Correlation Matrix – A square in which all correlation-coefficients of a study are located. All quantified variables are correlated to one another. The purpose is to discover unanticipated, hidden relations. Many such correlation-coefficient are irrelevant, nonsense, or involve pseudo-relations. For example: frequency with soup is purchased as related to shoe size in a sample of housewife. See Also: Correlation-coefficient/Matrix
Correlation Ratio – A measure of coherence between two variables. It is a kind of correlation-coefficient and is applicable for:
- The combination of one variable on a nominal scale and one variable on an interval scale.
- Non-linear regressions (irregular broken lines in a graph).
Cross-sectional Analysis– Analysis of material that may have been acquired from a cross-sectional study of the population. An analysis is made, for example, of different age groups in relation to income, education, ownership of durable consumer goods. It is also feasible to conduct, for example, a cross-sectional analysis by area. Provided that the distinct groups are of sufficient size, it is possible to use the results of an “ordinary” sample that is nationally representative for this analysis. See also: Cross-sectional research
Density – a ratio, as expressed as a percentage, that indicates how many objects, persons or data occur in a population is 50% (one out of two people owns a typewriter). See also: Viewership/Universe
Descriptive Statistics – A form of statistics that involves the description of the data found. Various kinds of measurements are included, such as central tendencies, dispersion, correlations. The calculation of a mean average, group, etc., also belongs to the field of descriptive statistics, as it describes (with a single number) a phenomenon, group, etc. See also: Statistics/Central tendency/Dispersion/Correlation
Dispersion – Data in dispersed state (numbers, observations, scores). Two series of numbers may have the same average but a different dispersion; therefore, dispersion also yields information about a group of numbers. See also: Measure of Dispersion
Extreme Values – 1. The largest and the smallest number in a series, and therefore, the maximum difference between two numbers in this series.
- The ultimate values in a series which do not necessarily have to be the largest and the smallest numbers, though usually they are.
First Quartile – The value below which 25% of the observation fall in, for example, the experimental test results. See also: Centile/Quartile
Geometric Mean – A central tendency that is used to indicate the relative change in index numbers. For example: A study in 1983 indicates that there are 50 owners of television sets in a village; a study in 1984 registers that the number has increased to 4050 owners. How much has the (theoretical) relative increase per year been? Step 1: assume that the relative increase per year is equal to factor R. Step 2: calculate the constant increase. See also: Central tendency/ Arithmetic average?Ratio Scale
Interaction – A measure of the degree of relation between two or more variables.
Intercorrelation – The mutual correlation between a number of variables, as distinct from the correlation between these variables and an “outside” or dependent variable. See also: Correlation/Variable
Matrix – Syn: Grid A rectangular two dimensional table with columns and rows, which permits special algebraic calculations. It is a comprehensive total of, for example, all or a number of quantitative research variables. See also: Correlation Matrix
Mean Deviation – measure of dispersion derived from the mean deviation of the observations (data) of a central tendency. These deviations are not considered absolutes, that is, the algebraic symbols (+ or -) are not taken into consideration. The central value can be arithmetic average or the median. See also: Dispersion/Measures of dispersion/Central tendency
Measures of Dispersion – Statistical measures that represents the dispersion of data 9numbers, observations, scores) in one single number. The most important of these measures are: variance, standard deviation, range and midrange. These measures are established with he aid of a formula. See also: Standard deviation/Variance/Range?Midrange
Median – The figure that forms the central number in a series of numbers classified (specifically for these calculations) from low to high; for example: 1, 3, 5, 7, 9. Here the median is 5. when it involves classes of data (for example; 13 -19 years, 20-25 years), the median must be calculated. Median formula is used for this purpose.
Midrange – A measure of dispersion. The highest score minus the lowest on scale divided by two. The midrange is, therefore, simple to calculate. A danger is posed by the extremity (low or high) of values; they cause a distorted image. See also: Range/Measures of Dispersion
Mode – A central tendency. The mode is the number that occurs most frequently in a series of numbers. For example: in 2, 3, 5, 5, 5, 7, 7, the mode is 5. When a frequency distribution shows “peaks,” either upwards or downwards, it is frequently more meaningful to determine the mode instead of the mean average, since the latter is, in this instance, no longer an undistorted datum. For example: 10 millionaires in an otherwise very poor village of 10,00 inhabitants would have great influence on the calculation of average income per capita of the population. The average yields a distorted image. Wit the aid of the formula Mo =3Mdn – 2M, the mode can be calculated by means of the median and the average. See also: Unimodal/Bimodal/Central Tendency
Moving Average – The average of statistical data calculated over a progressively changing interval, such as, in time series. See also: Time series