Research Glossary

  Previous topic Next topic JavaScript is required for the print function Mail us feedback on this topic! Mail us feedback on this topic!  


Here's an overview of some of the most often used research terms.


Algorithm - A statistical procedure or formula.


Arbitron - The primary ratings service.  They initially contact people by phone, then place a one-week diary for each person over 12 years old in the house.  Participants write down their radio station listening for a week; this is compiled into monthly rolling trend and quarterly (4X/year) ratings reports.  Advertising rates and revenue is determined by success in Arbitron.


AQH (Average Quarter Hour) - This is the number of people, on average, that are listening to a station during a designated time period.  It's called "quarter-hour", because they have to listen for at least 5 minutes out of any 15-minute "quarter" of an hour to be counted.


Auditorium Music Tests (or AMT's) - Unlike weekly Callout research, some stations conduct once or twice-yearly AMT's.  Similar to callout, representative listeners are recruited over the phone (based on station listening, demo, gender, ethnicity, etc.)  However, instead of responding to song hooks over the phone, they are recruited to show up at a nearby hotel ballroom with about 100 other listeners, where they grade up to 700 song hooks.  They're generally compensated $50-75 for this.  AMT's are generally used to test oldies and recurrents, whereas weekly Callout is used for researching current songs.


Callout - The process of calling passive radio listeners on the phone, playing song hooks to them over the phone and getting them to score each. Unlike "active" forms of research, such as taking requests, responding to e-mail or website polls, callout reaches out to the passive listener who would not otherwise respond to other polling means.  It's been proven that the tastes of these passive listeners (the bulk of a station's audience) are vastly different than the more out-going active listeners.


Closed-ended question - The closed-ended question provides response options in the survey that can be answered quickly and easily. Examples of closed-ended questions are those of the Yes/No or PickList type.


Confidence Level - The number of times out of 100 that the true value will fall within the confidence interval.


Cume - Cumulative audience.  Cume is like circulation of a newspaper.  It's a count of how many different people (in a specified demo and/or time period) listened to a station for at least 5 minutes.


Cume Duplication - The percentage of estimated Cume Persons for one station who also listened to another station.


Demographics - Age groups.  Typical Arbitron age groups are 12-17, 18-24, 25-34, 35-44, 45-54, 55-64 and 64+. Some cells can be broken down further; for example 25-34 can be split into 25-29 and 30-34.


Exclusive Cume - The estimated number of Cume Persons who listened to only one station within a reported daypart.


HDA (High Density Area) - A Zip-code defined area that may be established in a county within a the Metro of an ethnic-controlled market (High Density Hispanic or High Density Black).


Hook - Song Hooks are the most familiar/memorable :07-:10 seconds of a song.  This is what is tested in Callout or Auditorium music testing.  Generally (but not always), it's the part of the song where the song title is sung.


Hot Zips - Arbitron distributes diaries to people in Zip Codes, based on population.  Each station will perform better in Arbitron in certain Zip Codes in their market (they'll receive mentions in more diaries in some Zip Codes than in others.)  These are the stations' "Hot Zips"; we try to target these Zip Codes when ordering a list of names/phone numbers from list vendors (SSI or InfoBase).  This assures a higher probability of reaching listeners for that station (and/or their direct competitors).  Stations can determine their Hot Zips (and pass them along to ComQuest), by running a 4-book trend of Zip Code Diary Distribution from their Arbitron Programmer's Package.


In-Tab - The number of usable interviews tabulated in producing a report.


Likert Scale - Respondents are asked to indicate their degree of agreement or disagreement on a symmetric agree-disagree scale for each of a series of statements.


Mean - The arithmetic average of a set of scores.


Measurement Error - An inconsistency produced by the instruments used in a research study.


Median - The midpoint of a distribution of scores.


Metro Survey Area (MSA) - The primary reporting area for local radio.  MSA's generally conform to the federal governments' metropolitan areas.


MetroMail - Another list vendor (see "SSI")  This is the company Arbitron used to use for contacting diary keepers.  After more than 20 years, they moved to SSI in Summer 1998.


Mode - The score that occurs most often in a frequency distribution.


Multivariate Statistics - Statistical methods that investigate the relationship between one or more independent variables and more than one dependent variable.


Open-ended question - An open-ended question presents no response options to the respondent, but rather collects a verbatim response.


Ordinal Level - The level of measurement at which items are ranked along a continuum.


P1's, P2's, P3's.  - The "P" stands for "Partisans".  Every station's audience can be segmented into groups based on how much time they spend with each station, relative to all the stations they listen to.  A station's P1's are those listeners that spend the most time with that station (over all other stations they might also listen to.)  P2's are listeners that spend a secondary amount of time with that station, but listen to another station more.  P3's are people who tune into that station, but it's not the one they spend a lot of time with.  In research, stations attempt to get a mix of (as many) P1 participants combined with P2 (and/or P3) respondents in their research.  P1's are also known as the "core" audience. A station can garner as much as 80% of it's Average Quarter Hour Ratings from the most loyal 20% (P1's) of it's audience.


Panels - People who previously took the test.  A percentage of people taking the test each week will most likely be people who previously took the test.  The advantage in using panels is it is easier to get previous test-takers to take the test again than it is to get new people.  Most PD's only allow a certain percentage of panels, because after awhile it seems like you're just researching the same people over and over again.


Probability Sample - A sample selected according to the laws of mathematical probability.


Qualitative Research Method - A description or analysis of a phenomenon that does not depend on the measurement of variables.


Quantitative Research Method - A description or analysis of a phenomenon that involves specific measurements of variables.


Random Sample - A subgroup or subset of a population selected in such a way that each unit in a population has an equal chance of being selected.


Rating (AQH or Cume) - The estimated percentage of the demographic population listening to a given station or to total radio during a specified period of time.


Recurrents - Songs that were previously played on a contemporary radio station as a current (several weeks or months ago).  Recurrents are important to a station because they are generally the most familiar and (formerly) popular songs.  However, program directors like to know when these songs eventually "burn-out" so they can either reduce the frequency they're played, or quit playing them altogether.


Sample-specific Results - Research results that are based on, or specific to, the research sample used.


Sampling Error - The degree to which measurements obtained from a sample differ from the measurements that would be obtained from the population.


Share - The percentage of those listening to radio in the Metro who are listening to a particular radio station.


Song Scoring Attributes - Here are the different attributes we are able to determine from the collection of song scores:


Familiarity - the percentage of people that know the song

Positive Acceptance - the percentage that said they liked it.

Burn Factor - the percentage of people that are tired of the song

Dislike - the percentage that don't like the song

So-So - the percentage that are indifferent

Favorite - the percentage that said this was one of their favorite songs.


SSI (Survey Sampling Inc.) - One of many list vendors; companies that sell names/ phone numbers of residences in specific demographics/ zip codes.  This is the company that Arbitron Ratings now uses for contacting diary keepers.  As such, some stations prefer to use this company's sample for loading names into their ComQuest systems.


Time Spent Listening (TSL) - An estimate, expressed in hours and minutes, of the amount of time the average radio listener spent with a station during a particular daypart.


See also

Computer Glossary