Statistics 101: Multiple Linear Regression, Two Categorical Variables

[@jonathansiegel2818]

Brandon Folt is absolutely the best... the very best... teacher in fundamental statistics. What a gift you are, Brandon, to the world!

[@BrandonFoltz]

Very kind of you. I appreciate your kind words. However learners like you are the true gift. Pay it forward when you can.

[@americanbluediamonds]

I am also a KW real estate agent, your example is so excited to me and I can use to explain to my clients why certain area the price increases subtantially.

[@knightsky8378]

I am in the MS program in Applied Economics. Your videos has helped me so much, especially in further understanding econometrics. This is some really great stuff right here. Thanks again!

[@akanlunadewuyi2588]

Good morning sir, I finish college this year read statistics as option. I still do not understand statistics until i found your videos. It's really great. Thanks for good work and your passion to help. God bless you sir.

[@beckyhuber8137]

I am a graduate student in a stat class. I have followed Brandon Folz video as my resources. Thank you.

[@JoshFlorii]

SUCH a good video. Spent like 5 minutes googling this subject only to turn to youtube, and you have once again not disappointed.

[@zhehabeshascience3066]

you are the best teacher i understod for statistics via youtube

[@andresparradrumaths3651]

great! i'm a teacher from Colombia and your videos are very useful for my classes...thank you!

[@juanaa.2111]

Thank you for sharing your knowledge. I am using the IBM SPSS, but the values and meanings are the same. Thank you for including the interpretation of the outputs, that really brings it all together.

[@cathycirina-chiu6972]

i so appreciate your encouraging intro and wonderful teaching style

[@getitdone913]

I know this is an old video, but thank you so much for uploading it. It's so helpful!

[@Nico-jc7zr]

Thanks for the video. I have a masters in economics and I have to get into regressions again for work. This is a nice reminder of the basic concepts. Nicely illustrated too.

[@vanessachen2330]

Watching your videos help me to improved my grade in Stat class, thanks for your passion of helping and providing such great series on Statistics.

[@lisameretekristensen3181]

Dear Brandon, I want to thank you for providing such excellent explanations and examples of statistical scenarios and how to go about analysing and prepping analysis. The latter is especially tough to find in textbooks in my experience. Thank you.

[@LoizidesGeorge]

Brandon Thanks a lot for all your the videos in the statistics series. You really saved me hours and days! Whenever you are in Cyprus I will devote 1-2-3 days to you + a an old house in the mountains of Marathasa, returning your courtesy to share them! Just let me know 1-2 hours before you arrive at the airport! Regards Γ [ Loizides George ]

[@rajkumar-xr5im]

Great ..Good service to society ..Please keep doing..

[@ActionSportsExtreme]

I'm really happy I stumbled upon you!

[@BrandonFoltz]

*NEW* video is up! Part B will be up later this evening (US EST). Thank you all so much!

[@SuzanneAmsalem]

Thank you:))

[@pascaleyram]

Could you make a short video on logistic regression analysis?

[@BrandonFoltz]

I may have the first Logistic video up tonight :)

[@pascaleyram]

***** Thanks a lot

[@madhurjyadeka5569]

Hello Sir, I'm estimating apartment prices based on 5 factors and in your earlier videos where you took 3 independent variables there were 7 regression analysis. So in my model that I'm about to build where there are 5 variables... Do I need to take the variables and analyse them : 1 at a time 2 at a time 3 at a time And so on.. untill I find the best result ?

[@rasyimahramli5b125]

Such a great video with the clear explanation. Thank you for your good work.

[@wendypaul3910]

Still saving lives in 2021, thank you Sir for blessing us with your knowledge and time☺

[@BrandonFoltz]

So nice of you. Not sure about saving lives but I want you to have the chance to make the one you want.

[@wendypaul3910]

@@BrandonFoltz literally true, passed my quantative statistics exam, I do not need to resit, could not do this without you, thank you is not enough. Please be encouraged and please continue to produce these amazing videos. 👏🏾👏🏾❤️

[@mkme2358]

Great videos very detailed! For better learning... I would suggest tests questions and answers.

[@shreyanshishukla6309]

I found very difficult to understand statistical use in research. your video makes it easy to understand the basic concepts. Pease Keep Uploading videos

[@mecha_studio_official]

Hi Brandon! Thanks for the excellent videos. You are a great teacher! Just a suggestion. It would be good if you can provide the datasets for us to do our own practice using statistical softwares adopted by our college.

[@mostafalotfi1818]

Great tutorials. Enjoyed learning from your videos a lot.

[@fabio7621]

Great video Brandon! Please, keep it up, you are helping so many people, more than you can imagine! :-)

[@mesfintesfaye7943]

Thank you very much i found the tutorial very important and i like the way you explain, easly understood for bigginers like me . Thank you keep it up sharing is caring !!!

[@florramirez2740]

Your videos are amazing, Im love in it. Im actually seeing since I have a job to do, however, I would like to know the machinary work in order to reproduce the graphics and others stuffs. I will appreciate so much if you can teach us. Very cool videos!!

[@bunyaweepunch2785]

Thanks for your video. They help me a lot!

[@darshitparkhiya1223]

Thank you sir for giving such a great videos, can you please provide 100 row data which you have used in Video

[@kayyalo9621]

Your videos are very much helpful

[@simratahluwalia965]

Great work Brandon .. keep it up....

[@VSP4591]

Well done. Thank you.

[@kobic8]

again, so HAPPY I came across your page here, I do have a question though, what if the dependent variable is categorical, and has more than one option e.g. direction (N /E / W / S) ?

[@nehabhatt1285]

Great Video.

[@anooppaul1]

Thank you for Videos. It is very helpful to me

[@TheCablebill]

Thanks for the videos. I am finding them helpful. Around the 9:30 mark of this one, the scatter plot is showing regression lines for sqft~price for each of the four categories. The visual patterns clearly indicate that a generalized function to predict price from ft^2 and region is not represented well by the simple linear template in use. Specifically, their is a potential missing multiplier term. the region effect could be expressed as a modifier of the slope of the regression line for sqft~price in addition to affecting y-intercept. In other words, some region factor could also be included with the existing coefficient for the sqft term. I believe this would generate a more effective prediction model for this circumstance. So I suspect that I'm not the first to note this possibility, and my question is: What am I talking about? Is there an established technique for the type of curve-fitting I describe, and what is it called? Thanks again for a great lecture series.

[@BrandonFoltz]

TheCablebill An interaction between region and square-footage could be analyzed using other methods such as ANOVA / ANCOVA where one or more of the independent variables are nominal. There are also some coding techniques used to examine interactions. That is a bit beyond my goal for the video but a good point!

[@kasunpathirana9410]

good explanation

[@riddhirekhawat]

It would be really helpful if you upload some videos on sample survey. SRS, stratified, double sampling and all such.

[@jessecuster7246]

Just awesome. Thank you for you work.

[@jessecuster7246]

Jesse Custer your*

[@ThePmac14]

Thanks Brandon

[@divneetbagga8258]

Hi Brandon!! I wish to run the problem given in the video by myself so that i can tally the results. For the same, I would require the entire data In the video there are only 15 entries given. Thanks!

[@sheebee8398]

Outstanding! Thank you!

[@ProfessorJoaoArantes]

Dear Brandom, could you please make the dataset available for download? Thank you!

[@bensonmoima6872]

Dear Sir, thanks for the great work. Doing my bachelor thesis in Germany and this video has really come in handy. One problem though, I have failed to plot this data exactly like you did on an excel scatter plot. Sgft on x-axis & prices on y-axis but I have failed to implant the categorical variables (yes/no) . How did you do that sir? Did you use excel for it as well?

[@hamadsuleiman5177]

use SPSS

[@ayiteajavon1894]

Very helpful. Thank you!

[@sullainvictus]

Great videos, Brandon! They are very informative and easy to understand. I have a question regarding dummy variables. The method you outline in this video and the previous one (Part 4) seem to deal with changing not the slope of the line, but the intercept. What if you have a situation where a categorical variable changes not just the position of the line (the intercept) but also the slope of the line. For example, what if the relationship between sqft and price were somehow a negative relationship in exemplary school districts? Would this method capture that effect?

[@nkvd1000]

Dear Mr Foltz I would be grateful if you could post a link to the dataset that you used in the example. Many Thanks

[@BrandonFoltz]

+nkvd1000 Hope to do that soon on my blog.

[@helen4805]

Hi Brandon, did you ever post this dataset? I would love to use it in my program. i am following along programming everything in Matlab.

[@adityaravi1876]

@@BrandonFoltz , can you please share the link of the data sheet... it will help us to practice in Minitab.

[@yvonnessy]

Thanks for the videos! I'm learning a lot. I have a question here.. Can I do a linear regression if the dependent variable is a categorical variable? If yes, how can it be done??

[@BrandonFoltz]

Yvonne Szeto If the dependent variable is categorical regular multiple regression cannot be used. It will require logistic regression (I have a video series on that as well). There is binary, ordinal, and multinomial logistic regression depending on the structure of the dependent variable. My videos are about binary logistic regression. Hope that helps!

[@yvonnessy]

***** Thanks for the quick reply! I will check those videos out. :)

[@Lbanin]

Hi Brandon, would you paste here the title of the video you refer above as I'm currently running an analysis over categorical dependent and independent variables (all variables are categorical). Thank you so much!

[@danyouse409]

Did you post a link to the dataset? If so, I cannot find it. Thank you!

[@YogeshprabhuJ]

Awesome... You should do some on logistic regression too..

[@BrandonFoltz]

Thanks! Logistic regression is my next topic actually 😃

[@janaria1985]

okay I understand the concept of dummy variable but with example of the region with 4 variables how do you know which one to use. In this case you omitted East, how did you decide on that? confused on what to use and what to take out

[@brunofischer808]

Hello. You actually do not care. Pick the combination you prefer. When you getto interpret the results, you will get to the same conclusion, independently from the set of dummies you selected.

[@leahhazanovich2556]

Thanks! This is super cleae!!

[@LukasStammler]

thanks a lot for these superb lectures. I do all your examples in R and now I ask, if it is possible to get the home price dataset for the lecture Multiple Regression Part 5.

[@1985dv]

with example of the region with 4 variables how do you know which one to use. In this case you omitted East, how did you decide on that? confused on what to use and what to take out

[@mahendrabhattarai5903]

Same confusion here. Did you find the reason now?

[@RodrigoTechador]

It's completely arbitrary. You can choose whichever category you want to be your reference category. The results will be the same.

[@khaledabdu2]

can you do a video on confounding and interaction for medical examples?

[@moom-sey]

Thank you so much!!! your video help me so much :)

[@dinethprabash1001]

Thanks, if you can number your videos (index number) it would be more helpful. After downloading, its hard to figure out which video comes first.

[@vanishingtears]

Dear Brandon Foltz I have a data set of sales, advertisement then dummy variables (years, months and quarters)..How to find out which month was the most and least successful? what is annual growth, quarterly growth? Please provide help as to what approach we should use when we have such timeseries element incorporated in the cross section data in the form of time as dummy variables? How would our interpretation of the model will differ?

[@vardeh1]

Thanks Brandon. In the equation, how do we calculate the constant?

[@BrandonFoltz]

vardeh1 No problem! I go over the calculation and interpretation in Part B which I should have uploaded tomorrow evening/night. So check back then. Thanks for watching! :)

[@sibghaafzal247]

Hi! Thank you for this helpful video! I am a psychology stats student and have a question with regards to the way that variable are coded, as I am a little confused. Am I right in thinking that if you code something as 1, then in the regression equation, it will mean that the outcome is always higher for that variable in comparison to the variable coded as 0, which essentially means we can influence our findings based on how the variables are coded? I hope this can be clarified, as I feel like this is not necessarily the case but I am not sure why - thank you in advance!!

[@nabajyotidey5613]

So basically ordinal data instead of nominal data.Your observation is good.

[@sonallagad284]

How do you plot the scatter plot of the exemplary schools where the plot shows the red and blue colors for the diff values.

[@iamshauno]

Hi! Is it possible to do regression using an independent variable with 7 units and a dependent variable with 20 units? Or should both variables have the same number of units?

[@sofieseymour481]

It looks like in the surface plot at 13:20, the directions are labeled incorrectly. Shouldn't it be east, north, south, west? If not, could you explain, because this doesn't seem to match the data shown in the scatterplot at 9:33.

[@Big-guy1981]

Hi. Hi can we apply these great videos to predicting the outcome of a sports event, say a baseball game?

[@BrandonFoltz]

Hi! For a win/loss prediction you could utilize Logistic Regression since your outcome is binary. The reality is most betting organizations already do this. Once they figure out the probability they then adjust the payout odds. So I always just recommend going with the conventional wisdom unless you know something everyone else does not. :)

[@lettersforkumar]

how does the scatter plot look like if there are more than 2 continious predictor variables? in your example if we want to add age of house as predictor variable, where doest it lie on the plot?

[@vinopavankumarathasan6736]

Hi I want to do a statistical analysis with two independent variables (IV) and both are categorical and dependent variable is interval. I have chosen the multiple regression. Guide me whether my choice is right...

[@郭巧生-w4o]

what if the hause is at southwest or in the middle of the town?

[@MostHitMan]

郭巧生 make more variables SW NW ES etc

[@gauravms6681]

sir can u list the books which helped you in these videos(machine learning and statistics) please it would be very helpful

[@cococnk388]

Statistics for business and economics by David Anderson , Business Analytics by Jeffrey D.Camm , The Hundred-Page Machine Learning Book by Andriy Burkov Brandon shared this books in his recent live on youtube. Hope it helps.

[@srijaP1]

Hi, I have around 42 independent variable (genotypes )and 8 dependent variable (cognition score) with age and gender as co variate. However my dependent variable are positively correlated so I have done PCA and have 2 component now. What kind of statistical analysis I should do?

[@8625gaurav]

Thanks a lot...

[@hangnisa2176]

Could I know how to calculate b2 coefficient in multiple regression?

[@tanmaybhayani]

But what to do if we dont know the number of categories, and the number of categories is not fixed. eg:-model of a car, to predict car prices.

[@utopiasolutions8797]

In your next video the equations all have the same slope. How is it possible?

[@tiannadermody4761]

Good explanations but it would be good if you could provide the code for these scatterplot outputs

[@BrandonFoltz]

Hello! Thanks for watching. The scatter plots were actually done in Minitab or JMP (It's been a while sorry) so there is no code to share. They are both traditional stats software packages.

[@varundeshpande3674]

Sir, as we haven't added any variable for East region how will we account for in any house is situated in the East?

[@xuchuan6401]

Every other region will be compared to East. East is the baseline

[@varundeshpande3674]

@xuchuan6401 helped 👌

[@burhankl2331]

***** Hi, one quick question , is it possible to perform regression analysis on ONLY categorical variables? in other words-can one perform a regression analysis if all the independent variables are categorical?

[@xuchuan6401]

Yes, and this will be equivalent to ANOVA

[@yegonb]

I have enjoyed your lessons, but I could not reach you on tweeter under your handle. I am doing a research and I would like to get your perspective on a few things.

[@adazeeviohana3495]

thanks for the clear videos! but what about the subtitles, there are tons of mistakes, seems that whoever did them did not really listen to what was said and made no effort to do a good job... sometimes actually really funny !

[@RodrigoTechador]

The subtitles are automatically generated by Google's voice recognition technology.

[@polomarco1256]

How to know the minimum amount of sample from huge population i.e. a nation?

[@mercygeorge2961]

why n-1?

[@x_kingsas-_-]

The South region has the steepest form not the west

[@davids8347]

I fail to understand how a university class that you are paying hundreds of dollars to be in can take 1.5 hours to complicate and make confusing a topic that a free KZbin video can take 15 minutes to clearly explain... 🤦‍♂

[@janaria1985]

sorry I get it now