The determinants of green consumption: a study of socio-demographics factors as determinants

3.3.3 Hypothesis 1: socio-economical characteristics have a positive effect on consumers buying decision of green product

For this hypothesis the researcher has established the null hypothesis as:

- H0 = the socio-economical characteristics are not explaining the consumption of green products

- H1 = the socio economical characteristics permit to explain the consumption of green product

After implemented the hypothesis, three boards were obtained, those tables would permit to determine if the independent variable (socio-economical characteristics) has an effect on the dependent variable.

Table 3.5 H1 Model summary

Récapitulatif des modèles*

Modèle		R		R-deux	R-deux ajusté	Erreur standard de l'estimation
dimensio n0	1		,111^a	,012	,006	1,09501

a. Valeurs prédites : (constantes), socio

*Model summary table translation. R-deux means R-square and R-deux ajusté, R-square adjusted. The last column means standard mistakes according the estimation.

As the researcher was using a French version of SPSS, all the different tables are in French. A translation is provided for each table.

Firstly, the summary model table has to be studied. In this table, the most interesting indications are the R and the R square (=R-deux). The first, R, represents the simple correlation between the two variables, in our case it is 0,111, which indicates a low degree of correlation; the correlation is strong when it's close to 1. (Laerd Statistics 2007)

The R-square (R-deux) refers to the proportion of variance in the dependent variable (green consumption) which can be explained by the independent variables (socio-economical characteristics). «This is an overall measure of the strength of association and does not reflect the extent to which any particular independent variable is associated with the dependent variable». (UCLA University 2007)

In this particular case, R-square is equal to 0,12 this means that only 12% of the variance of green consumption could be explained by the socio-economical characteristics; therefore it is not really important.

That's why the researcher has divided the socio-economical characteristics in order to see which one is affecting, or not, the consumption of green products.

Table 3.6 H1 ANOVA Table

ANOVA^b**

Modèle	Somme des carrés	ddl	Moyenne des carrés	D	Sig.
1 Régression Résidu Total	2,214 177,460 179,673	1 148 149	2,214 1,199	1,846	,176^a

a. Valeurs prédites : (constantes), socio

**ANOVA Table translation: the first column means sum of squares, the third one is mean of the squares and the following one is F in English.

The second table is the ANOVA table; it refers to the analysis of the variance. To be relevant, the improvement obtained with the independent variable must be large and the residual between the observed and the regression line, low. (Eric Yegereau 2009) Therefore we can observe that the part of variance none explain by the independent variable is much more important, 177.46, than the part explain by the independent variable, 2.21. So it seems that the socio-economical characteristics don't have an effect upon the green consumption.

In this case, F (=D) is 1.846 and we get p-value = 0.176 > 0.05, in other words, at the p = 0.05 level of significance, there exists enough evidence to conclude that the slope of the population regression line is close to zero and, hence, that socioeconomical characteristics isn't useful as a predictor of green consumption. (Statistical Sciences and Operations Research; 2010)

So there isn't a statistically significant relationship between the dependent variable and the independent variable.

We can conclude that the model with a predictor, soio-economical characteristics, doesn't permits to predict the variable, green consumption, better than a model without a predictor. (Eric Yegereau 2009)

Table 3.7 H1 Coefficients table

Coefficients^a***

Modèle			Coefficients
	Coefficients non standardisés		standardisés
	A	Erreur standard	Bêta	t	Sig.
1 (Constante)	4,048	,446		9,067	,000
socio	-,218	,161	-,111	-1,359	,176

a. Variable dépendante : green_consump

**Coefficient table translation: the first column means unstandardized coefficients with A and standard error. The second column means standardized coefficients. The last two columns remain the same.

The last table permits to see the relative importance of each independent variable to the dependent variable and to draw the regression equation.

For this hypothesis, the regression equation could be drawn as followed: Green consumption = 4,048-0,218*socio-economical characteristics.

The coefficients, also, permits to look at the p-value (=sig), we reject H0 if p< 0.05 (Jeff Sinn, 2008)

In this case, p = 0.176 therefore we get 0.176 > 0.05, as a consequence we can't reject H0 and we have to say that generally the socio-economical characteristics don't permit to explain the consumption of green products.

3.3.3.1 H1a: the gender has a positive effect on green buying

For this hypothesis the researcher has established the null hypothesis as: H0 = the gender is not explaining the consumption of green products

H1 = the gender has an effect on the consumption of green product

Table 3.8 H1a Model summary

Récapitulatif des modèles

Modèle		R		R-deux	R-deux ajusté	Erreur standard de l'estimation
dimensio n0	1		,344^a	,118	,113	1,03448

a. Valeurs prédites : (constantes), 2

For this hypothesis, we could observe that the correlation between the variables, gender and the consumption of green products is not really strong 0,344. Moreover, R-square is equal to 0,118 this means that only 11.8% of the variance of green consumption could be explained by the gender; therefore it seems that the gender is not, wholly, explaining the consumption of green products.

Table 3.9 H1a: ANOVA Table

ANOVA^b

Modèle	Somme des carrés	ddl	Moyenne des carrés	D	Sig.
1 Régression Résidu Total	21,291 158,383 179,673	1 148 149	21,291 1,070	19,895	,000^a

a. Valeurs prédites : (constantes), 2

b. Variable dépendante : green_consump

The part of variance none explain by the independent variable is much more important, 158.4, than the part explain by the independent variable, 21.3. So it seems that the gender don't have an effect upon the green consumption.

In this case, the D (F) value is 19.895 and is significant at p <0.0005. In other words, at the p = 0.05 level of significance, there exists enough evidence to conclude that the slope of the population regression line is not zero and, hence, that gender is useful as a predictor of green consumption. So there is a statistically significant relationship between the green consumption and the gender. However, according to the previous observations, it is not a strong relationship between those variables;

as a consequence it appears that gender doesn't have a strong effect upon green buying decision.

Table 3.10 H1a: Coefficients Table

Coefficients^a

Modèle			Coefficients
	Coefficients non standardisés		standardisés
	A	Erreur standard	Bêta	t	Sig.
1 (Constante)	2,358	,260		9,085	,000
gender	,717	,161	,344	4,460	,000

a. Variable dépendante : green_consump

For this hypothesis, the regression equation could be drawn as followed: Green consumption = 2.358+0,717*gender.

For the p-value, in this case p = .000 therefore we get .000 < 0.05, as a consequence we reject H0 and we have to say that the gender can explain the consumption of green products. However, as it was previously explained, there isn't a strong relationship between those variables; as a result, the gender doesn't seem to predict totally the consumption of green product.

3.3.3.2 H1b: the level of income or revenue is positively linked to

consumers green buying behavior

For this hypothesis the null hypothesis is:

H0 = the level of income is not explaining the consumption of green products H1 = the level of income has an effect on the consumption of green product

Table 3.11 H1b: Model Summary

Récapitulatif des modèles

Modèle		R		R-deux	R-deux ajusté	Erreur standard de l'estimation
dimensio n^0	1		,291^a	,084	,078	1,05424

a. Valeurs prédites : (constantes), 2

For this hypothesis, we can observe that the correlation between the variables, level of income and the consumption of green products is not really strong: 0,291. Moreover, R-square is equal to 0.084 this means that only 8.4% of the variance of green consumption could be explained by the level of income; therefore it seems that the level of income is not explaining the consumption of green products.

Table 3.12 H1b ANOVA Table

ANOVA^b

Modèle	Somme des carrés	ddl	Moyenne des carrés	D	Sig.
1 Régression Résidu Total	15,182 164,491 179,673	1 148 149	15,182 1,111	13,660	,309^a

a. Valeurs prédites : (constantes), 2

b. Variable dépendante : green_consump

The part of variance none explain by the independent variable is much more important, 164.491, than the part explain by the independent variable, 15,182. So it seems that the level of education don't have an effect upon the green consumption. In this case, the D (F) value is 13,660 and is significant at p <0.0005. In other words, at the p = 0.05 level of significance, there exists enough evidence to conclude that the slope of the population regression line is close to zero and, hence, that the level of income isn't useful as a predictor of green consumption. In this case, we keep the null hypothesis formulated above. So there isn't a statistically significant relationship between the green consumption and the level of income.

Table 3.13 H1b Coefficients table

Coefficients^a

Modèle			Coefficients
	Coefficients non standardisés		standardisés
	A	Erreur standard	Bêta	t	Sig.
1 (Constante)	3,965	,163		24,313	,000
2	-,217	,059	-,291	-3,696	,309

a. Variable dépendante : green_consump

For this hypothesis, the regression equation could be drawn as followed: Green consumption = 3,965-0,0.217*level of income.

For the p-value, in this case p = .0.309 therefore we get .0.309 > 0.05, as a consequence we keep H0 and we have to say that the level of income can't explain the consumption of green products.

3.3.3.3 H1c: the level of education is positively linked to the

consumption of green products

For this hypothesis the null hypothesis is:

H0 = the level of education is not explaining the consumption of green products H1 = the level of education has an effect on the consumption of green products.

Table 3.14 H1c Model Summary

Récapitulatif des modèles

Modèle	R	R-deux	R-deux ajusté	Erreur standard de l'estimation
1 dimensi on0	,104^a	,011	,004	1,09589

a. Valeurs prédites : (constantes), 2

For this hypothesis, we could observe that the correlation between the variables,
level of education and the consumption of green products is not strong at all:
0,104. Moreover, R-square is equal to 0.011 this means that only 1.1% of the

variance of green consumption could be explained by the level of education; therefore it seems that the consumption of green products is not dependent of the level of education.

Table 3.15 H1c ANOVA Table

ANOVA^b

Modèle	Somme des carrés	ddl	Moyenne des carrés	D	Sig.
1 Régression Résidu Total	1,930 177,743 179,673	1 148 149	1,930 1,201	1,607	,207^a

a. Valeurs prédites : (constantes), 2

b. Variable dépendante : green_consump

The part of variance none explain by the independent variable is much more important, 177.743, than the part explain by the independent variable, 1.930. So it seems that the level of education don't have an effect upon the green consumption. In this case, the D (F) value is 1.607 and is significant at p <0.0005. In other words, at the p = 0.05 level of significance, there exists enough evidence to conclude that the slope of the population regression line is close to zero and, hence, that the level of education isn't useful as a predictor of green consumption. In this case, we keep the null hypothesis formulated above. So there isn't a statistically significant relationship between the green consumption and the level of education.

Table 3.16 H1c Coefficients Table

Coefficients^a

Modèle			Coefficients
	Coefficients non standardisés		standardisés
	A	Erreur standard	Bêta	t	Sig.
1 (Constante)	3,873	,343		11,288	,000
2	-,126	,100	-,104	-1,268	,207

a. Variable dépendante : green_consump

For this hypothesis, the regression equation could be drawn as followed: Green consumption = 3.873-0,126*level of education.

For the p-value, in this case p = .207 therefore we get .207 > 0.05, as a consequence we keep H0 and we have to say that the level of education can't explain the consumption of green products.