same exam board as you, i just know how to do spearmans because i do a-level maths. the cw i did on edexcel had no statistical analysis.

what was your hypothesis?

Reply 4

welshprincess

My hypothesis was that there would be significant correlation between attractiveness scores and my friendliness, happiness, intelligence, optimism and vanity scores. My teacher started explaining to me on wednesday about what to do if i had 2 of the same scores but had to rush off before she'd finished so i'm none the wiser. I cant believe we have a day off on monday its the 1 time i actually wish i had school!!!

Reply 5

Opalfire

Ok heres what you do if you have 2 scores;
here is an example list
3 so you rank 1
4 rank 2
5 rank 2.5
5
6 rank 4

Does that make any sense? Its been a long time since i did a level statistics! Hit me up if you need any more help x

Reply 6

welshprincess

I can't get my head around it at all, I'll post my results as an example...

4.4 - rank 10
2.9 - 5
3.8 - 8
4.3 - 9
2.8 -
2.8 -
3.0 - 6
2.7 - 4
3.1 - 7
2.4 - 1

What would both 2.8's be? Would they be 2.5 or 3.5 or something like that? I'm so confused :frown:

One set of my results also has 3 results which are the same and i have no idea what to do about it, its giving me a headache :frown:

Thanks for all your help so far i'm really grateful and i'm trying to take it all in my results are just so awkward!

Reply 7

Robbolo

I think you need to put them in numerical order before you assign rank values.

Reply 8

beggyboo

when there are two numbers the same, their ranking would be the same but it would be a .5 (point 5) number, for example, the two 2.8s you have would both be 3.5 because they are the same and then you would miss out the 4th ranking and so the 2.9 would be 5....i hope that makes sense....!xx

Reply 9

psyc_postgrad

Here are a couple of examples of how you'd rank scores if you had several the same (I've shown the scores on the left and the ranks on the right, with a - inbetween)

12 - 8
9 - 7
7 - 6
5 - 4.5
5 - 4.5
4 - 3
3 - 2
2 - 1
(for the scores of 5, you add the relevant ranks and divide by 2: 4+5 / 2)

9 - 8
7 - 7
5 - 6
4 - 5
2 - 4
1 - 2
1 - 2
1 - 2
(for the scores of 1, you add the relevant ranks and divide by 3: 1+2+3 / 3)

Reply 10

welshprincess

Thank you all so much for your help :smile:

I finally get it its only taken me like a week lol. I've done all my ranking and i've done my equations now i just need my teacher to have a look to check they're ok. My results show no correlation between my attractiveness scores and happiness, friendliness, intelligence, optimism and vanity score but when you look at my rankings a lot of them are the same for attractiveness and the personality trait. Someone said that spearman's rho doesnt show correlation because i only had 10 participants so even though my ranks show theres a connection, i'd need more participants for the spearman's rho test to show that. I don't know if thats correct or not :s
Thank you all again :smile:

I really appreciate it
xxx<3