|
Cruise Scientific Visual Statistics Studio Table of Contents |
Visual statistics encompasses traditional statistics and modern computer assisted data analysis. It is modern continuation of the classical epistemology, providing foundations for a general theory of scientific inquiry. It begins with description of quantities, their relationships, and structure.
A data matrix is a rectangular arrangement of numbers symbolizing properties of phenomena under scrutiny. These numbers are located at intersections of data matrix's rows and columns.
Columns of data matrices are called variables, or, in general, attributes. Rows of a data matrices are typically subjects, but can be also other entities.
Variability and Information
The numbers in data matrices on the one hand must not be the same and on the other hand must not be random to carry information about phenomena that they describe. This variability of numbers in data matrices can be expressed as a quantity, called the variance. The goal of data analysis is to analyze the variance contained by the data and thus make the data meaningful.
A matrix of data from the real studies is usually much larger than the matrix shown above. This matrix must be entered it into a computer and analyzed. Events that take place after the data matrix is submitted to computer for data analysis are the subject of this book.
Within social sciences, an element of a data matrix is often created when a subject encounters a statement: a test item, a question on a survey, or an inquiry about a personality characteristic. These queries are channeled through subject's cognitive systems that, in turn, direct psychomotor reactions of recording responses to statements.
The row marginal referents of data matrices carry information about identities of subjects generating the responses. The column marginal referents contain information about the meaning of the statements. Thus, an element of a data matrix typically carries a quantified reaction of a subject to a statement.
Data matrices are bordered by row and column marginal referents. From the standpoint of formal data analysis, the exact character of these marginal referents is immaterial. Some authors prefer to call the marginal referents of data matrices attributes and entities to stress this point. By associating attributes and entities with numbers, we can generate a variety of designs. The possibilities are unlimited; statistical data analysis is not restricted by discipline, or by the character of measured entities or attributes.
There is a dictum that 'anything which exists, exists in some amount, and therefore can be measured.' Statistical data analysis attempts to translate measurements into indices that can be interpreted and structures that can be described. Following data analysis, a story can be told and a paper can be written. The findings are closer to reality and more believable than if based solely on author’s personal experiences, introspections, or beliefs.
