- Write 1 to 2 well-developed and research-supported paragraphs discussing the importance of statistical software in marketing research data analysis.
The first task of selecting methods of marketing research is to introduce a single method that can be used in collecting and analyzing marketing information. Then, taking into account the resource potential of the most appropriate set of these methods.
First of all, we give the general description of how to conduct market research.
The most widely used methods of market research are the methods of analysis of documents, methods of consumer research (the totality of which with a certain degree of conditionality can call the methods of sociological research, since they were first developed and used social scientists), expert evaluation, experimental methods and mathematical economics.
The main difference between the methods of sociological research on expert judgment is that the first focused on the mass of the respondents are very different competencies and skills, while expert assessment - to a limited number of professionals. What these two groups of methods primarily the fact that in both cases the processing of the collected data are used the same statistical methods.
Several groups of economic and mathematical methods used in marketing research:
- Statistical methods for data processing (definition of average estimates, the values of errors, the degree of agreement of the respondents, etc.
- Multivariate methods (primarily the factor and cluster analyzes). They are used to support marketing decisions. They are based on analysis of multiple interrelated variables. For example, the determination of the volume of sales of a new product based on its technical level, prices, competitiveness, costs for advertising, etc.
- Regression and correlation methods. They are used to establish relationships between groups of variables, statistically describe marketing activities.
- Simulation methods. They are used when the variables affecting the marketing situation (for example, describing the competition), cannot be defined by means of analytical methods.
- The statistical decision theory (game theory, queuing theory, stochastic programming) are used to describe the stochastic consumer response to changing market conditions. We can distinguish two main areas of application of these methods: for statistical testing of hypotheses about the structure of the market and assumptions about market conditions, such as research loyalty to the brand, forecasting market share.
- Deterministic Operations Research techniques (primarily linear and non-linear programming). These methods are used when there are many inter-related variables and the need to find an optimal solution, such as the delivery option of the product to the consumer to maximize profits, one of the possible channels of goods.
- Hybrid methods that combine deterministic and probabilistic (stochastic) characteristics (eg, dynamic and heuristic programming), are primarily used for the study of problems of goods movement.
These seven groups of quantitative methods, of course, do not exhaust all their diversity.
- Review and interpret the SPSS-generated cross-tabulation results and write 1 well-developed paragraph analyzing the appropriate components of the output and draw inferences from the data for CarCare, the organization conducting the research.
According to the Cross-tabulation output we see, that there were 50 observations(cars) considered, 24 of those were cars from Germany, 26 from USA, and there were 9 import Germany cars, 15 domestic, there were 16 import USA cars and 10 domestic.
If we look at the frequency distribution table, we can make a conclusion, that the mean value of the amount the respondent spends on car care products is 358.3495, this value is quite close to the median value of 359.0576. The attitude to the new car is 13.2009 at average, with median of 12.8904
- Review and interpret the SPSS-generated t-test output results and write 1 well-developed paragraph analyzing the appropriate components of the output and draw inferences from the data for the organization.
As we can see, the independent sample t-test was performed, to compare means of two sets – Spend of Germany cars and Spend of USA cars. Equal variances assumed. According to the Levene’s Test for equality shows p-value equal to 0.785, hence, at 5% level of significance the null hypothesis about that the variances are equal will not be rejected, and we proceed with this assumption.
Since p-value is 0.337, we can say, that at 5% level of significance we have no enough evidence to state, that there is a significant difference in means of these two sets.
- Review and interpret the SPSS-generated ANOVA and write 1 well-developed paragraph analyzing the appropriate components of the output and draw inferences from the data for the organization.
In this question we compare means of two data sets – Att and Attnew, divided by Country groups. We can see that p-value for Att is low(<0.001), hence, there is a significant difference in means of Att of Germany cars and Att of USA cars. In this time, the p-value of Attnew for these sets is very large(0.827), hence, at 5% level of significance, there is no difference between the means.
This might be the evidence of the following fact: the attitude for new cars is approximately equal, regardless of country of car owners. But the attitude for the current cars is different for Germany owners and USA owners.
- Review and interpret the SPSS-generated multiple regression outputs and write 1 to 2 well-developed paragraphs analyzing the appropriate components of the output and draw inferences from the data for the organization.
In this question we constructed a linear regression – for prediction Spend variable based on Attnew variable. The results are very bad – R-square(coefficient of determination) is almost 0, the p-value of model significance is 0.571. We can make a conclusion, that there is almost no linear association between Spend and Attnew variables, and this model is useless.
The second regression is a multiple regression. The dependent(response) variable is Attnew, and independent variables are Att, Spend, Country. According to the Model Summary output, the model is a little bit better(higher R-square, adjusted R-square). The ANOVA p-value is 0.329 and the p-value of Att significance is 0.081. But generally, the model is still bad, and can’t be used for any normal forecasts. The analysis of coefficients gives us a hint, that the bivariative regression of Att and Attnew could be more useful.
- In conclusion, write 1 to 2 paragraphs discussing how your understanding of the application of statistical analysis has changed or developed through this course, and how you could envision this type of statistical data analysis being useful for an organization that you are familiar with.
Statistical methods play a very important role in market researches. As we can see from the completed assignment, these methods help us to understand if there any associations between market indicators and how to create forecasts to predict important data.
Argyrous, G. Statistics for Research: With a Guide to SPSS, SAGE, London, ISBN 1-4129-1948-7
Levesque, R. SPSS Programming and Data Management: A Guide for SPSS and SAS Users, Fourth Edition (2007), SPSS Inc., Chicago Ill. PDF ISBN 1-56827-390-8
SPSS 15.0 Command Syntax Reference 2006, SPSS Inc., Chicago Ill.
Wellman, B. "Doing It Ourselves: The SPSS Manual as Sociology's Most Influential Recent Book."pp. 71–78 in Required Reading: Sociology's Most Influential Books, edited by Dan Clawson. Amherst: University of Massachusetts Press, 1998.