Pearson r Correlation in SPSS - Part 2
HTML-код
- Опубликовано: 2 фев 2014
- How to calculate the correlation coefficient in SPSS is covered in this video. The correlation is also tested for significance and a scatterplot is constructed.
For more SPSS help: tinyurl.com/nywhxw3
Correlation
Pearson's r Correlation
Correlation video in SPSS
Video Transcript: To run the correlation we go to Analyze, and then Correlate, and then Bivariate. And bivariate if you think about it, 'bi' just means two and variate it is just another name for a variable so this is the correlation we use when we have two variables, like we do here. So select Bivariate and then here notice hours media and college GPA are on the left-hand side here of our Bivariate correlations box. You should see on your screen hours media is already selected, so go ahead and click the right arrow button to move it to the Variables box. And then select college GPA and click the right arrow button once again to move it over as well. Now notice here under Correlation Coefficients that Pearson is chosen by default. That's Pearson's r, that's what we want to run here in this analysis. If you want to run something like say Spearman's you could check that box, but what we want to run is the Pearson correlation, which is the most commonly used correlation coefficient by far. So that looks good Pearson is selected. Everything else looks great let's go ahead and click OK. And when we click OK, the correlation analysis is presented in our output here. And the way you read this here is you want to see where hours media and college GPA meet or intersect. So these two variables if you see they meet right here in this box here and it's also shown here once again as it's redundant. So either box is fine to look at. Notice that our correlation where it's shown here is Pearson correlation is negative, so it is negative, .66 if I round to two decimal places and then the p-value is .001. So using an alpha of .05 and a two-tailed test where a two-tailed test allows for a positive correlation or a negative correlation, we can use the following decision rule as we have for all of our tests: if p is less than or equal to .05, the test is significant which here would indicate that there is a significant relationship between hours of media watched and college GPA. Whereas if p is greater than .05, the test is not significant, indicating that there is not a significant relationship between hours of media watched and college GPA. Our p-value in our output is once again .001. So applying the p value to our decision rule, we can see that .001 is in fact less than .05, which indicates the test is significant. In other words in our example there is a significant relationship between hours of media watched and college GPA. We could write our results as follows: There is a significant negative relationship between hours of media watched and college GPA. And then we have r, which is for Pearson’s r, we have 18 degrees of freedom. Now the degrees of freedom interestingly are not reported in this table here instead we just get an N. But the degrees of freedom for Pearson's r are equal to the total number of people in your study, or your sample size, minus two. So N - 2. But we have 20 people, N - 2 or 20-2 is 18, so that's where I got this here. Next we have our correlation coefficient of negative .66 which you can see in the output here when we round to two decimal places. And then we have a p-value of .001, which is shown in our table here. Notice as well in the output here that SPSS uses asterisks to indicate if the test is significant and they might be a little hard to see but there are two asterisks here. And notice the note down here where the two asterisks are showing and it says correlation is significant at the .01 level, two tailed. Now if you saw one asterisk showing instead of two, this would read correlation is significant at the .05 level, two tailed. And it's just the way SPSS outputs it by default to make it a little more user-friendly. We could turn that off if we wanted to in the Bivariate correlations box but we just left it as is.
RUclips Channel: / statisticsinstructor
Channel Description: For step by step help with statistics, with a focus on SPSS (with Excel statistics videos as well). Subscribe today!
RUclips Channel: / statisticsinstructor
This is an excellent resource. I have a clear understanding of the Pearson r Correlation now. Thank you so much!!!!!
+Angela Thomas
Very glad to hear it, Angela; thanks for letting me know!
9 Years later this video still holds up. Thanks a lot!
Great video! Clear and concise explanations
always set this to 2 times speed for maximum efficiency
"The degrees of freedom are equal to the total number of people in your study, minus 2."
THANK YOU. It is absurdly difficult to get this as a straight, simple answer and it just should not be.
This lecture has demystified the wrong perception I had about SPSS. I am grateful
Thank you! Your explanation was very helpful 😀 I wish I could send you a case of cupcakes. I have reread the text over and over, taking a lot of notes. This video makes perfect sense. Thank you thank you. Do you tutor?
THANK YOU! Your video is incredibly helpful! THANK YOU!
hi, can u help explaining mine, my p value is above the 0.01, its .287 while the correlation is .085, theres no significant relationship. so, how do i write it in a report? is it " there is no significant relationship between x and y, thus the null hypothesis is not rejected. is it?
You're a lifesaver. Thank you
Thank you for your video! This was easy to understand and extremely helpful!!!!
My pleasure, Sharlene. Glad to hear it helped!
I think your video does a great job of explaining things. I do think it is important to note that you should only run Pearson's correlation coefficient if your data meets the assumptions that Pearson's requires. Even in an introductory statistics class, one should never run a statistical test if there are any assumptions that may not be met/violated.
I can't disagree with your there. Also to use only on linear relationships, as it will underestimate nonlinear relationships.
Thank you very much for sharing this helpful video regarding this subject matter. Hope you do more of this video. Really appreciated
I love it. Thanks very much for the clarification.
Thank you so much! This was so clear, well done, and very helpful!!!
Thanks for the feedback, Ali. Glad to hear you enjoyed it!
This is great - really helpful with my statistics assignment! To be honest, this helped me more than my lecturer and seminar leader :)
empyreantic
Very glad to hear it helped. Thanks for letting me know!
Excellent! Thank you very much. A life saviour indeed
thank you ! really helpful!
Dude, thanks for the APA format!
+David Keane
You bet, David. Glad to hear it helped!
you guys are awesome! thanks a lot for the videos! Ive subscribed :-)
thank you! very helpful tutorial :)
thank you! i am inspired
thank you, this is so useful and simple to understand.
María José Camacho
Thanks, Maria. Glad you found it helpful!
Thank you so much for these videos. I did not have a good Statistics/Research Methods Class and with these videos I was able to grasp the main concepts in order to run my data on SPSS. Thank you for talking in a very simple way that people like me who are not intelligent in Math could understand as easy as 1,2,3. Thanks!
Hi Jose,I'm glad to hear they helped - thanks for letting me know. I wish you great success in all your classes!
Thanks, I am schedule to defense my thesis on June 26.
Best of luck with your thesis defense, Jose!
statisticsinstructor Thanks!
EXCELLENT! Thanks
This is great! Thankyou
Thanks for excellent explanation! What deos 'p' stands for in 'p=.001'?
Hey, thank you for this. I was just wondering how can I use this as a reference so i can credit you?
Is there a demonstration of Spearman correlation coefficient test? because I want to test two ordinal variables but When i run the Spearman correlation test on spss I get a missing output. the " level" under the table (** correlation is significant at the 0.01 level) does not appear and so I can't get the scatterplot any solutions please?
This makes so much more sense. Thank you so much for this video. How do you test for reliability?
+Deb Gates
Hi Deb,
You are welcome - it is my pleasure! Good question -- we usually don't test for it (since it should be far above zero), but just treat is like a descriptive statistic, where higher is better.
QS
omg I LOVE YOU!
Hi, i want to know how can we deal with negative values like for example -127.79. i cannot type in the SPSS a negative value . how can i deal with it ? please help
thanks you helped me out
thank you!
Wonderful video
Hi thanks for your video
How can I show the equation of correlation please?
Hello
Thanks for video. Can you help me to explain my data (Sig 0.000
Pearson Correlation 0.407) . So can we say Sig = 0,000 is no relationship between them ?
Thanks
why did you use =a.05 as the rule for assessing?
if the ** is not stated, how to interpret the p-value? The result of analysis not stated the **
Please accept my gratitude!
It would be nice to know how many parts there are (i.e., Part 2 of x).
Thank you, It was really helpful. Can we do the same procedure for the categorical variables?? Does it make sense to use Pearson index for categorical variables???
+Ali Hoseini
Hi Ali,
My pleasure. Yes, if both variables are dichotomous, then calculating the correlation would be fine (also called the phi coefficient in this case). You can also do calculate Pearson's r with one dichotomous and one continuous (called the point biserial).
With two ordinal variables, then you would want to use Spearman's rho instead.
QS
very helpful!
Tareq Furas
Thanks, Tareq!
Hi thanks for your video,
I want to know how you set the Alpha, can you explain that to me please! You said in the video that the "Decision Rules for assessing that test is significant is Alpha = 0.05"
Is this constant? Is it editable on SPSS? should I worry about it? Is this a pearson's rule for significance that is unchanged? What is the APA format? are there other formats?
Thanks!
+Rami Massoud
Hi Rami,
Alpha of .05 is standard in practice. Some people use alpha = .01. APA format is the format of the American Psychological Association. It is common and used in many disciplines, but there are certainly others as well.
thank u!
Jiani Wang
My pleasure, Jiani!
Many thanks
Glad you found it helpful!
Great video and explanation. I have been watching lots of videos to understand the interpretation of Pearson r and most of them gave me a real headache. Yours is just WOW *-* .. Thank you so much. Could you do one explaining regression too as i have to submit my dissertation but i am unable to explain the results of my coefficient / anova for regression in SPSS.. i would be so grateful as i need to submit in one week.
If ever you are able to do it, it would be nice if you could add the written results like in this video.
Thanking you in advance for your reply.
Kreshila Kats Hi Kreshila,
Thank you for the feedback - I'm glad to hear you found the videos helpful. We have a regression series available (with written explanations) here: ruclips.net/video/thO_Wfdc6O4/видео.html
Please let others know about us as well if you get a chance!
And good luck with the dissertation defense!
Ron @ QS
Quantitative Specialists Hey Ron,
Thanks a bunch!
You are like a saviour ^_^ ..
Well, i will surely let others know about your awesome team.
Good continuation in what you guys are doing.
Take great care.
Have a great day ^_^
Sir,
Students t test interpretations are quite puzzling to me.
Group A mean=76.68,sd=19.54,
t=-5.44 , p = .58
Group B mean = 79.84,sd= 21.48,
Inspite of differences in the mean why is this t test not significant.
Thanks for the help.
Many thanks for the video its really helpful. I was wondering if you can help me with my doubt. I have measured latent variables x(with 10 observed items in the scale) among father and son dyad where both the members are responding to the same scale individually. Now I want to check the correlation between the responses of both the respondents to check the non independence. How can I check the Pearson co relation of the 10 item scale and two respondents. Would really appreciate your help.
Hi Pallavi,
My pleasure. I'm glad to hear you found the video helpful.
In SPSS, each row should be a different dyad, with, for example, father in column 1 and son in column 2. Then you would have 10 rows of data, one for each family dyad. Run Pearson's r on these data and that should do the trick!
QS
Quantitative Specialists many thanks for the reply. just to confirm if I understand it right, so with 10 items in the scale, I will have 20 column (10 for father 10 for son) for each dyad. So will I need to run Pearson's r each column wise for example...column x1 of father with column x2 of sun for dyad 1 for item x1? or should I calculate total of the scale for father and total of the scale for son and then run pearson's r on that total??? would really appreciate your reply. many thanks
pallavi singh If you want a single measure of agreement (one value returned), then it would have to be the total score ran on each dyad. Do in other words, row 1 in Data View, son total - dyad 1 (column 1), father total - dyad 1 (column 2). Then row 2 - dyad 2 son and father, respectively, and so on. Then run Pearson's r on these data.
Otherwise, you would need to run pearson's r on each individual dyad and then the simplest approach would probably be to take an average of all of these values (and report other descriptives - min, max, std).
It gets a little more tricky due to the multiple measurements across multiple dyads.
Good luck!
QS
Why 18-2? Can someone please explain?
Why is it 20-2 when writing down the correlation relationship in APA style?
Alana van Diggele
Hi Alana,
It's the way degrees of freedom (df) are calculated for the test. In APA format, df typically goes in parenthesis following the name of the test statistic that is used, for example: r(18), t(18), and so on.
An easy way to think of it as N-2 (where N = the total number of pairs of scores or numbers) is because there are 'two' variables being calculated in correlation.
I hope this helps.
QS
what if i get a p value of .000?
is it normal or is there any problem with my data entry
yoven ragu
Hi Yoven,
That is not a problem at all. What it means is that your p-value is very small (less than .005) and SPSS just rounded down to .000. It is a bit of a pet peeve of mine, as no p-value is ever truly zero (there is always some probability of getting the results obtained if the null was true), but this is the way it is reported in SPSS.
QS
nicee..thank you so much:)