Discriminant Analysis

Dependent variable: online buyer or none buyer

Independent variables: Education, Gender, Monthly Income, owning a Credit Card, E-banking, use of social media sites and Age.

The significance of univariate F ratios indicates that when the predictors are considered individually, only Gender, Credit Card, E-banking, Use of SNS and Age significantly differentiate between those who shop online and those who do not.

Because there are two groups, only one discriminant function is estimated. The eigenvalue associated with this function is 0.691 and it accounts for 100 percent of the explained variance. The canonical correlation associated with this function is 0.639. The square of this correlation, (0.639)2= 0.408, indicates that 40.8% of the variance in the dependent variable is explained or accounted for by this model.

Cluster Analysis

Depending on the reasons for a person to be online, consumers can be clustered into homogeneous groups.

The various attributes used in CLUSTER Analysis have been coded as follow:

V1: News or Information
V2: Websites of company regarding product
V3: Travel and leisure
V4: Spent time in social media sites like Facebook
V5: Online shopping sites such as Flipkart
V6: Education related sites
V7: Official works, email

The three resulting clusters can be described as follow:

Cluster 1: internet users who are Leisure Hunter (relatively high values on variables V1, V4 and V5)
Cluster 2: internet users who are Regular Web Person (medium values on the variables)
Cluster 3: internet users who are Dedicated Surfer (relatively high values on variables V2, V3 and V6)

Users can further be clustered on the basis of factors which influence them while making an online purchase:-

The various attributes used in CLUSTER Analysis have been coded as follow:

V1: Brand Name
V2: Service delivery time
V3: Website Content
V4: Recommendation by friends
V5: Online Ads – posters/banners
V6: Online reviews by users of product
V7: Ease of payment and security

The four resulting clusters can be described as follow:

Cluster 1: The Surgical Shopper (relatively high values on variables V4 and V6)
Cluster 2: The Enthusiast Shopper (relatively high values on variables V1, V2, V3, V5, and V7)
Cluster 3: The Casual Shopper (relatively high values on variables V1, V2, V3, and V7)
Cluster 4: The Reluctant Shopper (relatively low values on all the variables)

Factor Analysis

To find the major factors on which customer’s online buying characteristics can be loaded, Factor Analysis was done on a 5-point Likert scale (1= Strongly Agree and 5= Strongly Disagree)

The various attributes used in factor analysis have been coded as follow:

V1: I prefer making a purchase from internet than using local malls or stores
V2: I can get the latest information from the Internet regarding different products/services that is not available in the market
V3: Online shopping is more convenient than in-store shopping
V4: Online shopping saves time over in-store shopping
V5: It is safe to use a credit card while shopping on the Internet
V6: Online shopping allows me to shop anywhere and at anytime
V7: I trust the delivery process of the shopping websites
V8: Products purchased through Internet are of guaranteed quality
V9: Internet provides regular discounts and promotional offers to me
V10: Cash on Delivery is a better way to pay while shopping on the Internet
V11: Sometimes, I can find products online which I may not find in-stores
V12: I have faced problems while shopping online
V13: I continue shopping online despite facing problems on some occasions
V14: I do not shop online only because I do not own a credit card

Attributes loading on various factors/components:
Loaded on factor 1:- V5, V6, V7, V8,
Loaded on factor 2:- V1, V2, V3,
Loaded on factor 3:- V12, V13,
Loaded on factor 4:- V4, V9, V10, V11
Loaded on factor 5:- V14 (negative loading)

Depending on the eigenvalues >1, there are 5 resulting factors which respondents look for:
Factor 1: Trust
Factor 2: Convenience
Factor 3: Risk propensity
Factor 4: The Power Shopping
Factor 5: Neglect

Regression Analysis

Dependent variable: Average Amount spent per purchase made online

Independent variables: Frequency of Purchase of products and services online, Family Income, owning a Credit Card, Marital Status, Gender, Occupation, Education and Age.

The value of R square is quite low and so it can be said that the regression model does not fit into the data very well. Also, the sum of squares of regression is lesser than the sum of squares of residuals and this reiterates the findings of R square. This is because if the sum of squares of regression is lesser than the sum of squares of residuals, then the independent variables do not explain the variation in the dependent variable well. While cross tabs suggest a positive relationship between multiple pairs of factors, the linear correlation model, with all factors together, does not fit in with the outcomes.

Data Interpretations and Analysis

Cross-Tabs With Chi-Square

H1: Owning a credit card does not have any impact on the frequency of online purchase. 

As the p-value is lesser than 0.05, which is our assumed level of significance, we do not accept the null hypothesis, i.e. for the sample population, owning a credit card has an impact on the frequency of online purchase.

H2: Age of the respondent does not have any impact on the frequency of online purchase.

As the p-value is greater than 0.05, which is our assumed level of significance, we accept the null hypothesis, i.e. for the sample population, Age of the respondent does not have any impact on the frequency of online purchase.

H3: Gender does not have any impact on the average amount spent per purchase made online.

As the p-value is greater than 0.05, which is our assumed level of significance, we accept the null hypothesis, i.e. for the sample population, Gender does not have any impact on the average amount spent per purchase made online.

H4: Gender does not have any impact on the frequency of purchase of online products and services

As the p-value is lesser than 0.05, which is our assumed level of significance, we do not accept the null hypothesis, i.e. for the sample population, Gender has an impact on the frequency of purchase of online products and services.

H5: Income of respondents does not have any impact on the frequency of purchase of online products and services.

As the p-value is greater than 0.05, which is our assumed level of significance, we accept the null hypothesis, i.e. for the sample population, Income of respondents does not have any impact on the frequency of purchase of online products and services.

H6: E-banking does not have any impact on the frequency of online purchase.

As the p-value is lesser than 0.05, which is our assumed level of significance, we do not accept the null hypothesis, i.e. for the sample population, E-banking has an impact on the frequency of online purchase.