Spearman Correlation - SPSS (part 1)

hi welcome to how to stats comm in this video I'm going to demonstrate how to perform a Spearman rank correlation in a previous video I demonstrated how to do a Pearson correlation which is appropriate when your data are scored on a interval or ratio scale however if your data are rank ordered or measured on an ordinal scale then you are left with you are left with not being able to do a Pearson correlation because a Pearson correlation assumes that your data are measured on a more informative basis which is the interval ratio scale now in this case the variable that I'm going to examine two variables that I'm going to examine are rank order and one of them is education in the statuses bank loan from from the SPSS sample data sample data that you can get and EDD or education I'll look at the variable view and it's um it's got a label and it's telling me its level of education and it goes from did not complete high school high school degree some college college degree and post undergraduate degree so this is the rank ordering of this is a variable it's a rank ordered ordinal variable so this suggests to me that I should be doing a Pearson correlation because my other variable is also or no scaled and what I've done is I've made an income category variable from the income variable I would never suggest that somebody actually does this because you're losing information by doing so but in this case what I just to do this Spearman correlation I've created an income category variable which is goes 1 2 3 4 5 6 from very low to wealthy so it's a very little low medium high very a high wealthy I don't think this perfect saw the ideal representation because basically what I did was zero to $25,000 a year household is very low - is 26,000 250 medium 51 275 high 76 to a hundred wealth very high 101 225 and then anything over but more than one hundred twenty-six thousand dollars or more is considered wealthy I I just did that in the first go I think I would change things if I wanted to make it a bit more accurate more reflective of reality but for the sake of demonstrating a Pearson correlation this is this is the kind of data you frequently are presented with people measure household income but they only measure it in a in an rank-ordered way rather than a fully informative way of measuring exactly how much people earned or how much they're reporting that they earned for the for the year they just have category bands from very low to meet to very wealthy or they'll just say zero to twenty five twenty five to fifty and people select which one best represents them so if you wanted to examine the association between years of association years of education rather which is basically a rank-ordered type of variable ordinal variable categorized across levels of diploma and end degrees and then income category you have to use a Spearman rank correlation arguably and I'm going to show you how to do that you're going to analyze then you go in to correlate and then you're going to bivariate and then you put the variable you're interested in so we've got a level of education actually the it's got the right symbol income categories actually got the nominal variable and actually let me change that and see if it doesn't automatically so I'm going to change that to ordinal and I'll save that alright so analyze and correlate bivariate got a level of education and then income category now to default in SPSS is to have a correlation coefficient estimated that's a Pearson correlation I don't want the Pearson correlation in this case I want the Spearman correlation alright I'm going to click OK and here is the correlation presented level of education is increases as level of education increases across the ordinal categories income also increases it's not a very strong correlation it's only point two one in the previous video where I did Pearson correlation the correlation between age

Channel: how2stats

Views: 283,220

Rating: 4.5813084 out of 5

Keywords: Spearman, rank, correlation, SPSS

Id: r_WQe2c-ISU

Length: 4min 55sec (295 seconds)

Published: Thu Sep 01 2011