Statistics: Correlation and Regression Analysis in SPSS

Agron Kaci
意见 224 344
97% 865 23

This video shows how to use SPSS to conduct a Correlation and Regression Analysis. A simple null hypothesis is tested as well. The regression equation is explained despite the result of the hypothesis conclusion.
Buy Andy Field's textbook here: amzn.to/2yxomuQ
Buy SPSS (Student's version) here: amzn.to/2g19Ofc
This book is written by Dr. Everett Piper, President of Oklahoma Wesleyan University. Analyzes the current higher education system: amzn.to/2y6tpRk
Watch Using Excel to find the Correlation Coefficient r here: cnvid.net/video/视频-y3bgaLwdm50.html
Watch ANOVA in SPSS here: cnvid.net/video/视频-Bx9ry1vBbTM.html
Watch Sampling Distribution of Sample Means here: cnvid.net/video/视频-anGsd2l5YpM.html
Watch Using Excel Charts to calculate Regression Equation here: cnvid.net/video/视频-qZjTtnyaV70.html
Watch Using Excel to calculate Regression Equation here: cnvid.net/video/视频-LDC0p9iZY8g.html
Watch ANOVA in Microsoft Excel (One-Way) here: cnvid.net/video/视频-WhBkgWL3_3k.html
Useful stuff:
Robot Vacuum Cleaner: amzn.to/2xpNGCH
Roku Express: amzn.to/2yvvAPQ
Mini Coffee Maker: amzn.to/2y7S1tq
Visit my Men's Fashion Store here: mensfashionstore.siterubix.com/




correlationregression analysislinear correlationSPSSstatisticshypothesisnull hypothesis testingstatstatscollege statisticscollege statsuniversity statshigher educationstatistics educationhow to do regressionregression helpcorrelation help






评论 120
Shubham Lahan
Man, This is the most clarifying tutorial I've come across. Thank you. Thanks a lot. You're doing a fantastic work. Keep it up. Best wishes to you. :)
QueenRose238 年 前
Same, this is extremely helpful.
Agron Kaci
Agron Kaci 年 前
Shubham Lahan... Thank you very much. It's encouraging to hear that. Good luck to you!
J 个月 前
Good and simple explanation.
Relaxing Music & Nature
🙋 ! Very good channel & cool video ! 👍👍👍852👈 🙋(I'm RelaxingMusic&N.)👈W👈E👈L👈 C👈O👈M👈E👈💃👨‍👩‍👧‍👦) 😊
lil coby
lil coby 3 个月 前
thank you so much!!!
Correlation and regression analysis were explicitly clarified to a logical conclusion in this video. Thanks a lot
Manal Samamah
Manal Samamah 4 个月 前
Thank you very much ..do you have a video on multiple regression
John Doyle
John Doyle 4 个月 前
Thank you
Bimala Jadhari
Bimala Jadhari 6 个月 前
Thank you so much for this tutorial. You made is so easy to understand.
Newman Omigie
Newman Omigie 8 个月 前
Thank you
Adem Tamo
Adem Tamo 11 个月 前
Crystal clear. Thank you!
arielle margueritte
arielle margueritte 11 个月 前
Thank you sooo much Sir !!!!!
passang sherpa
passang sherpa 11 个月 前
nice explanation
kuganeshvari eshvari
kuganeshvari eshvari 11 个月 前
thanks you ..i need it
Maruf Mamurovich
Amazing explanation! Thank you very much.
Jamille Richardson
This is by far the most enlightening and helpful tutorial I have viewed on the topic. I will seek more of your tutorials. Thank you so much!
Anissa Yuniashaesa
Hi, I have the same analysis. However, my age dataset is in range (ex: 0-7, 7-9, etc.). What should I do to analyse this? And should I choose R pearson correlation analysis? Thanks!
Agron Kaci
Agron Kaci 年 前
Anissa Yuniashaesa Variable Age as you have it, makes an ordinal variable. Please make sure that you meet all the assumptions, one of which is the variable should be at least on an interval level of measurement.
drakaay 年 前
Excellent work.. i am really embarrassed with your video. x show
Giorgio giorgetti
shume e sakte. thank you profersor. i have a question: if i want to see ho serum amyloid A beheaves in relation to the clinimetric index called BDCAF (continuous varibale) would it be better to use simple linear regression, or pearson correlation. and if i use spss, is nomrality automatically evaluated, or should i first check for normality distribution through shapirowilk test. thanks in advance.
Giorgio giorgetti
last question, although it has nothing to do with this video, at least not directly. ROC curves are for diagnostic purposes only, or they can also be employed for a biomarker that assesses disease activity/severity. I apologyze for the disturb, and hope you keep doing more videos on statistics and spss, since they are very precise, complete and simple to understand, especially for beginners. Enjoy your time Prof.
Agron Kaci
Agron Kaci 年 前
Giorgio giorgetti, thanks! Usually you'd perform the normality test before you test your Correlation and Regression hypotheses. If you need to test a predictive model you definitely need to run regression.
sumita pradhan
your explanation is clear and precise...thanks a lot
manar keshk
manar keshk 年 前
thank you for the amazing tutorial :) it is really simple and easy to digest :) you are amazing statistics teacher really thank you !
Hannah Jung
Hannah Jung 年 前
It is so clear. Thank you so much.
imran khan
imran khan 年 前
pls someone send me your email addresses so that i can forward these outputs for helping out me....
imran khan
imran khan 年 前
Can any one pls make regression table and correlation table from my Spss output,,,and resend me...here i am attaching these outputs....
Nyakoe Lamech
Say we have a project in finance that is tacking the effect of omni-channels business strategies on competitiveness of commercial banks. The channels are e-transaction, e-communication, e-banking, and e-finance, and these make up the independent variables of the study. In development of the questionnaire, each variable has 6 items/statements on a 5-point likert scale. Moreover, the dependent variable on competitiveness has six statements that should be rated on a likert scale. Now my question is, how do i carry out correlation analysis or regression analysis going by the fact that there exist so many statements under one variable?
Nyakoe Lamech
That is well understood. How do you go about variables of the study (independent) in a questionnaire having particular statements, say five each and requiring responses on a 5-point likert scale, and you need to carry out regression analysis? In SPSS, all these statements/items of individual variables will be there once you click on regression then linear. Help here.
Agron Kaci
Agron Kaci 年 前
I would think that you may need analysis of association, hence chi-square (cross tabulation), because your data is ordinal, not ratio.
Doreen Shamala Antonio Merino
Thank you very much.
Sanjit Agrawal
Its really good. Thanks a lot. Waiting for more videos from you.
Aminath Manal
Finally I understood something. Its great that you have used the ppt parts to actually tell us what all those stand for teaching the basics. My dissertation draft is due tomorrow and this helped me a lot. Thank you so much.
Hassan Mohamed
I need more lecture to understand spss internationally
Eriselda Mino
Faleminderit 🇦🇱
ikram noor
ikram noor 年 前
you have just made my concepts clear.. awesome tutorial...
Lim_May 年 前
Thank you so much!!! Your video help me a lots when i am suffering. Million thanks have to express for you. Appreciated
nangbaa bright
nangbaa bright 2 年 前
i love this tutorial
Nino  Alfonso
Nino Alfonso 2 年 前
Thank you very much Sir, your explanation is very clear!
Hossam Hossam
Hossam Hossam 2 年 前
Thank you Mr. Argon but usually we look to the Beta and not B which is 0.271 (slope)...please clarify
Ліза Буксанчук
Thank you! Now I know what to do with my weak correlation!!!
Saint Grégoire d'Arménie Monastère
Excellent video, with CLEAR explanations, especially for SPSS novices and statistics allergics! Thank you for sharing!
Aleksandra Milinkovic
Thank you for uploading this. It made things very clear!
Pius Ambros
Pius Ambros 2 年 前
It really gives the clear concepts on improving the skills of statistical inference
Agron Kaci
Agron Kaci 2 年 前
Thank you!
Moon Light
Moon Light 2 年 前
Are correlation and regression analysis applicable as nonparametric tests? does it go with convenience type of sampling?
Katherine Fowler
Thank you! This helps a lot! Do you have a video on confidence intervals?
Zeeshan Rafique
Zeeshan Rafique 2 年 前
It was very helpful Sir. Your effort is appreciated ! Thank You
ganesh rajput
ganesh rajput 2 年 前
Alicia Prasana
Alicia Prasana 2 年 前
wonderful video. explains everything anyone needs needs to know for correlation and regression. thank you very much.
ili lu
ili lu 3 年 前
me falni per kte pyetje idjote por kam nje ushtrim qe nuk po e zgjidh dot ,me duhet te llogaris ekuacionin e vijes lineare te dy variablave gjithashtu edhe r ,si behet kjo ne spss
Mamta Chaube
Mamta Chaube 3 年 前
thanks .it was helpful
bara khan
bara khan 3 年 前
Very helpful, thanks
Rajiv Sindwani
Rajiv Sindwani 3 年 前
Hi, I am doing research on automated banking service quality. In my work I have found in total 6 dimensions (factors)....4 service quality(SQ) factors,1 customer satisfaction(CS) factor and 1 loyalty factor. All factors are having multiple items. Now, I want to apply regression analysis using spss and make an equation of the form CS=a+b1(SQ factor1)+b2(SQ factor2)+.... Note: Independent as well as dependent both are factors and not items Please explain how I can apply regression on all factors. I used Command in factor analysis “save variables as factor score”. As dependent variable is also a factor score and not a single item, so applying linear regression in spss is not giving proper results. If dependent variable would have been a single item then it would have shown proper results. Please answer 1) Which method shall I apply to compute scores representing factors for using them in regression and ANOVA. 2) Which regression method will be most appropriate one ...linear, PLS etc and how to apply the same on factors? 3) Can I use mean value as a representative value of factors for regression (any citation for that)? Request you to please reply at the earliest. I shall be thankful to you. Rajiv
Марио Христов
Great video and great explanation.It makes Regression Analysis sound easy.Thank you!
T hu
T hu 3 年 前
Thank you so much! Very helpful!
Mahmud Hossain
Mahmud Hossain 3 年 前
I tried several times to understand the correlation and regression but failed. This video helps me to understand
anes dj
anes dj 3 年 前
Thank you for sharing these information and simplify it, I am dealing with multiple regression to assess the impact of the provenance of students (Home/International) and the course studied (MBA/MSc) on the employability of graduates in a University in the UK, as there are many independent variables, I am confused whether the correlation coefficient represents the correlation between all the variables (dependent and independent)!?
mayani71 3 年 前
Great! Video and explanation.....Thank you so much for such a clear and concise tutorial.
Princess Amrin
Princess Amrin 3 年 前
It is very useful but I am having the problem at the end. Mean to calculate regression equation( Y=a+b*X) Which means Y = Dependent Variable, A= Intercept(const), B= Slope (Coefficient Value) and X= Independent variable. You got Y=2.444+.40*X(Age 21) Y= 2.444 + 0.84 Y= 3.284 Which I understand. But I can't make mine following Y=a+b*X( I have constant =-.351 then Service quality =.181 ) However, my dependable variable Y= Customer Loyalty Y= -.351+.181*X(What should I put for this X?) Is it the number of items I have for service quality? I would really if you or anyone can help me.
Agron Kaci
Agron Kaci 3 年 前
Yes, that is X, the independent variable. You must include in the equation the measurement unit(s) of the independent variable.
Fikirte Demissie
it is very helpful! Thank you so much!!
Tom 4 年 前
I wish everything was explained like this!
Jew Bear
Jew Bear 4 年 前
Awesome video. I am currently writing a research paper using these measures. How would I differentiate between correlation and regression? Just that correlation is based on observing if there is a relationship and regression deals with being able to predict the dependent based on the independent?
Agron Kaci
Agron Kaci 3 年 前
That is basically it! However, mathematically, these concept are just about the same thing. Statistically, although you may see a high correlation between two variables, it does not necessarily mean that there is causation. Correlation, basically, tells you whether there is an association (existence), if this association is positive or negative (direction) and the strength of this association.
mobeen f
mobeen f 4 年 前
This is the best regression tutorial I've EVER watched. Everything was very clear and well explained. THANK YOU!!!!!!
Yinan Yang
Yinan Yang 4 年 前
We have learned for a whole semester about doing applied linguistics research and reading papers, yet our so-called professors never clearly impart us with the necessary SPSS techniques and reasonable analysis. They might themselves get baffled in understanding the data. It shames me to say that I am already an English postgraduate from an "eminent" university in China. So I am about to learn SPSS by myself. Thank you!
Dars Aijaz
Dars Aijaz 4 年 前
+Agron Kaci thanks for the tutorial, however I have a question , if I have to find the relationship between age and salary and the given salary is current annual salary and starting annual salary which test to use for it and which salary to choose as variable with age.
До Должин
Thank you, I Need it.
Johan Pieter de Borst
In your example you show firstly there is no correlation between the two variables, but, then, using the same variables, your regression equation shows a relationship between the two, hence the ability to predict. Is that possible with the same data? Or am I missing the point (that's fairly likely!)?
Johan Pieter de Borst
+Agron Kaci Thanks Agron!
Agron Kaci
Agron Kaci 3 年 前
Johan! Apologies for replying so late. Your question is excellent. The correlation coefficient, r= 0.271, is very small, therefore showing a weak correlation between the variables. The regression definitely shows the same thing, since mathematically, they crunch the same numbers the same way. Therefore, it logically follows that the relationship is insignificant. When the relationship is insignificant at the alpha level, we must retain the null hypothesis, therefore we find that we do not have enough evidence to accept the hypothesis that the relationship is significantly strong. Now, why did I bother and explained the regression equation when the model is insignificant? For pedagogical purposes! I did not want to create another video with other variables or datasets where I could find a significant association for regression purposes. That's all...
Johan Pieter de Borst
+Murtaza Mansoor Excellent, thank you.
Murtaza Mansoor
Murtaza Mansoor 4 年 前
Please note that there is a correlation between the two variables. It is just that such correlation is 'weak' (i.e. 0.271) which is less than 0.3. The regression analysis further supports this notion as it indicates that the strength of relationship given by the two variables used is only 7% meaning that there are other factors (amounting to 93%) that determine the relationship between the two variables.
Jew Bear
Jew Bear 4 年 前
That's what I am trying to figure out...
Cindy Baksh
Cindy Baksh 4 年 前
you are so cute and smart. thanks for this vid, saved my life!
Nicholas Webb
Nicholas Webb 4 年 前
If you've accepted your null hypothesis, then why would you then use your line formula as a predictor for age and GPA? I thought we just determined that age doesn't explain GPA with statistical significance?
anna melissa
anna melissa 3 年 前
why when i try doing my reliability test,i get cronbach alpha number section but i don't get any number from cronbach alpha if item deleted section? it just show a small letter (a) saying it was negative value?and say my coding is wrong?i check triple time already and nothing is wrong with my coding and key in data? :( the weird thing is i able to do the rest of analysis with that data :(
Agron Kaci
Agron Kaci 3 年 前
+anna melissa Anna, yes, you can combine variables into one variable and then use the combined variable for statistical variables. However, my video does not demonstrate SPSS' commands. A simple search for how to combine variables will get you to where you want to be.
anna melissa
anna melissa 3 年 前
+Agron Kaci i have 4 question in questionnaire that represent dependent variable and 3 question in the questionnaire represent independent variable 1,how can i combine all this question to represent depended variable and independent variable 1 because i want to use dependent variable with Independence variable in correlation analysis?
Agron Kaci
Agron Kaci 4 年 前
Hello Nicholas! Firstly, thank you for sharing my video! Much appreciated! Secondly, we never accept the null hypothesis, since we cannot create a new fake notion about the initial idea; we just fail to reject it, in which case we do not have enough evidence to accept the alternative hypothesis, which was our main research idea. Thirdly, the reason I explained the regression equation is for academic purposes only; just to show my students (in the video) how to properly set up and solve such an equation, the conclusion of the hypothesis notwithstanding. Thanks!
Peraya 4 年 前
Thank you so much !
Jeffrey Anderson
The best video I've seen on correlation and regression. Thanks.
Juan Carro
Juan Carro 4 年 前
prashant ubarhande
best for learners........
Agron Kaci
Agron Kaci 4 年 前
Thank you! Appreciated!
Nardi Gjergji
Nardi Gjergji 4 年 前
Thank you for this videos Prof. Agron. I am preparing my business statistics exam and your videos are very helpful. Si shqipetar nuk ju fshehi dot edhe emocionin qe me krijoj kur pash emrin tuaj, faleminderit shume!
Agron Kaci
Agron Kaci 4 年 前
You're very welcome Nardi! Hope your business statistics is going well...
Robert Manibola
Robert Manibola 4 年 前
This was a very helpful video. Thank you for posting it
Victor Velayos
Victor Velayos 5 年 前
I believe you should use a p-value of 1/100= 1%. Your confidence will be higher that you have the correct result ;)
syafika ghazi
syafika ghazi 5 年 前
Thanks for showing the easiest way to understand about variables correlations. However my independent variable show -.143. So does the variable is not correlate with my dependent variable? does it show negative relation between two variables? after that i did multiple linear regression and again the number shown for my variable 1 is positive and my 2nd variable is negative. What does it mean? thanks for the help.
Eka Listy
Eka Listy 5 年 前
Thanks for the very CLEAR explanation! It does help!!
Ganasoundari Joshua
Your video is very simple and answered some of the basic questions I had. With regard to the value of constant, I was told that it refers to other factors that are affecting the independent variable other than what we have considered (in this video, GPA). Please let me know if I am right or pl explain
Agron Kaci
Agron Kaci 5 年 前
If p is smaller or equal to 0.05 you must reject the null hypothesis, if your alpha level (reference probability) is 5%. So, yes, reject it even if p=.05.
Badrudin Yasini
Badrudin Yasini 5 年 前
if p value is exactly 0.05, should we reject or accept the null hypothesis? thanx in a dvance
Gert Pere
Gert Pere 5 年 前
Thanks Mr. Kaci !! This was very useful to me :) Best wishes from Albania!
khym france
khym france 5 年 前
Hi Sir, May i firstly thank you for such a clear and concise lecture on correlation and regression and secondly for some advice. I am currently comparing 4 indep variables to 1 depen to test correlation and which has highest sig. However all four have resulted in .000 sig. How do i determine which is now more statistically significant between the four? Your guidance would be much appreciated.
Ray LK
Ray LK 5 年 前
Excellent explanation thank you so much!
MY Fun
MY Fun 5 年 前
My friend would you please tell me when we use correlation and when using Regression? as far as I know correlation is to examine the relation between IVs & DV but then what the regression is for? please simplify the answer as am not from statistic background.
manglien singson
Thank you very much, it was very helpful
transformeris7 5 年 前
great video! :) Helped a lot :)
smallgold88 5 年 前
Sir, how to test the intercept and slope estimates for statistical significance at 1% significance level ?? (by refer to a t-test table ) i not understanding bout tis...can u explain for me ?? what is F-statistic ?? thanks
Nishantha Priyananda
Dear sir, Can you explain how we can analyze liker t scale questionnaire and use the result seeking mean,median,and standard divination including relation ships.thanks a lot sir
Agron Kaci
Agron Kaci 5 年 前
The p-value represents the chances that your Null Hyp is true. The smaller the p-value the less likely it is that your null will be true; therefore you reject it if the chances are less than 1 in 20, or 5 in 100 (hence 5%, or alpha 0.05) that it is true. P-value larger than 5% (this reference probability is chosen in advance - depending on the field or research, or on the seriousness of your work and test) show a greater likelihood that the null would be true, that's why you do not reject it.
Agron Kaci
Agron Kaci 5 年 前
You're welcome!
Agron Kaci
Agron Kaci 5 年 前
You're welcome!
Agron Kaci
Agron Kaci 5 年 前
Vivien Chai
Vivien Chai 5 年 前
Hi, so if my P value is more than 0.05, we keep our hypothesis? So if mean to say my P value is 0.000 shown in the output in SPSS, I have to reject my hypothesis? eg. Hypothesis : X and Y have a positive relationship. If after Regression, my p value is 0.000, I have to reject my hypo and my hypo should be : x and Y does not have a positive relationship? Is that right? Please help me! :) Thanks in advance!
Agron Kaci
Agron Kaci 5 年 前
Your comment is very much appreciated! I'm glad you have liked it.
MY Fun
MY Fun 5 年 前
Hi frankly speaking and allow me to say that your video is the most valuable, clear and simple video on YouToub. > simple and clear language > focus on the important areas (why, when, how..) > slow enough to understand and absorb > good example KEEP IT ON YOU ARE GOOD AT IT MAN (THUMB UP)
Sam Lai
Sam Lai 5 年 前
thank you. very useful for my thesis.
Anika Henry
Anika Henry 5 年 前
Great...Stats made easy :-)
Swapnil mehta
Swapnil mehta 5 年 前
stats simplified...
namechosen1 5 年 前
Agron Kaci
Agron Kaci 5 年 前
Thank you!
Kinful Lartebea Aryee
Great lecture