HThe problem is that I recorded these variables in the field via qualitative methods. An example question from my data collection booklet is “How dense are the buildings in the area?” and pre-determined responses were “Dense”, “Intermediate”, “relatively sparse” , “sparse” and “no buildings”.
This particular variable sounds like it's ordinal. For a bivariate relationship between this variable and temperature you could look at something like Spearman's rho... but it's not ideal, because I imagine that getting a precise idea of the strength of the relationship will be key here. The spearman's rho correlation between ranks on the ordinal variable and ranks on the temperature variable does not give you a very direct measure of how much of a difference in temperature the different categories make.
One thing that strikes me is that the urban heat island effect is quite a well-known scientific problem that's attracted quite a lot of study (not to mention dubious media attention...). Maybe looking at the research techniques of others in the field might be helpful? I think climatologists have quite sophisticated quantitative methods for looking into problems like these, which us here on the forum might not be v familiar with.
You're a good man, thanks a million. Yeh, proper UHI researchers do have sophisticated methods of measuring these variables (ground cover and albedo with a pyranometer for instance) and a lot of it is done with modelling also. I didn't have the right gear to record exact values but since this is only a sub-component of my study (the main part was to examine the magnitude of the UHI which I have done) and it's undergraduate level I will be fine as long as I can demonstrate an understanding of the issue and a sense of how these variables are relatable to temperature. I've tried Spearman's rank and I think i'll be OK with that so thanks for the recommendation.The only thing i'm not sure of is how to calculate the p significance value for the coefficient in excel.
I am currently doing my dissertation at the moment and using spearmans rho to test for a relationship between heart rate and emotion. I am also using Kedall's Tau-b which is another option for comparing an ordinal level and continuous data like you are doing. I did a lot of the graphs I'm using in Excell, but for the actual statistical analysis I used SPSS. If you don't have in on your own PC I would be surprised if your uni computer didn't have it (maybe try visiting the maths/stats department if your uni has one?) On SPSS you can easily copy the information from Excell and paste it right into SPSS and then it can be used to calculate the correlation coefficients and the p-values for you very easily (again, if you find someone friendly in the maths department they should be able to show you in a matter of minutes). I hope this helps
Cheers Pat, I will check out Kedall's Tau-b - never heard of it before now. Yeah there's definitely SPSS in the geography department but I guess I've been set in my ways and stuck to excel. I will definitely be experimenting with if after your recommendation since I really need to calculate the coefficients and p-values. Is it easy to use for a first-timer?