- A - Definitions
Accuracy (in sampling) The closeness of a sample estimate to the true population value.
Approximate Almost correct, but not completely so.
Arithmetic mean See mean.
Assumption Thing accepted as true or as sure to happen, but not proved.
- B -  
Bias The systematic difference in collected statistical data from the actual characteristic(s) of the target population. Sources of bias are the sample frame, sample selection, respondents and data processing.
Briefing The instructions, information and advice given to staff. See also debriefing.
- C -  
Case study An enquiry in which a small number of study units are investigated in great detail. Usually – but not necessarily – the study units are selected purposively.
Census An investigation which covers individually every unit in the population being studied.
Classification The set of codes used to assign a response to a particular category.
Closed question A question which has a number of pre- specified response categories.
Coding The values assigned to collected data so that it is arranged into classes according to specified characteristics.
Computer hardware The physical components of a computer system.
Computer software A general term used to describe the set of instructions given to a computer to perform operations or applications used to perform operations.
Consumer A person who uses goods and services. A person who purchases goods and services.
Consumer Price Index (CPI) A measure of the average change of prices of goods and services bought by households.
Cost of living index An index measuring the change in the cost preserving a particular standard of living. Such an index would take account of changes in the mode of living as well as price changes. The term is frequently, but incorrectly, used to describe the CPI.
Coverage See survey population.
CPI coverage The definition of CPI in terms of goods and services represented, and of the population group and geographic areas to which it relates.
- D -  
Data entry A part of data processing where the responses from questionnaires (or the coded responses) are entered into a computer.
Data items The items of information collected.
Debriefing A discussion reviewing the tasks and activities undertaken. The debrief is usually conducted by senior staff and reviews the instructions, information and advice given to staff as well as their experiences in conducting the tasks and duties assigned to them. See also briefing.
Descriptive statistics The compilation and presentation of data in tables, graphs and diagrams.
- E -  
Edits The checks that are made to ensure the correctness of the data.
Enumerate To count.
Enumerator A person employed to count (and usually to collect data).
Estimate An approximate calculation.
- F -  
Field edits The checks that are made to ensure the correctness of the data collected from respondents.
Field staff Enumerators, field supervisors, managers and any other people recruited to collect data from respondents.
Fieldwork The process of collecting data from respondents.
Filter A mechanism (or operation in computers) that allows only certain things to pass.
Formulate Create something in a precise form, express clearly and exactly.
Frame (listing or register) A list of statistical units (people, groups of people, businesses).
- G -  
Gantt Chart A matrix of activities and time used to plan, manage and monitor a project.
Goods Physical items, as distinct from services.
- H -  
Hardware See computer hardware.
Histogram A graphical presentation of data by the areas of rectangular blocks, arranged in vertical or horizontal columns.
Household income and expenditure survey (HIES) A survey conducted to determine the income and expenditure patterns of private households. An important source of data for the weighting pattern of a CPI.
Hypothesis An Idea or suggestion that is based on known facts and is used as a basis for reasoning or further inquisition. Hypothetical, based on a hypothesis, not necessarily true or real.
- I -  
Index In mathematics a number scale by means of which levels can be compared with some base value, usually over time.
Inference Drawing conclusions about a population on the basis of evidence obtained from a sample.
- J -  
Jargon Technical or specialised words used by a particular group of people but difficult for others to understand eg. inter-penetrating subsamples. Should be avoided as much as possible.
- K -  
Kurtosis Kurtosis (the term first used by Pearson, 1905) measures the "peakedness" of a distribution. If the kurtosis is clearly different than 0, then the distribution is either flatter or more peaked than normal; the kurtosis of the normal distribution is 0.
- L -  
Listing See frame.
Literature review A search of published and unpublished reports, papers and articles for information on the topic of interest.
Logic edits Verify the correctness of a particular response using the response(s) to other queries. Eg. children cannot be older than their parents. See also edits.
- M -  
Matched sample The use of the same sample of items (or respondents) in successive periods.
Mathematical statistics The tools of statistics, the methods used to analyse statistical data and to make estimates or draw conclusions from statistical data.
Mean (arithmetic mean) The average value of a number of observations, calculated by dividing their sum by the total number of observations.
Median The value of the middle item in an ordered set of data. Fifty percent of observations lie above and fifty percent below the median.
Mode The data value which occurs with the highest frequency.
Monitor To keep check of something, to make continuous observations of something.
- N -  
Non-sampling error That part of the error in a statistical collection other than the error introduced by only collecting sample data, i.e. ambiguous phrasing of questions and enumerator error in recording respondents’ answers.
Numerical Expressed in or representing numbers.
- O -  
Open question A question which does not have a range of pre-specified responses, i.e. respondents can answer as they wish.
- P -  
Percentage change The change in an index series from one period to another expressed as a percentage of its value.
Pilot survey (or pre-test, or dress rehearsal) A practice or trial run of the methods and procedures to be used for the whole survey.
Population The entirety of a defined set of units, eg. the total human population of a village or a state, the total number of plants in a field, the total number of children enrolled at primary schools in a province, etc.
Probability sampling Sampling in which every individual has a known probability of selection.
Proportion The number of units possessing some characteristic, divided by the total number of units with or without that characteristic.
Proxy respondent A respondent who answers questions on behalf of the intended respondent.
Purposive sampling A type of sampling in which the selected units are chosen not by chance but deliberately – in order to include units with particular characteristics, or for reasons of convenience.
- Q -  
Quality control The process of keeping the standard of outputs within acceptable limits of error.
- R -  
Random Without conscious choice.
random numbers Numbers selected at random, that is, selected without conscious choice.
random sampling Sampling in which every individual is selected at random.
Range The difference between the highest and lowest values in a data set.
Ratio One number or variable divided by another.
Recall period The length of time into the past for which a respondent is asked to recall information.
Reference period The period for which a respondent is asked to give information about events occurring within it.
Register See frame.
Reliable Good quality, dependable.
Representative Typical of a group, or kind.
Respondents Individuals, businesses, authorities, etc. from which data are collected.
Respondent burden The trouble or inconvenience to a respondent of being included in a survey.
- S -  
Sample (of goods or outlets) A representative selection of goods and services to be priced, or of retail outlets from which to collect prices.
Sample estimates The estimated measures of population values (means, totals etc) achieved by performing computations on the sample data.
Sample population See survey population.
Sample survey A data collection which collects data from a subset of units in the population being studied.
Sampling error That part of the difference between a sample estimate and the true population value (derived from a census) which arises because of estimating from a sample rather than from the whole population. It represents the uncertainty due to sampling.
Sampling frame A list of ordered arrangements of all the units in the population, which is used for selecting a sample.
Sampling unit The type of unit or element selected in a sample, and which is then the unit to be measured, observed or interviewed.
Sampling without replacement A type of sampling in which, when a unit has been selected for the sample, it is removed from the pool of units for selection and cannot be chosen again.
Scope See target population.
Self-weighting sample A type of sample in which all units have an equal chance of being selected, so the sample values do not need to be weighted to compensate for different probabilities of selection.
Standard error A measure of the precision of an estimate.
Statistical data The values observed in collecting information.
Statistical method The collection, analysis and interpretation of statistical data.
Statistical table A matrix (rows and columns) used to arrange and present statistical data.
Statistical unit A person, group of persons or things about which we collect information or make observations. The individual items in a population about which we want to collect information.
Statistics Numbers that are systematically collected and analyzed. Also the science or skill of dealing with statistical data.
Standardise To make something conform to a fixed standard.
Stratification The division of the population into sub-populations, which are then sampled independently.
Survey population (or coverage) The set of statistical units information is collected from.
Survey proposal A document which contains the objectives of the statistical survey, its main concepts and the methods to be used.
- T -  
Table See statistical table.
Tabulate To arrange figures in the form of a statistical table.
Target population (or scope) the group of statistical units we ideally want to collect information from.
- U -  
Unit of study An entity or event defined to be the subject for the collection of data. Commonly a household or a person, or an event such as a market transaction.
Urban households Households in the townships.
- V -  
Verify Check that something is correct.
- W -  
Weight The measure of the relative importance of an item in the index regimen.