Media-Screen: Glossary of Terms

A-D | E-H | I-M | N-Q | R-U | V-Z

Ad hoc survey - A custom survey fielded to the target audience without prior contact by the researcher.

Base - The number of people answering each question on a questionnaire.

Bias - Extent to which, over repeated samples, the mean of the sampling distribution differs from the true mean. Bias is generally hard to quantify, but is likely to increase if the sampling frame is deficient and/or the response rate is low.

Booster - Where one is interested in the views of young people, for example, an extra sample may be selected to investigate the responses specific to that group.
Extra sample points may be selected to investigate the responses of specific groups within the universe dataset. These points are picked so that they are biased towards a specified group within the population and help to highlight the responses of the specific group under investigation.

Bulletin Board Focus Group - offers a unique opportunity to engage people online in in-depth discussions over an extended time frame. Typically involves 8-12 often busy professionals who may normally lack the time to attend a real-time group. Groups can bring together respondents from all over the world and typically last about three days.

CATI - Computer Assisted Telephone Interviewing.

Coding - Turning respondents' written answers into numeric answers.

Confidence Interval - A range of values constructed around a point estimate that makes it possible to state that an interval contains the population parameter between its upper and lower confidence limits. The most frequently used confidence interval is the 95% confidence interval. This can be interpreted as there is only a 5% chance that the sample is so extreme that the 95% confidence interval calculated will not cover the population mean.

Correlation - A measure of the strength of linear association between two variables.

Demographics - Information about the sample that includes areas such as age, sex, social class, presence of children, etc.

Depth Interviews - Involve one-to-one contact with respondents and are usually conducted face-to-face, although telephone interviews are sometimes used. The depth interview is very different from the tightly structured interview used in quantitative surveys or opinion polls: it is not only longer (45-60 minutes), but also more discursive and open-ended. The interviewer also has a greater level of flexibility: since they are not constrained by the order or wording of questions, s/he is able to cover the issues specified in the topic guide in a more context sensitive way, as they 'naturally' and conversationally emerge, and to probe responses to gain a full understanding of their meaning and/or implications.

Diary Study - Asks people to keep a diary or journal regarding targeted aspects of their life. Can be conducted online or offline, through written entries or through tape or video recorders. Diary studies provide longer-term feedback and can be useful for tracking trends.

Back to Top

Ethnography - Involves the study of a small group of subjects in their own environment. Ethnographic studies involve both descriptive and interpretive elements and offer deep investigations into subjects over a longer period of time.

Factor/Cluster Analysis - A multi-variate, data reduction technique. It aims to summarize a large number of variables with a small number of factors.

Filter - When only a section of the total sample are required to answer the question, e.g. if the question asks why people are dissatisfied with a particular service, only those who are dissatisfied should answer the question. Those who are satisfied will skip to the next question that is to be asked of all respondents.

Focus Group - Are open ended, discursive, and are used to gain a deeper understanding of respondents' attitudes and opinions. Typically involve between 6-10 people, and last for 1-2 hours. A key feature of group is that participants are able interact with, and react to, each other. In order to facilitate this group dynamic it is therefore important to ensure that participants do not know each other beforehand and that they are broadly 'compatible'.

Back to Top

Incidence - Any figure referring to the percentage of people in a category. Examples: incidence of users, incidence of people qualifying for a study.

Margin of Error - A measurement of the accuracy of the results of a survey. Example: A margin of error of plus or minus 3.5% means that there is a 95% chance (confidence interval) that the responses of the target population as a whole would fall somewhere between 3.5% more or 3.5% less than the responses of the sample (a 7% spread). However, for any specific question, the margin of error could be greater or less than plus or minus 3.5%.

Mean - The arithmetic average of a set of data in which the values of all observations are added together and divided by the number of observations

Median - The outcome that divides an ordered distribution exactly into halves.

Mode - The single category among the categories in the distribution with the largest number of observations.

Multiselect questions - Where respondents are allowed to give a number of responses to one question.

Multi-variate analysis - A wide range of techniques which are used to analyse many variables and the relationship between them.

Back to Top

Omnibus - Surveys conducted regularly by a research agency, but which can be sponsored by a number of clients who do not share the same data but who buy access to a survey and questions to meet their individual needs. A multi client survey run regularly among samples that are nationally representative of the general public.

Online Real-Time Focus Group - Brings together people from across the country or around the world, without the expense of travel, for a real-time discussion online. Can involve the presentation of multi-media objects if necessary.

Open-ended data - Data from questions where a respondents' answer has been written in.

Panel - A variation of the quantitative survey approach. Where it differs however, is that participants become part of an ongoing process of regular consultation: as a member of a panel they may take part in several surveys or events over the course of a year.

Pilot - Where research is conducted in advance of the actual study to assess the logistics of the study (i.e. the sample, methodology, question wording etc.). The aim is to highlight any areas of weakness which can then be eradicated in the actual study.

Population - The entire set of persons that have at least one common characteristic of interest to the researcher. The sample is selected from the population.

Pre-coded data - Data that is generated from questions with fixed answers written on the questionnaire.

Qualitative/Qual - Gives an in-depth understanding of why people hold particular views, and how they make judgements, in a way that structured quantitative research cannot, but is not intended to be statistically reliable. Newer types of qualitative research attempt not just to understand but also to inform participants views by actively supplying them with information which allows them to reach a more considered view.

Used to identify not only what people think but also, more importantly, the reasons why they hold such views. While these techniques do not produce results which are statistically reliable (quantitative methods are needed for this), findings can - if participants are broadly representative - be strongly indicative of the population as a whole and have a valuable diagnostic value

Quantitative/Quant - Designed to produce data that tell us how many people do or think something, and to be statistically reliable

Quotas - Specify the number of respondents that fall into the categories required (e.g. so many women, so many men etc.). Interviewers are given quota sheets to show them how many of each type of person are to be interviewed. The sample quota profile is a profile of all the individuals that will be interviewed.

Back to Top

Reliability - The extent to which different studies of the same concept produce consistent results. To provide results that are statistically reliable a reasonably large and representative sample is required.

Representative sample - A representative sample is one which has been selected in such a way that, as far as can be ascertained, the main characteristics of the sample match those of the parent population; i.e. the population from which the sample has been drawn.
Respondent - Those taking part in the research.

Response rate - The proportion of those people originally drawn at random from the population who actually end up taking part in a survey. This will help indicate whether the data being collected accurately reflects the views of the population being interviewed. If the response rate is low it is important to examine the profile of the sample and assess whether weighting or other methods would help to improve the quality of the data.

Sample - A subset of cases or elements selected from a population.

Sample profile - Using Census and other data gives a profile of people living in a certain area i.e. the percentage of 16-24 year olds, the percentage of ABs etc. Ensures those interviewed are representative of the local population.

Significance testing - Using statistical tests to determine whether the observed difference between sample proportions could occur by chance in the populations from which the samples were selected or not.

Single select questions - Where respondents are force to give a single response to a question.

Standard deviation - The variance of values from the mean.

Subgroups - Sub-divisions of the sample (e.g. men, women, those aged 15-25, etc).

Survey research - The method of collecting information by asking a set of pre-written questions in a predetermined sequence in a structured questionnaire to a sample of individuals drawn so as to be representative of a defined population.

Tracking surveys - Surveys repeated using the same questions and methodology in order to monitor changes in attitudes or behaviour.

Trend data - Repeat surveys showing how views on a particular topic have changed over time.

Topic guide - A check list/aide memoir which simply lists the main topics for discussion. It is not a structured list of questions to be followed in the order they are presented as is the case with the questionnaire. It helps the researcher keep track of the central research objectives while at the same time giving flexibility for improvisation.

Topline Results - Usually unchecked figure results in the form of a marked-up questionnaire, which are based on the total sample only.

Back to Top

Validity - The degree to which a variable accurately reflects the concept it is intended to measure.

Variable - Any characteristic or attribute of persons, objects, or events that can take on different numerical values.

Weighting - Weighting is the process by which data are adjusted to reflect the known population profile. This is to counter any effects of differential refusal rates, interviewers falling short on particular quotas, or to correct for any over-sampling of minority populations. A 'weight' is the percent assigned to a particular criterion. We need to weight the sample is the responses show that particular groups (for example younger people or those living in a particular area) are under represented in the sample. If this is not carried out then the results will not properly reflect the views of the population being considered. Where data has not been weighted, this is referred to as 'unweighted' data.

Back to Top

lens | philosophy | netpop | services | methods | news | about us | contact us




			A-D \| E-H \| I-M \| N-Q \| R-U \| V-Z Ad hoc survey - A custom survey fielded to the target audience without prior contact by the researcher. Base - The number of people answering each question on a questionnaire. Bias - Extent to which, over repeated samples, the mean of the sampling distribution differs from the true mean. Bias is generally hard to quantify, but is likely to increase if the sampling frame is deficient and/or the response rate is low. Booster - Where one is interested in the views of young people, for example, an extra sample may be selected to investigate the responses specific to that group. Extra sample points may be selected to investigate the responses of specific groups within the universe dataset. These points are picked so that they are biased towards a specified group within the population and help to highlight the responses of the specific group under investigation. Bulletin Board Focus Group - offers a unique opportunity to engage people online in in-depth discussions over an extended time frame. Typically involves 8-12 often busy professionals who may normally lack the time to attend a real-time group. Groups can bring together respondents from all over the world and typically last about three days. CATI - Computer Assisted Telephone Interviewing. Coding - Turning respondents' written answers into numeric answers. Confidence Interval - A range of values constructed around a point estimate that makes it possible to state that an interval contains the population parameter between its upper and lower confidence limits. The most frequently used confidence interval is the 95% confidence interval. This can be interpreted as there is only a 5% chance that the sample is so extreme that the 95% confidence interval calculated will not cover the population mean. Correlation - A measure of the strength of linear association between two variables. Demographics - Information about the sample that includes areas such as age, sex, social class, presence of children, etc. Depth Interviews - Involve one-to-one contact with respondents and are usually conducted face-to-face, although telephone interviews are sometimes used. The depth interview is very different from the tightly structured interview used in quantitative surveys or opinion polls: it is not only longer (45-60 minutes), but also more discursive and open-ended. The interviewer also has a greater level of flexibility: since they are not constrained by the order or wording of questions, s/he is able to cover the issues specified in the topic guide in a more context sensitive way, as they 'naturally' and conversationally emerge, and to probe responses to gain a full understanding of their meaning and/or implications. Diary Study - Asks people to keep a diary or journal regarding targeted aspects of their life. Can be conducted online or offline, through written entries or through tape or video recorders. Diary studies provide longer-term feedback and can be useful for tracking trends. Back to Top Ethnography - Involves the study of a small group of subjects in their own environment. Ethnographic studies involve both descriptive and interpretive elements and offer deep investigations into subjects over a longer period of time. Factor/Cluster Analysis - A multi-variate, data reduction technique. It aims to summarize a large number of variables with a small number of factors. Filter - When only a section of the total sample are required to answer the question, e.g. if the question asks why people are dissatisfied with a particular service, only those who are dissatisfied should answer the question. Those who are satisfied will skip to the next question that is to be asked of all respondents. Focus Group - Are open ended, discursive, and are used to gain a deeper understanding of respondents' attitudes and opinions. Typically involve between 6-10 people, and last for 1-2 hours. A key feature of group is that participants are able interact with, and react to, each other. In order to facilitate this group dynamic it is therefore important to ensure that participants do not know each other beforehand and that they are broadly 'compatible'. Back to Top Incidence - Any figure referring to the percentage of people in a category. Examples: incidence of users, incidence of people qualifying for a study. Margin of Error - A measurement of the accuracy of the results of a survey. Example: A margin of error of plus or minus 3.5% means that there is a 95% chance (confidence interval) that the responses of the target population as a whole would fall somewhere between 3.5% more or 3.5% less than the responses of the sample (a 7% spread). However, for any specific question, the margin of error could be greater or less than plus or minus 3.5%. Mean - The arithmetic average of a set of data in which the values of all observations are added together and divided by the number of observations Median - The outcome that divides an ordered distribution exactly into halves. Mode - The single category among the categories in the distribution with the largest number of observations. Multiselect questions - Where respondents are allowed to give a number of responses to one question. Multi-variate analysis - A wide range of techniques which are used to analyse many variables and the relationship between them. Back to Top Omnibus - Surveys conducted regularly by a research agency, but which can be sponsored by a number of clients who do not share the same data but who buy access to a survey and questions to meet their individual needs. A multi client survey run regularly among samples that are nationally representative of the general public. Online Real-Time Focus Group - Brings together people from across the country or around the world, without the expense of travel, for a real-time discussion online. Can involve the presentation of multi-media objects if necessary. Open-ended data - Data from questions where a respondents' answer has been written in. Panel - A variation of the quantitative survey approach. Where it differs however, is that participants become part of an ongoing process of regular consultation: as a member of a panel they may take part in several surveys or events over the course of a year. Pilot - Where research is conducted in advance of the actual study to assess the logistics of the study (i.e. the sample, methodology, question wording etc.). The aim is to highlight any areas of weakness which can then be eradicated in the actual study. Population - The entire set of persons that have at least one common characteristic of interest to the researcher. The sample is selected from the population. Pre-coded data - Data that is generated from questions with fixed answers written on the questionnaire. Qualitative/Qual - Gives an in-depth understanding of why people hold particular views, and how they make judgements, in a way that structured quantitative research cannot, but is not intended to be statistically reliable. Newer types of qualitative research attempt not just to understand but also to inform participants views by actively supplying them with information which allows them to reach a more considered view. Used to identify not only what people think but also, more importantly, the reasons why they hold such views. While these techniques do not produce results which are statistically reliable (quantitative methods are needed for this), findings can - if participants are broadly representative - be strongly indicative of the population as a whole and have a valuable diagnostic value Quantitative/Quant - Designed to produce data that tell us how many people do or think something, and to be statistically reliable Quotas - Specify the number of respondents that fall into the categories required (e.g. so many women, so many men etc.). Interviewers are given quota sheets to show them how many of each type of person are to be interviewed. The sample quota profile is a profile of all the individuals that will be interviewed. Back to Top Reliability - The extent to which different studies of the same concept produce consistent results. To provide results that are statistically reliable a reasonably large and representative sample is required. Representative sample - A representative sample is one which has been selected in such a way that, as far as can be ascertained, the main characteristics of the sample match those of the parent population; i.e. the population from which the sample has been drawn. Respondent - Those taking part in the research. Response rate - The proportion of those people originally drawn at random from the population who actually end up taking part in a survey. This will help indicate whether the data being collected accurately reflects the views of the population being interviewed. If the response rate is low it is important to examine the profile of the sample and assess whether weighting or other methods would help to improve the quality of the data. Sample - A subset of cases or elements selected from a population. Sample profile - Using Census and other data gives a profile of people living in a certain area i.e. the percentage of 16-24 year olds, the percentage of ABs etc. Ensures those interviewed are representative of the local population. Significance testing - Using statistical tests to determine whether the observed difference between sample proportions could occur by chance in the populations from which the samples were selected or not. Single select questions - Where respondents are force to give a single response to a question. Standard deviation - The variance of values from the mean. Subgroups - Sub-divisions of the sample (e.g. men, women, those aged 15-25, etc). Survey research - The method of collecting information by asking a set of pre-written questions in a predetermined sequence in a structured questionnaire to a sample of individuals drawn so as to be representative of a defined population. Tracking surveys - Surveys repeated using the same questions and methodology in order to monitor changes in attitudes or behaviour. Trend data - Repeat surveys showing how views on a particular topic have changed over time. Topic guide - A check list/aide memoir which simply lists the main topics for discussion. It is not a structured list of questions to be followed in the order they are presented as is the case with the questionnaire. It helps the researcher keep track of the central research objectives while at the same time giving flexibility for improvisation. Topline Results - Usually unchecked figure results in the form of a marked-up questionnaire, which are based on the total sample only. Back to Top Validity - The degree to which a variable accurately reflects the concept it is intended to measure. Variable - Any characteristic or attribute of persons, objects, or events that can take on different numerical values. Weighting - Weighting is the process by which data are adjusted to reflect the known population profile. This is to counter any effects of differential refusal rates, interviewers falling short on particular quotas, or to correct for any over-sampling of minority populations. A 'weight' is the percent assigned to a particular criterion. We need to weight the sample is the responses show that particular groups (for example younger people or those living in a particular area) are under represented in the sample. If this is not carried out then the results will not properly reflect the views of the population being considered. Where data has not been weighted, this is referred to as 'unweighted' data. Back to Top
			lens \| philosophy \| netpop \| services \| methods \| news \| about us \| contact us