Question 1.

Which is the dependent variable in this analysis and what type does it have?

Purchase intention is the dependent variable in this analysis and it is ordinal data.

Question 2.

What is the average purchase intention among your respondents? Is the distribution of variable skewed or symmetrical?

Testing the degree of skewness is important to see whether the distribution of the data is       symmetric and can be used for further analysis. As we can see from the table below the        skewness of the data is equal to 0.096, which is close to 0 so the distribution of data is          considered to be reliable and symmetric. In order to draw a conclusion off the opposite, the  data would have been negative or positively skewed,  it had to deviate outside the -0.5 to 0.5

range.

Descriptive Statistic table (Question 2)

In the frequency table below we see that the data is bell shaped and symmetric. Hence the average purchase intention is assumed to be normally distributed.

Histogram  (Question 2)

Question 3.

Is there a difference between the purchase intention of men and women?

In order to find out if there are any differences between the purchase intention between men and women, we first need to make a Levene’s test in SPSS. This will be done by comparing the nominal variable “gender” (male or female) with the ordinal variable “purchase intention” . When entering these two variables into SPSS, we get the following outputs;

Group Statistics (Question 3)

By looking at the first output from SPSS we are able to see that the average answer from       females is 4.02. Male participants' average answer was 3.68, so by doing a quick comparison of the mean from the two sample groups (female and male), we are able to see that there are in fact some differences in their purchase intention regarding meat substitutes.

Levene’s Test (Question 3)

(Note! The last part, t-test for Equality of Means, is connected to the picture above with the Levene’s test. The SPSS output was too wide to fit properly, hence we cut it in half.)

With the above given Levene’s test, we are also able to investigate whether or not the two groups share their purchase intention. From SPSS we were given a Levene’s test with a    significance level of 0.209, which is more than 0.05, i.e 0.209 > 0.05, meaning that the     variance between the females and males is not significantly different.

Question 4.

Definition of a null hypothesis and an alternative hypothesis for the relationship between purchase intention and gender of the respondent.

Ho : The gender of the respondents has no significant impact on their intention to purchase meat substitutes

H1: The gender of the respondents has a significant impact on their intention to purchase meat substitutes

Question 5.

A linear regression to test our hypothesis.

Variables Entered/Removed and Model summaryfor Question 5

From the first table we can see that the “intention to purchase meat substitutes the next 4 weeks” is the dependent variable that depends on the independent variable “gender” .

Model Summary (Question 5)

The model summary is used to see how much of the model can be explained due to the variance in observed data. The simple linear correlation R is 0.08 which can be interpreted as a low positive degree of correlation between the DV (Purchase intention to buy meat substitute the next 4 weeks) and the IV (Respondents gender). R2 value illustrates the proportion of variance in the DV which can be explained by the IV. The reason for why adjusted R2 the same is that we have not tested multiple variables in the regression model yet.

As we can see from the table this is a very low result and equal to 0.006 or 0.6%. Hence ,  0.6% of the variability in the “purchase intention to buy meat substitute the next 4 weeks” can be explained by the “gender” .

Anova Table (Question 5)

The anova table provides us necessary information of the overall significance that the model has, and if we could make inferences about the population or not. In this case from the table above we see that the 0.2 % chance that we are wrong if rejecting the null hypothesis, which is a very significant number in hypothesis testing.

If we for example using a = 0.01 : At 1 degrees of freedom in the numerator and rounding upwards using ∞ degrees of freedom in the denominator, the critical value is  6.635

Therefore we find evidence rejecting H0, since 9.897 > 6.635.  and the p value  < alpha=

0.01.

Conclusion: We are 99% confident that the gender role has a significant impact on purchase intention to buy meat substitutes coming next 4 weeks in the population (in this example      data), hence we reject H0.

Coefficients Tablefor Simple Linear Regression (Question 5)

The (constant) is the point where the linear regression line intersects the Y-axis which in this case is 4.361. Unstandardized B tells us that if IV gender increases by one, 1→ 2 (i.e. from   woman to male) the “Purchase intention for buying meat substitute” is decreasing by approximately 0.339.

The standardized regression coefficient which is -0.08, we can do a similar interpretation. However, as the name reveals, it is a standardization of the variables into z-value, so the average becomes 0 and standard deviation is one. It is used to analyze multiple variables with differential units or scales which we will cover more in the next question. But the standardized regression coefficient -0.08 is the amount of standard deviation the dependent variable (purchase intention to buy meat substitute) is changing per whole increase of standard deviation of independent variable (gender).

Question 6.

Multiple linear regression three independent variables.

Below is the table (table 6) of variables entered.

Independent variables are social influence, gender and sustainability consciousness and we     will analyze their effect on the dependent variable which is purchase intention as we can see   in the table 6. Purchase intention regards the subjects intention to purchase meat substitutes in the coming 4 weeks.

Table 6. “Variables entered and removed”.

A) Analyzing and explaining output tables.

In the model summary (table 6. 1.) we can see that the adjusted R square is 0.223. The            adjusted R square shows us how much of the variance in the purchase intention of the sample is explained by the independent variables that were entered in the table 6.

In this multiple linear regression we can see that our independent variables explain 22.3% of the variance in purchase intention.

Table 6.1. “Model Summary” .

The ANOVA table shows the significance of our findings. The significance value is lower    than 0.001 which indicates that the findings are significant, since it is lower than all p-values in social science at 0.05, 0.01 and 0.001. Therefore we would reject the null hypothesis in     favour of the alternative here. The null hypothesis would be that there is no impact of the      independent variables on the dependent variable of purchase intention. But what we found is that at least one of the factors of social influence, gender and sustainability consciousness     have an effect on the purchase intention.

Table 6.2. “ANOVA table”

In the table of coefficients (6.3.) we can analyze the independent influence and significance of every independent variable on the purchase intention. Dissimilar to the simple linear regression we will look at the standardized coefficient beta for the direct effect of a 1 unit   increase in independent variables on the dependent variable.

A 1 unit increase in gender, or a change from woman to man since woman is coded as 1 and  man as 2, affects the purchase intention by -0.063. We can therefore conclude that men in our sample have less intention to buy meat substitutes than women. However, this finding is not  significant on all levels since it is larger than 0.001. But for 0.05, 0.01, gender is a significant factor since 0.05 > 0.01 > 0.006.

Sustainability consciousness is significant since its p-value at < 0.001 (table 6.3.) is smaller than all significant levels. We can also see that social influence has an increasing impact on purchase intention. Every 1 unit increase in social influence increases purchase intention by

0.133.

Social influence is much like sustainability consciousness since it is significant at all levels and an unit increase also has a positive impact on the purchase intention of meat substitutes. A 1 unit increase in social influence increases the purchase intention by 0.44 (table 6.3.)      which is the largest impact of all factors investigated.

We can conclude that at the 0.05 and 0.01 significance levels all factors have a significant   effect on the purchase intention and we can therefore apply this on the population that we    analyzed. However gender is not a significant factor of the purchase intention on the 0.001  level and if we assume or conclude that gender has a significant effect, we risk conducting a type 1 error.

Table 6.3. “Table of coefficients”

B) Comparison of this model to the Q5

In comparison to the model in question 5, this multiple linear regression does a better job explaining the intentional differences.

First off we can see the difference between the R square in the model summary of question 5 and the adjusted R square from question 6. From the adjusted R square in table 6.1 we see    that 22.3% of the variance in the purchase intention could be explained by the independent   variables whilst the R square from question 5 was only 0.6%. So the multiple linear               regression covers much more of the variance in purchase intention than the simple linear      regression did.

Secondly, we concluded in the analysis of the analysis of the coefficients from table 6.3 that the significance of the impact of gender was only 0.006 whilst the significance of gender in  the simple linear regression was 0.002. Although none of these values are seen as significant at the 0.001 level, the difference is large. That difference in significant values could make us falsely reject a null hypothesis if we only used the simple linear regression model. But with the multiple linear regression model we could see that gender was not as significant as we would have thought from the analysis in question 5.

We can also see that the difference in purchase intention between the genders is not as great when we included two more independent variables. From the different tables of coefficients we get -0.339 (Q5) and -0.063 (table 6.3), and that difference is -0.339-(-0.063) = -0.276. So by adding two more variables into the regression model, the effect that gender has on            purchase intention of meat substitutes has decreased by -0.276.