Research jargon explained



Analysis - Systematic sorting and arranging of data to interpret the data and address the objectives of the research.

Attitude - A deep-seated belief, based on the accumulation of education, background, experiences.  (See also 'opinion'.)

Audit - An independent and systematic check that the required procedures are followed, the required features are present and functioning.

Benchmark - An initial measure against which subsequent results are compared.

Bias - Inaccuracy because the sample does not have the same profile as the population it is meant to represent, or because questions are worded to encourage a particular answer. 

Closed questions - Questions to which there is a limited and pre-determined range of answers, eg 'yes' or 'no'.

Concept - An idea; a suggestion.  See also 'proposition'.

Conclusions - Judging what the results mean and how they answer the research objectives.  Not to be confused with 'summary' (see below)

Confidence limits - Because survey research is based on samples, rather than whole populations, all results are estimates.  Confidence limits indicate the likely accuracy of a particular result.   

Correlation - The degree to which two sets of data depend on each other. 

Cross-tab - A table showing the links between answers to two different questions. 

Data - A collection of facts or statistics.  Data are analysed and interpreted to provide information (see 'information').

Demographic profile - Description of a population or sample based on personal characteristics such as age, gender, working status, social grade, etc.

Desk research - Research based on data that have alreadt been collected and analysed for another purpose. Aka 'secondary' or 'desktop' research.    

Distribution - (1) The channels through which consumers obtain products or services.  (2) The spread of responses to a question.

Frequency - The number of times the same answer is given to a question.

Hall test - Interviews conducted in a public hall or hired room, usually so that test items can be viewed or tried in a controlled environment, or because limited numbers of test items are available, or because test items are fragile or difficult to move.

Hole count - Simple count of all answers given in a survey. Originates from when data were inputted to computers via punch cards.  Means literally a count of holes punched.

Image - The impressions and perceptions of a brand, product, service or company that the market has, as opposed to the impressions and perceptions that the company or supplier intends.

Information - News or intelligence.  Obtained by analysing and interpreting data.  See also 'knowledge'.

Insight - Mental penetration, understanding.  Another word for knowledge, interpretation or conclusions. 

Interpretation - Inferring meaning from a set of data.

Knowledge - Understanding, learning.  The result of combining information with skill and experience.

Leading question - A question that pushes respondents to giving a particular answer. 

Lifestage - Profile characteristics based on the most recent key milestone, eg marriage, parenthood, retirement, etc.

Loyalty - Repeatedly purchasing a particular brand in preference to competitor brands.

Map - A picture summarising people's perceptions of the differences and similarities between brands, products, services or companies.

Mean score - One way of calculating the average response to a question where answers are recorded on a scale, eg income, age, service ratings.  Other ways of calculating the average result are the Mode and the Median.

Media/Medium - 'Media' is the plural of 'medium'.  A medium is the means by which something is carried or communicated.  'Media' has come to mean the news and broadcasting industry.

Mission - An activity or set of activities to achieve an ultimate goal or vision.  Not a research term, but often crops up in corporate development research.  

Mystery shopping - Researchers pose as buyers or users of a product/service to check the sales/service experience. 

Objective - (1) Something that you intend to achieve.  (2) The opposite of subjective.

Open-ended questions - Questions where answers cannot be predicted and are therefore written down in the respondents' own words.  Typically 'how' or 'why' questions.  Not intended to be used for questions where the range of answers could mostly be predicted. 

Opinion - A point of view.  Can be easily influenced and changed.

Panel - A sample of people who are repeatedly re-interviewed, either on the same or on different topics.

Positioning - The image that a brand, product, service or company has relative to its competitors.

Pre-coded questions - Questions where potential answers are listed for ease and speed of recording responses.  In interviews, it does not necessarily mean that the list of possible answers is revealed to respondents. (See 'prompt'.)

Pre-test - Testing people's reactions to something, eg a new product or advertisement, before it is launched.

Primary research - Any research that collects data directly from the subject, eg through interviews or experiments.

Profile - The characteristics of a population or market.

Probe - The interviewer asks respondents for more detail by asking a supplementary question, eg 'what else?'.  Not to be confused with 'prompt'.

Projective techniques - Interviewing techniques designed to de-personalise the subject matter or draw out information that is difficult for interviewees to articulate.

Prompt - The interviewer suggests possible responses, eg by showing or reading out a list of possible answers.

Proposition - A concept with a persuading argument or reason to buy. 

Qualitative research - Research methods that explore topics through loosely structured group discussions (aka focus groups) or depth interviews with small samples of people.  Designed to investigate attitudes, behaviours, motivations, etc before trying to measure them.  Particularly important if the subject matter is new or sensitive.

Quantitative research - Research methods that produce statistics based on counting or measuring (eg opinions, behaviours). Involves interviewing or observing large samples. 

Random sample - 'Random' means having no pattern.  A random sample is one where there is no pattern influencing who is selected, so everyone has an equal chance of being included. 

Recall - What people can remember about something.

Refusal rate - The number of people who decline to participate in a survey as a proportion of all those invited to participate.

Representative sample - A sample that has been designed to have the same characteristics (eg age, gender, brand users) as the total population.

Response rate - The number of people who participate in a survey as a proportion of all those invited to participate. 

Sampling error - Aka 'standard error' and 'standard deviation'.  Because survey research is based on samples, rather than whole populations, all results are estimates.  Sampling error measures the precision of the estimates.  See also 'confidence limit'. 

Scale - A range of mutually exclusive answers organised in a logical progression, eg smallest to biggest, best to worst, etc.  Typically used to collect information such as respondents' age or income, satisfaction ratings, strength of opinion, prices, etc.      

Secondary research - Research based on data that have already been collected and analysed for another purpose.

Segmentation - Dividing a population or market into segments for the purposes of explaining or predicting behaviour and, ultimately, for targeting. Everyone within a segment should have the same characteristics and be different from people in the other segments.  

Significant - If two percentages or scores are said to be 'significantly different', it means that the difference is real and unlikely to be due to sampling error.  It does not necessarily mean that the difference is important.

Social class - Classifying people according to their occupation.

Social grade - Classifying people according to the occupation of the head of their household/chief wage earner.

Spontaneous - As in 'spontaneous awareness', 'spontaneous recall' or 'spontaneous response'.  Answers to questions given without the interviewer showing or saying anything to jog interviewees' memory or direct interviewees' thoughts. Eg 'What brands have you heard of?' rather than 'Have you heard of brand x?'  Aka 'top of mind'. 

Standard deviation - See 'sampling error'.

Standard error - See 'sampling error'.

Statistics - Numbers obtained from research.

Stimulus material - Items used to generate a reaction or discussion in an interview or focus group.

Strategy - The plan of activities by which you will achieve your vision or ultimate goal.  Not a research term, but one which crops up in corporate development research.  Not to be confused with 'stratagem', which is a scheme to deceive someone.

Summary - The main points in fewer words.  Not to be confused with 'Conclusions' (see above).

Survey - Systematic collection of data (counts or measures) from a place or population.  Surveys are usually based on samples, so key elements of any survey are the sample design and the data collation method.

Top of mind - See 'spontaneous'.

Trend - A pattern of results over time.  There must be at least three measurements, from three different points in time, to begin to build a trend.  Not to be confused with 'tendency'.

Vision - The ultimate goal or state of affairs you are aiming for in the future.  Not a research term, but often crops up in corporate development research.

Weighting - Giving something greater or lesser importance.  Usually a way of re-balancing the characteristics of a sample to remove bias and ensure it is representative.