Multivariate Analysis

Simple explanations generally cannot be applied. Multivariate Analysis (MVA) is the complex of methodologies of multidimensional statistics enabling to better understand relations between particular questions. These methodologies do not study each surveyed fact separately but concentrate on determined quantity as a whole.

Benefits of multidimensional methods

Study of the list of questions as an indivisible complex brings new dimension in research. It includes new types of information we can get from the research. Understanding relations between particular questions offers solutions to following problems:

  • What matters people evaluate in similar way and which ones in the other way round?
  • What are the main motives guiding peoples’ opinions and attitudes?
  • Are there any segments with similar opinion and attitudes structure in the target group?

Extensive process of research outputs can MVA replace by short, well-arranged and summarizing output.

Practical examples of usage

  • Identification of main effects guiding customers’ behaviour.
  • Segmentation of customers in accordance with preferred type of product represented by list of characteristics, segmentation characteristic and representation in population.
  • Reinforcement (improvement) of a product in the areas preferred by customers.
  • Analysis how customers sense product or company in comparison with competition.
  • Decisions whether the individual approach to specific target groups pays off.

Multidimensional methods include several isolated analyses that can be used separately or can be combined in various ways.

Factor analysis

  • Helps to identify main motives, so-called factors affecting customers’ decisions.
  • Generally describes these “factors” and assign concrete characteristics.
  • Works based on relations (correlation matrix) between monitored characteristics.
  • Enables to assign also other, yet non-monitored characteristics to “factors” and to estimate respondents’ attitude to them.

Gathered analysis

  • Divides population into the groups in accordance with their opinions and attitudes.
  • People from one group are similar to each other, there are significant differences among groups.
  • Identifies customers with similar preferences and designs optimal approach to them.

Discriminant analysis

  • Assigns subjects to predefined groups.
  • Assumes that part of the population is divided into groups, for the second part of (non-assigned to any group) selects the most similar group.
  • Creates new groups based on similarity to already existing ones.

Corresponding analysis

  • Graphically illustrates relations between between two characteristics, example of usage: comparing brands towards several characteristics.
  • Depicts relative distance between brands and characteristics in one chart.
  • Identifies which characteristics are typical for your product and which ones for competition.
  • Analyses which competitor is the most similar to you and on the other hand which one differs from you essentially.

Multidimensional tests

  • Compare groups, e.g. demographically in accordance to their attitudes to list of signs and characteristics.
  • Analyse whether there is a statistically significant difference among groups and in what aspect the opinion of the groups differs.

Multidimensional methods provide new, more comprehensive view of research outcome. They help better understanding to customers’ desires, help to improve communication with them and provide other recommendation for marketing.