Multivariate Analysis
Guillaume, Knippenberg, and Brodbeck (2014) theorize that social classes or groups are an important starting point to determine the effects of societal diversity on individuals’ attitudes and behavior. People with similar backgrounds whether social, economic, or cultural are more comfortable and more likely to perform better both psychologically and behaviorally (Guillaume et al., 2014). Besharov (2014) states organizational identification occurs when members place a high value on organizational membership and define themselves accordingly. Additionally, creativity, helping behaviors, job satisfaction, and social support are other actions, which contribute to defining classes or groups (Besharov, 2014).
Organization identification benefits are fleeting when member sub-groups develop identities and values opposite of those of the main-group (Besharov, 2014). This mutual dis-identification of values can result in conflict and degraded organizational performance (Besharov, 2014). Besharov (2014) identifies the formation of sub-groups follow 1) inconsistent organizational identities and values or 2) individual differences of what values are important.
The research question follows:
How does the location of an individual’s alcohol consumption affect the development of mutually exclusive household values influence medical, diet, and housing consumption behavior?
The outcome of this research is to properly identify the target population for government information and educational campaigns based on alcohol consumption patterns and group values. The long-term objective is to improve the health of individuals by deglamorizing alcohol, improving diet, decreasing health costs, and offering alternative housing.
This research will use The Expenditure and Food Survey (EFS) data set. Mazzocchi (2008) describes the EFS dataset as a subset from the 2004 – 05 UK Expenditure and Food Survey. The file includes a random sample of 500 households and 420 variables out of the 1952 in the officially released data set (Mazzocchi, 2008). The UK collects consumer information from each person living in a household above the age of 16. Each participating person maintains a record of expenditures and then integrates the information into a consolidated household survey. This includes the collection of socio-demographic information, data on regular household bills, large items of expenditures and ownership of consumer durables to name a few (Mazzocchi, 2008). The UK targets and distributes the questionnaire to over 12,000 households based on postal codes. Consumer errors reduce the sample to 7,000 households (Mazzocchi, 2008).
Codebook
The codebook used in analyzing the data is illustrated below:
In answering the research question, the study focused on three main aspects. Firstly, the study aimed at establishing how household income affects alcoholic beverage/tobacco consumption. The objective was attained through the use of alcoholic beverage/tobacco consumption being dependent variable and the household income being the independent or explanatory variable since it attempts to predict the former. One-way ANOVA was used to establish the relationship between household income and the consumption of alcohol beverage and tobacco. In this analysis, the dependent variable was a group containing alcohol beverage and tobacco consumption whereas the independent variable was household income. The descriptive statistics for the two sets of variables are illustrated below:
The descriptive statistics for the consumption of tobacco are illustrated in the following table:
It is also important to obtain a descriptive statistics of household income, which is the independent variable in the analysis. The following table illustrates the descriptive statistics for the household income.
Performing ANOVA analysis on the effect of household on the alcohol beverage and tobacco consumption, the results obtained are as follows:
Instead of using the specific forms of alcohol beverage and tobacco consumption data, the ANOVA analysis used the EPS Total alcohol beverage and tobacco consumption amongst the participants. From the ANOVA tests analyzing the effect of household income on the consumption of alcohol, the p-value of the F-test under the p352 (which represents the Gross Current Household Income) is greater than 0.05 (0.21 > 0.05), which means that there effect of household on the consumption of alcohol is significant. Therefore, from the ANOVA tests, the study concludes that there is very little impact of the gross current household income on the consumption of alcohol beverage and tobacco as well. It means that household income as illustrated in the above analysis significantly impacts on the alcohol beverage and tobacco consumption.
Secondly, the study aimed at establishing how the socio-economic level affects the location of alcoholic consumption. This objective aimed at identifying whether the individuals drink more at home if they are poor or rich, whether the individuals drink more at the pub than home if they are rich or more, or whether they drink more if they are a poor female or male than a rich female or male. The attainment of the objective was through the use of MANOVA with the independent variable being a grouping variable of household type, ethnicity, and gender whilst the dependent variable was the consumption of alcohol or tobacco. The results of the MANOVA for both the alcohol and the tobacco usage are illustrated in the following tables:
On the basis of the above results, the p-value is less than the significance level value, that is, p < 0.05, which shows that the effect of gender on alcohol beverage consumption is insignificant. In other words, the consumption of alcohol beverage away from home is not determined by the gender of the individual.
With respect to age, all the tests apart from the Roy’s Largest Root are significant, that is, p-value > 0.05, which shows that on average the age of the household defines or affects the consumption of alcohol beverage.
Since the household income had been defined in the earlier sections, the above results illustrate a MANOVA analysis of household income on the consumption of alcohol beverage. From the result, it is clear statistically that the household income does not affect the type and where the consumers drink. Therefore, with the help of SPSS, it has become easier to perform an analysis on the relationship between numerous dependent variables and numerous independent variables. The use of multivariate analysis of variance has assisted in understanding how two or more independent variables explain two or more dependent variables as illustrated by the results shown in the above sections.
References
Besharov, M. L., (2014). The relational ecology of identification: How organizational identification emerges when individuals hold divergent values. Academy of Management Journal. 57(5), 1485-1512.
Guillaume, Y, R, F., Knippenberg, D, V., & Brodbeck, F, C., (2014). Nothing succeeds like moderation: A social self-regulation perspective on cultural dissimilarity and performance. Academy of Management Journal. 57(5), 1284-1308.
Mazzocchi, M. (2008). Statistics for marketing and consumer research. Thousand Oaks, CA: Sage. 2008. ISBN 9781412911221
