**Attribute** – A qualitative characteristic of an individual, usually applied to distinguish a variable (or other quantitative characteristic). For example,sex is an attribute, but age a variable. Attributes are often dichotomous but not in every instance (for example, blood groups). See also: Variable/Dichotomy

**Atypical** – A case, person, object or datum that clearly deviates from norm. A person wearing clothes in a nudist camp would be atypical, as would a communist in a company of liberals. The same holds true for a person who stutters or a beggar in a Western industrialized society; however, they would not be considered atypical in a club for stutterers or beggars’ guild.

**Cluster** – A group of coherent elements in a statistical population. For example, a group of people living in the same house or building, a number of sequential observations in series or number of adjacent properties.

**Company Statistics** – Various graphs and tables that help from a picture of how a company functions: for example, the output in number of units per week, absenteeism due to illness, cost price of raw materials on a monthly basis, etc. Company statistics are often used for desk research. See also: Desk Research

**Constant** – 1. A mathematical value, used in statistical formulas and tests/experiments that remains the same at all times.

- An experimental factor that remains the same within the framework of research. See also: Statistical test/Experiment

**Contrasting Group **– A statistical group that differs strongly or to a maximum degree from another group (after analysis). For example, Coca Cola drinkers versus on- Coca Cola drinkers, home owners versus those who rent. See also: Contrast Analysis

**Decimal Notation** – A statistical system using 10 as a base. The metric system is based on this notation.

**Index Numbers** – A measure for the magnitude of an alteration of a variable. This measure corresponds to a specific basic value of the variable. The basic value is usually 100 (arbitrarily selected). Values in excess of 100 indicate an increase in relation to the basic value, while values below 100 indicate a decrease. See also: Consumer Price Index

**Indicator** – An instrument that provides information concerning a situation capable of being measured. The indicator may be a dial or meter of an electronic instrument, or the answer to a question or the totality of the answers. See also: Social Indicator

**Law of Large Numbers** – The principle that individual differences or deviations cancel out one another when the number of occurrences is sufficiently large. For example, in a case in which extremely careless people are part of sampling, their effect will be canceled out by the participants who are extremely precise.

**Margin of Error** – Syn: Margin of Tolerance The magnitude of error(s) permitted in reference to a standard or criterion. See also: Confidence Margin

**Margin of Toleran**ce – Syn: Margin of error See: Margin of error

**Markov Chain Analysis** – A theoretical process that makes it possible to make a prediction pertaining to the future. The prediction, which is based on knowledge of the current situation, is as precise as would be the case if it were based on the entire past. For this analysis, a simple matrix (for example: 3×3) is established. The only information required is that pertaining to the most recent fact of behavior of a person (for example:voting behavior at the most recent elections, most recent purchase, most recent move). The assumption is always:the history of a person is of no(direct) influence on choice or behavior. At times this may well be a tendentious assumption.

**Monte Carlo Method** – Named after the famous gambling casino in Monaco, an empirical study of statistics using a random numbers. Throwing dice is a simple example; each number has an equal chance of falling. Using this method, it is possible to determine whether after 60 throws, number one has come up 10 times, number two has come up up 10 times, etc. as anticipated.

**Objective Criterion** – A norm, measure of standard against which data can be tested. This standard is established outside the context in which the data to be tested originate. For example, an official tax assessment is an objective criterion against which one’s view of tax liability can be measured. See also: Criterion

**Permutation** -1. Transformation of research results according to specific statistical formulas.

- The sequence of a group of objects or questions on a questionnaire or changes in that sequence.

Quantifying – To express data, assumptions, indications or hypotheses in terms of numbers. When a hypotheses requires affirmation, it needs to be put to the test. This can only be achieved with the aid of statistical material (figures) obtained from a sampling of sufficient size.

**Random Numbers** – Random numbers do not have any significance. Different lists of random numbers are constructed to be used in taking samples. Each number from 0 to 9 appears with equal frequency in such a list; its composition is based on chance; therefore, there is no mutual connection between the numbers. Computer programs exist for the creation of this kind of list.