Why do Mallet Mounié Why do people buy...
Transcript of Why do Mallet Mounié Why do people buy...
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
Why do people buy shoes?And what are the differences beetween boys and girls?
Bérénice Mallet - Floriane Mounié
Econometrics : Application Project
May 21st 2014
1/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
Introduction
What individual characteristics can influence our shoespurchases ?
2/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
IntroductionFirst steps
We had to :1 Choose some individual characteristics that might
influence shoe purchase
2 Realise a survey3 Create a data base for Eviews with the survey answers
3/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
IntroductionFirst steps
We had to :1 Choose some individual characteristics that might
influence shoe purchase2 Realise a survey
3 Create a data base for Eviews with the survey answers
3/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
IntroductionFirst steps
We had to :1 Choose some individual characteristics that might
influence shoe purchase2 Realise a survey3 Create a data base for Eviews with the survey answers
3/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
IntroductionExplanatory variables
Explanatory Variables Expected EffectAge =Friends +Store =Budget +Shoe Budget +Couple -Job +Children -Facebook +Pairs +Sex -Sales +Events +Clothe Budget +Zone + 4/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
First modelA linear model : First Equation
At first, we tried to modelise shoe purchase with lineardependance on all other variables.
EQU 01
purchasei = β0 + β1agei + β2friendsi + β3storei + β4budgeti
+β5shoebudgeti + β6couplei + β7jobi + β8childreni
+β9facebooki + β10pairsi + β11sexi + β12salesi + β13eventsi
+β14clothebudgeti + β15zonei
5/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
First modelLinear Regression Output
6/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
First modelReal effects
Explanatory Variables Expected Effect Real EffectAge = -Friends + -Store + -Budget + =Shoe Budget + +Couple - +Job + +Children - +Facebook + -Pairs + +Sex - -Sales + -Events + +Clothe Budget + +Zone + + 7/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
First modelR-squared
For this estimation, R-squared has a value of 0.562We will try to improve this by not taking into accountsome redundant characteristics.The correlation matrix reveals that some characteristicsare strongly correlated, and so redundant.
8/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
First modelCorrelation matrix
9/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
First modelCorrelated variables
Two variables were removed from the model :Age was to much correlated to budgetChildren was also correlated to budget
10/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
First modelWald Test
Some variables might not have a significant influence on shoepurchase. We performed a Wald test on the nullity hypothesisof the βk .
11/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
First modelInsignificant Variable
A high p-value means that the nullity hypothesis is notrejected. Four variables are removed of the model :
StoreFacebookZone
12/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
First modelSecond Equation
Here is the new linear equation :
EQU 02
purchasei = 0.158 − 0.0004friendsi − 0.0001budgeti
+0.004shoebudgeti + 0.304couplei + 0.236job + 0.069pairsi
−0.0337sexi − 0.0001isales + 0.063eventsi
+0.0006clothebudgeti
13/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
First modelNormality, Quantile Quantile Plot
14/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
Women’s ModelDistinct Population
One of the reasons our model is not exactly accurate migth bethat boys and girls have a very different behavior for shoespurchase.
There are two different populations, so we need to build adifferent model for women !
15/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
Women’s ModelDistinct Population
One of the reasons our model is not exactly accurate migth bethat boys and girls have a very different behavior for shoespurchase.
There are two different populations, so we need to build adifferent model for women !
15/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
Women’s ModelLinear Regression
Using the same variable, another linear regression wasperformed focusing on women population. Using the samemethods,
The variables Age, Children and Clothebudget areremoved because of correlation.Store, Facebook, Couple and Zone are removed becausethey are insgnificant.
16/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
Women’s ModelWomen’s Equation
Here is the final linear equation :
EQU 03
purchasei = 0.0187 − 0.0011friendsi − 0.00001budgeti
+0.0083shoebudgeti + 0.524job + 0.0509pairsi
−0.0072sales + 0.0917eventsi
A few remarks :Of course, women buy a lot more shoes than menGirls are less sensible than boys to budget constraintThey buy a higher proportion of their shoes during sales.
17/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
Women’s ModelHeteroscedasticity
The White’s test result is that we do reject theheteroscedasticity hypothesis. (p > 0.05)The model can be considered homoscedastic.
18/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
Women’s ModelError
The Jarque-Bera’s test result (p > 5 % and Stat < 6)indicates that the residual error can be identified as following anormal law.
19/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
Women’s ModelQuantile Quantile Plot
Comparison with normal law :
20/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
Conclusion
Conclusion
The final model reveals that the more influcial charateristics onwomen’s purchase are :
her budget for shoeswether she has a job or notthe number of pairs she already ownsthe number of activities an leisures she have
But more surprisingly neither the place of living nor the beingin a relationship have an impact on shoe purchase.
21/22
Why dopeople buy
shoes?
Mallet Mounié
Introduction
First model
Women’sModel
ConclusionThank you for your attention !
22/22