Correlation Coefficients in SPSS/PASW

A short tutorial on calculating a correlation matrix in SPSS (now known as PASW). Describes the range and strength of associations, as well as positive and negative associations. Also describes the pearson correlation (r), the significance level (p), and the sample size (n).

Category:

Education

Tags:

• statistics
• tutorial
• how_to
• correlation
• coefficient
• matrix
• SPSS
• PASW

[theplanetofmatt]

Where can i get PASW 18 for my mac?

[bartonpoulson]

@theplanetofmatt Well, you have to buy it. Time-limited student versions are a little under $100 and you can get them at an online store called studentdiscounts. (By the way, it’s back to its original name, SPSS, and its up to version 20 now). If you’re not a student but you’re fast, you can download a trial version from IBM.

[KilloDel]

do you always have to use correlations when using spss?

[bartonpoulson]

@KilloDel Not at all; I just think they're a very handy statistic. If you're looking at association, you can also do regression, t-tests, ANOVA, chi-square, whatever. However, most of these also translate to correlations, which makes them useful.

[bartonpoulson]

Basically, that's right. The important thing to remember is that the .05 cutoff is just a convention; a very strong one, for sure, but ultimately arbitrary. That being said, most people would consider a higher p value (e.g., p = .30, p = .17, p = .06) to signify an essentially null relationship – that is, that the true correlation could easily be 0. Hope that helps!

[kbl3000]

greeting from uni of birmingham

[bartonpoulson]

@kbl3000 And right back at you from Salt Lake City in the heart of the Rocky Mounains!

[TehAznSmiley]

Barton Poulson, the man with the mission to make SPSS easier for dummies like me. thanks.

[sadiyazaheer]

please zoom the screen next time......:/

[henry1086]

hey so im working on prison violence and i put in the variables for total assaults and number of white C O officers and white inmates and other races also i.e. black and hispanic. is this test the right one that i should be running to determine/show a relation? please help

[trapadoored]

'm doing this and I just got a correlation of -.253 between two variables. Would that be considered a strong correlation? Or a weak one?

[mehranalimemon]

Hi, how i interpret the P values if they are correlated more thn 0.75. Any other way involved to interpret the higher correlation, or what will happen if thy r correlated more thn 0.70 or 0.80

[theplanetofmatt]

@bartonpoulson Thanks for the response. I am a student in the UK. I have tried to find trials on IBM but i have been unsuccessful. I will try studentdiscounts as you suggest, as i'm happy to pay. Also, your videos have been really helpful to me, so thank you :)

[chopin7tristesse]

note that it is about LINEAR CORRELATION that is stated in passing between 3:21--3:24.

[chimonimoni]

hi bart, i have low results as low as .299 and .319. what does this mean.