Spearman Correlation on SPSS with Write Up (APA Style)

hello so welcome to the video today we're going to take a look at how to do spearmint correlation analyses on spss so speedrun correlation analyses are like the non-parametric equivalent of pierce and correlation analyses so with the pearson correlation analysis you assume that you have normal data and that the data are linearly related to each other um but with the spearmint test uh those assumptions don't have to be met um so you don't need to have normally distributed data and even though you assume that one var has one variable goes up another variable goes up or the other variable goes down uh the relationship between those those two variables doesn't have to be linear and while you would normally use uh continuous data for a pearson test with a spearman test you can use original data so ordinal data are basically groups that you can order hence the name ordinal so for example um level of income could be like low medium and high so there's like three groups but there's an inherent order to those groups so that would be one example of ordinal data um so in this one if we take a look at the in this video if we take a look at the excel spreadsheet we're going to imagine that we're interested in the relationship between education salary and job satisfaction and so in this spreadsheet one is going to represent like a low level of education five will represent a high level of education and the same thing for salary so one is low five is high and with job satisfaction one equals dissatisfied and five equals satisfied so if we take a look at so if we go to sps we can take a look at how to enter these data into the program so we can go to variable view to start with and i'll just tell sbs the names of these variables so this one we can call for one of them education we can call one of them job underscore because we can't use the space satisfaction and the last one is oh this is wrong around actually this one is salary i'm gonna put this one as job satisfaction okay um so once i've done that i'm just gonna use this values column to give um the numbers some names we have those numbers between one and five we saw before so now i'm just going to give those numbers some names so we'll say one equals uh no education two equals high school three goals let's say bachelor to uh for this masters uh it's not clear i'm just clicking add every time i i uh enter something so you've had that appears there so five put that as a phd add okay so that's we have five numbers and we have five names for those numbers um so i'm just gonna do the same thing for salary and job satisfaction so one is going to be let's say low two can be quite low three will be medium for the white high and five will be high okay so i've entered those i'll just go to okay and then i'll just do the same thing for job satisfaction so one can be dissatisfied two can be quite satisfied three can be neutral four can be quite satisfied and five can be satisfied add okay all right so we've now told spss what those numbers are i'm just going to use this measures column to indicate what type of variables these are so these are all ordinal variables as i mentioned before and i'm gonna use the days view and we can see that these these names have appeared at the top of these columns then it's just a matter of uh copying and pasting this data from the spreadsheet so copy and paste and we should see that paste one second copy paste what's going on copy paste place paste okay there we go um so all those numbers so for example if we look here we have a one in the education top row but this has been replaced by the word num in spss if your file doesn't look like this if you just see numbers here even after you've given the number some names you can just go to view and check that value labels is checked okay so i mentioned that one of the assumptions of the spearman's test is that as one variable goes up another variable goes up or the other variable goes down but that relationship doesn't have to be linear so let's just take a look at how we can look into that assumption so we'll do this by going to graphs down to legacy dialogues and to scatter dots i'm going to choose simple scatter and then define i'm going to put education in the x-axis and job satisfaction y-axis i'll go to okay and we can see that basically as education goes up so this is education at the bottom this is job satisfaction on the side this is low education this is low job satisfaction so you can see that as education goes up so does job satisfaction so we kind of have a diagonal line like this this suggests that the relationship is going to be positive so a positive correlation just means as one variable goes up so just the other variable um it would also be also it would also be okay if we saw the opposite direction here so if we saw that uh the line went in this direction instead that would also be fine that would just indicate that the relationship is negative rather than positive so let's just do a couple more graphs for the other potential relationships so graphs legacy dialogues scatter slash dots simple scatter define i'm just going to move job satisfaction out i'm going to put salary there and that's just going to create the same type of graph again we can see that there's basically a positive correlation between them as one goes up so does the other and then we'll just do the the last one so graphs legacy dialogues scatter dots simple scouts are defined and this time let's go so we're going to put job satisfaction here from this little one we haven't checked yet job satisfaction and salary so so far we've looked at education and job satisfaction education and salary and now we're looking at salary and job satisfaction and we can see that we have the same thing again so as as job satisfaction goes up so the salary so that assumption that there is like a relationship between the variables as one goes up so to the other or down that that assumption has been met so once we've checked that we can run the analysis itself so i'll go to analyze then down to correlate down to five variants i'm just going to transfer all of these variables across to the variables box and check spearman and untick pearson and i don't need to check anything else so i'll just go to okay and that produces this table for us so if we take a look at education first so uh we go across from here and we'll get to a salary and we can see that value is 0.763 basically correlation values range from minus one to plus one and if they're close to plus one that means it's a strong positive correlation if they're close to minus one that means there's a strong negative correlation and if it's somewhere close to zero uh well if there's zero there's no correlation if they're let's say it's uh minus point one that would indicate a weak uh negative correlation so in this case we don't have a minus symbol so we know that this this um correlation is positive and we also know that it is significant because if we look at this sig value below it's below 0.001 so really the value spss only shows this value to three decimal places uh so the value isn't really zero it's just somewhere below 0.001 so we can say that there's a positive significant relationship between education and salary and then if we look at uh job satisfaction we can see that the same thing is true for that so education job satisfaction we have another value that's close to positive one and again we have a c value that is less than zero zero one so another positive significant correlation there um if we look at salary and job satisfaction so salary here we'll go across to job satisfaction and we have this uh spearmint's row value of 0.836 so this is a positive uh significant relationship and we know it's significant because again this value is below 0.001 um so those are the results that's how to interact with them and let's take a look at how we can build them so let's have a look at this paragraph first i've just started off by saying what test was run so we ran spearman's rank order correlations to examine the relationships between level of education salary and job satisfaction and these results are quite easy to report because we observe the same type of correlation between all of the variables so we can say there were positive and significant correlations between education and salary education and job satisfaction and salary and job satisfaction and uh we've included some stats here so let's just take a look at where those stats come from so we've got the rs equals 0.76 here so that's the first one we looked at before because we're looking at uh education and salary in this case uh so education will go across towards the salary it says 0.763 and we've just rounded that value here to two decimal places we've got n equals 20 next that is actually just the the number of participants um if we're not sure about how many participants we have we can also just check this table because it says n equals 20 and that's where that comes from and the last value then is the p equals point or less than 0 0 1 and that's just this this sig value here it's 0.000 if that value had been point zero zero one or above i would have used p equals and then whatever the value is so it was uh p if the same value was uh point zero five three i would have written p equals point zero five three and i wouldn't have used the less sign symbol um and then we just had the same thing for these these other relationships so hopefully that would be obvious where those values come from um in the case of this particular analysis i probably wouldn't present the results in the table just because there are only three correlations and it's very easy to report them just in text but if you had lots more correlations it might be easier for the reader to figure out what's going on if you included these stats in a table instead of in a paragraph so let's just take a look at what a table might look like so this is within apa style you can tell because uh well for example this table one is above the actual title the title is italicized it has capital letters uh with the exception of uh short prepositions such as and and articles but we don't have any articles in this title and also the table doesn't have any like versatile lines within it it's all just horizontal lines so that's sort of the apa style um so these are just the the rs values um that we looked at before so uh seven six is there and i have the same thing here so that just represents the relationship between salary and education and i put this little asterisk here which is the value just to indicate the value is significant so the asterisk tells the reader that the value is less than 0.001 if you had some values that were less than .001 and some that were less than 0.05 i might use two asterisks to distinguish between those so i could use two asterisks to stand for .001 or below that value and i'll probably use one as for x to refer to values below 0.05 uh so i think that's about all there is to the spearman's correlation analysis and how to present them um if anything's unclear please just let me know in the comments and thanks very much for watching

Channel: David Robinson, PhD

Views: 1,523

Rating: 5 out of 5

Keywords:

Id: pcSTjVFu-xI

Length: 14min 5sec (845 seconds)

Published: Fri May 28 2021