Multiple regression in SPSS procedures and interpretation (July 2019)
HTML-код
- Опубликовано: 1 июл 2019
- This video provides a walkthrough of how to carry out multiple regression using SPSS and how to interpret results. Included is a review of assumptions and options that are available for evaluating assumptions and identifying potential outliers and influential cases. A copy of the Powerpoint referenced in the video can be downloaded here: drive.google.com/open?id=1Rvn.... A copy of the dataset can be obtained here: drive.google.com/open?id=1ELu... . If you find the material and useful, please "like" the video and share! Thank you for watching!
Thank you, this is great. Clear, concise and yet detailed.
Oh my gosh I needed this so much. I had to redo all my results and I'm trying to relearn how to use spss so I can run a multiple regression analysis. This was so helpful.
Thank you sooo much for such a helpful video and your consideration to upload the slides. They are very helpful!
Excellent work with detailed explanation. Keep it up! Thank you!
This is great. Your channel is my go-to for any statistics problems. Thanks for your work :)
Hi Britt, thanks for watching and for your kind feedback! Best wishes!
Best video I’ve seen explaining multiple regression analysis. I have watched so many but this is the best this far. Thank you
Glad it was helpful, Racquel! Best wishes!
Excellent video, presentation, PowerPoint presentation, and SPSS data file. I am highly benefited. Thanks a lot.
Great work! It's really informative and clear. Thanks for sharing.
I love the way you teach. You make things very easy. thanks for your kindness
You are so welcome, Kingley! Cheers!
Your efforts are really commendable. Thank you
Excellent presentation and explanation.
thank you so much for this VERY helpful video.
Thank you Dr Crowson!
Thank you so much. God bless you! Thank you
It's very helpful. Thank you
great work and awesome contribution.
Hi Nasir, thank you so much for your feedback! Thanks for visiting!
thank you this has been very helpfull
Woow..Thanks for a clear presentation sir..!
Hi there. Thanks for visiting! Best wishes.
very useful indeed. thanks for sharing. GS Bawa
Hi everyone, I have just uploaded a new video (Sept 2021) on multiple linear regression using SPSS. I hope you consider visiting it at ruclips.net/video/0N4Q8zZijcI/видео.html .
Thank you so much. it is very benefical.
You are very welcome, Mehmet!
thanks a lot
very good
VERY gOOD PRESENTATION
In my case: the standardized coefficients beta and the square of the part correlations give me 2 quite different rankings of the dependent variables
Great lecture!
Hey, thanks Yulin! I appreciate the feedback!
@@mikecrowson2462 You're very welcome! Thank you for the generous sharing of knowledge!
Thank you so much
You are very welcome!
thumbs up
Hello Sir,
What if the variables are measured on different point likert scale. For example 5 point and 6 point. How to run a multiple linear regression in such a case.?
Hello 👋🏻 I’ve 1question to ask. As a beginner I currently run the spss. My hypothesis is to find “effect. From my knowledge regression test can do it. But, my data not normally distributed. Can i use regression to find my hypothesis with data not normally distributed?
How about there are more than one dependent variables? Shall I run Analyze>Regression>Linear for each dependent variables as much as number of DVs?
Same question
Hey Mike
nicely explained
Can we use regression for scale data for dependent and independent Variables are in scale... if Yes what will be the way to do so
Hi there. When you say 'scale ' data are you referring to the measurement setting in SPSS? That setting is sort of a catch-all for interval and ratio-level variables. You would use this setting if you are assuming your variable is continuous.
Least squares regression assumes that your dependent variable is continuous. And generally your IV's are continuous. [It is possible to model categorical independent variables via the use of some type of dummy or effect coding system.] So running the analysis in SPSS with your variable set to 'scale' is the same as what I demonstrated in the video. Cheers!
@@mikecrowson2462 Thank you so much Mike
How do you obtain the values of the independent variable..
I mean those numbers under interest and the other independent variable
Good job and best wishes...
How possible to get the power point that u mentioned in the video?
Hi there. Thanks for visiting! There is a link to the PowerPoint underneath the video description. Cheers!
So as far as I understood you can just mark a categorical variable as a continuous one in order to include in the model. What's the point of logistic regression then?
Hello, thanks for visiting and for your question.
First off, logistic regression is used when the dependent variable is binary and your goal is to predict the probability of a case falling into a target group (as opposed to a reference group) as a function of the predictors in your model. [there are actually other forms such as multinomial and ordinal, but I'm sticking to the most commonly used logistic regression model in this discussion]. OLS regression is used to predict a continuous dependent variable.
With respect to your statement about 'marking a categorical variable as continuous for inclusion in the model', I would point out that the decision to do this does not convert your previously 'categorical' variable into a continuous one. SPSS allows you to change how you indicate the scaling of your variable under the Variable View tab, but whether you decide to have the indicator set to nominal, ordinal, or scale will not impact the mathematics behind the computation when running your regression. The program will give you the same OLS regression output when you perform your analysis with the variables listed in either way. I think of that option in SPSS as largely something that encourages folks to think about the scaling of their variables.
Now, if you are asking about the decision by a researcher to treat an ordered categorical variable "as if" it was continuous (such as those cases where a researcher might treat a Likert type item, ranging from 1=strongly disagree to 7=strongly agree) in a model, then that is something different. Generally, the decision to do something like that is based on a variety of theoretical and pragmatic considerations, as well as the possible effects of doing this when it comes to meeting model assumptions. I'm not going to go into all the finer points here; but I will say that OLS regression technically makes no assumptions regarding the distributional characteristics of the predictors in the model. The only distributional assumptions pertain to the residuals (prediction errors).
I hope this is helpful to you. Cheers!
p.s., I do have a newer video on OLS regression here: ruclips.net/video/0N4Q8zZijcI/видео.html
I sure wish you were my neighbor
thanks a lot, can I have your email plz