# Trend

Point Cloud – The correlation of points in a scatter diagram. See also: Scatter diagram

Profile – Graphic representation of research results in which the objective is to show clearly the differences between elements or parts.

Profile Score – The individual score in a profile. See also: Profile

Quadrant – Square. Statistical data are often displayed by means of a matrix. Sometimes the matrix is divided into four so that the data can be “read” easily. For example, in the price/quality matrix below all data in the quadrant K1 refer to high price and law quantity. In K2 they refer to high price and high quality

Scatter Diagram – A correlation diagram that shows the relation between 2 variables (typically on the X & Y axis) of a graph. Whenever there are a talks of a point cloud. Each dot shows an individual relationship. See also: Correlation

Trend – A developmental direction. In a graph, the number of products sold is indicated by units of time (for example: months). After a number of measurements have been carried out, a picture is obtained of the direction in which sales are developing. Data yielded by panels are particularly suited for assessing trends, since trends involve alterations due to the influence of time. The group from which the data originate remains unchanged. See also: Panel/Trend Analysis

Trend Analysis – On the basis of a series of consecutive measurements, the progress of a graphic line is assessed. After the results have been studied, further movement is predicted.

Venn Diagram – Named after the British mathematician Venn: a graphic method that is used mainly to represent probabilities. Circles or ellipses indicate probabilities. The usual method is: colors or crosses = the probability for A; stripes = either the probability for A or no probability for A (but for B); blank = indefinite (can be either).

Above Average – A group of people or results that are above a certain norm in a research study. Usually those who are on the “right” of the standard norm; i.e. above the average. For example: if the average income in a particular survey is \$50,000, then those respondents with an income of \$60,000 are above average. See also: Norm/Distribution

Arithmetic Average – The sum of all quantities in a set, divided by the number of quantities. Example: the body weight of 4 persons in an automobile is, respectively: 60kg., 70kg., 75kg. And 91 kg. See also: Central tendency/Geometric mean

Attenuation – Syn: Stress

Attenuare (Latin) = diminution, decrease.

1. Because the influence of errors in measurements, the true value of a coefficient of correlation becomes obscured: the correlation-coefficient is usually an under-estimation. There are formula and tables available that indicate the real magnitude of the calculated correlation-coefficient. However, the calculated or indicated “rue” correlation-coefficient is, as a rule, too high.
2. Attenuation or stress arises when a computer handles a maximum of three dimensions (as in a cube), while the subject/respondents (wittingly or otherwise) utilize more than three dimensions. This distortion is called stress or attenuation. See also: Correlation-coefficient

Average – Syn: Mean   An often used, “but not entirely correct, short-form term for “arithmetic average.” The most familiar and most frequently used “central tendency.” See also: Arithmetic average?Central tendency

Base – A number (or magnitude) that can be employed as a standard of reference. It may occur as the denominator in a fraction or in a percentage.

Canonical Correlation – A correlation analysis of the relation between two sets of variables. Standard correlation techniques concern the relation between a dependent variable and one, or more, independent variables. In a canonical analysis, the dependent variable is set of variables (for example: purchasing frequency of a product, price, frequency of use) and not one single variable. See also: Correlation/Correlation-coefficient/Variable

Cell – A small homogeneous group within a larger sample. The cell is the smallest unit in a cross table. See also: Cross Table

Centile – Syn: Percentile

Centrum (Latin) = one hundred. One hundred part of a distribution. Statistical measure, based on a scale composed of one hundred equal parts. An individual in the 28th centile is better/more etc. than 27% of all individuals studied. The individual is worse/less etc. that 72%. The 50th centile is the median. The 25th and the 75th centiles are termed the first and third quartile. See also: Median/First Quartile/Third Quartile

Central Tendency – Syn: Location measure       A single number that indicates the position of the center of a number of data. The most important central tendencies are: arithmetic average/Mode/Median

Correlation – The statistical coherence between two, or more, variables (groups of numbers). The correlation does not have to indicate a causal dependence. The criterion for the coherence is the correlation-coefficient. See also: Correlation-coefficient/Variable

Correlation-coefficient – A statistical concept that indicates the magnitude of the relation between two variables. It is a standard for the calculation of the relation between phenomena. Symbol: r. It varies (by definition) between -1 and +1, where 0 indicates no relation, + full relation, -full relation, but opposite. With the passing time, quite a number of different formulas for the correlation-coefficient have been developed. The most familiar formula, used most frequently, is the product-moment correlation-coefficient. See also: Product moment correlation-coefficient/Correlation