5: Chapter The Arbitron Report
Before you can interpret the Arbitron report, it is necessary to become familiar with some important concepts and formulas associated with the estimates that are reported by Arbitron. Ratings A rating is thought of as percentage of potential audience. You may hear the terms "rating points" or "points". These are short for "percentage points". Ratings can be based on all listeners, on a specific subgroup, or on households (for TV). Finally, ratings are "historical" in that they show how a station or program performed in the past. Rating are based on the population. The population is the total potential audience (people or households) for a station or a program. It is the number of persons equipped with radios or TV sets. In the case of cable TV, ratings for national cable networks are based only on the number of homes that can receive the networks' programming. In calculating a rating, the base figure is the number of radio or TV households (for cable the number of subscribers) or the number of individuals in a particular demographic group (e.g., females 18 years or older, written as F 18+). For national TV ratings the base is the national universe of approximately 100 million TV households. Thus one rating point (1%) equals about one million homes. For radio the base figure would be all persons in a specific demographic group. In radio, estimates are reported for the basic time unit of the quarter hour. The average quarter hour (AQH) provides the average number of listeners over all the quarter hours making up a daypart. For example, M-F 6-10 am, known as morning drive, has 16 quarter hours per day or 80 quarter hours per week. The audience is reported as an average across those 80 quarter hours. Remember, a person only has to listen for 5 minutes in the quarter hour to be counted. Since a person could listen to two different stations for at least 5 minutes in a quarter hour, he or she could be counted in two program audiences in a quarter hour. The national People Meter sample (NTI) has the most narrowly-defined audience rating - the size of the audience in an average minute within a quarter hour. The average number of persons estimated to have been listening at least five minutes in any quarter hour is referred to a AQH persons. For radio, this is usually abbreviated as AQH (00). The (00) means that two zeroes are added to the figure that is reported in the Arbitron report. It is understood that an advertiser who buys a spot during a particular quarter hour will presumably reach the number of persons indicated by AQH (00). In radio the various time periods are referred to as dayparts. The listening day is divided into these established segments. They include morning drive (6-10 am), midday (10-3 pm), evening drive (3-7 pm) and night (7 pm-midnight). While radio will report audience size and composition in terms of broad dayparts, TV ratings are reported for each quarter hour of the day because TV advertisers want more precise information. Audience segments for radio and television are defined as demographic groups. These are standard age and sex groupings. In the Arbitron report there are various grouping such as M 18-24, F, 18-24, M 25-34, F 25-34, etc. The first estimate we will look at is the average quarter hour rating or AQH rating. It is a convenient way to represent the average size of the audience over a specified period of time. It is calculated using the following formula:
Average Quarter Hour Persons listening to a station X 100 = AQH Rating** Survey Area Population
** for a specific demographic group and time period
Another way of expressing the size of the audience is the projected audience. It is a basic gross measurement of audience. To know a station has a certain rating doesn't tell us how many people are in its audience. The projected audience is an estimate of absolute audience size; it makes audience size more interpretable. Audience projections can be made for specific programs, stations, all people using the medium at a point in time, etc. Most of the numbers reported in a Nielsen or Arbitron report are simply estimates of absolute audience size. Q: Station "A" has a 25 rating in one city and "station "B" has a 15 rating in another city. Station "B" has a larger projected audience. How is this possible? A: Station "B" is in a city with a larger population. Rating (%) X Population = Projected Audience
For example, we wish to know how many Men 18-49 (M 18-49) are listening to WABC during evening drive time (M-F 3-7 pm) in the Metro area. We know the following: WABC AQH Rating (M-F 3-7 pm; M 18-49) The total number of M 18-49 in the Metro = 2.6 = 62,600
WABC's projected audience = .026 X 62,600 which equals 1,628. Thus, we estimate that an average of 1,628 M 18-49 were in WABC's audience during an average quarter hour of evening drive in the Metro area. A different rating and estimate of the population would be used for the Total Survey Area. NOTE: In local market samples, weighting may be used: a statistical adjustment or balancing of the responses of underrepresented groups. These weighted responses would be used in projecting audience size. To express the absolute size of the radio audience at a single or average point in time we use the PUR (persons using radio) figure. Shares In addition to a rating, the other major estimate that is commonly used in radio and television is the share. A share is the listening/viewing audience expressed as a percentage of all persons listening to radio or watching TV at that time. The share of listeners is expressed in terms of: (a) a specific demographic group (or all persons), (b) at a specific period of time (program, daypart, quarter hour), and (c) in a specific geographic area. A program share is always larger than the program rating. A comparison of the formulas for rating and share confirm this. The share is used by advertising buyers/sellers to project future audience size. However, shares are of somewhat limited value in buying/selling audiences because
they indicate performance relative to the competition. They don't convey information about actual audience size in which advertisers are most interested. Average Quarter Hour Persons listening to a station X 100 = AQH Share** Survey Area Total Listening Audience **For a specific demographic group and time period Target Audience Efficiency A measure of how well a station covers a target audience is referred to as target audience efficiency (TAE). A more precisely targeted station is more efficient in delivering a desired audience. Stations AQH (00) (target) Stations P12+ AQH (00) **for a particular daypart For example, consider the following two stations (Area = Metro; time period = M-F 3-7pm): AQH W 18-34 700 2400 AQH P12+ 2200 4800 TAE 31.8 50.0 X 100 = TAE**
WBZE WTNT
The station with the higher TA is more efficient in reaching W 18-34. That is, a larger share of WTNT's audience is made up of its target the listeners it most wants to reach. CUME Thus are we have considered terms and formulas for gross measures. Now we turn to the cumulative measures. These measures play a larger part in radio audience measurement and advertising for radio. Cumes (described below) are preferred in radio more than TV because radio ratings/shares are generally low. A cume is the number of different persons or households who listen or watch for at least five minutes in an average quarter hour of a daypart (or a day, week, even a month). Each person is cited only once whether he/she listens five minutes or five hours. Cume is also referred to as reach or total unduplicated audience. The weekly cume is considered a measure of circulation or total audience. It is comparable to the circulation figures used by newspapers and magazines. Radio's main competitors are local newspapers. Radio advertisers want to know how many and what types of people they are reaching (and not how long someone is listening). Cume figures provide the necessary information. Cumes are reported for different subsets of audiences as well. Radio cumes (and TV cumes in local nonmetered markets) are limited to one week due to the use of a seven-day diary. The actual number of persons estimated to be in the cume audience is referred to as cume persons. It is often written as cume (00). Cume cannot be added together across stations or across dayparts on a single station because without specialized ratings reports, it is not possible to determine the percentage of audience that is common to the stations or daypart.
For every pair of stations in the radio market, Arbitron estimates the number of listeners who are in both stations' cume audiences. This is referred to as cume duplication. In other words, it is an indication of which stations have audiences in common. An exclusive cume is the estimated number of cume persons in a survey area who listen only to one station during any given daypart. All other things being equal, a large exclusive audience may be more saleable than one that can be reached by several stations. When a cume is expressed as a percentage of the total possible audience it is referred to as a cume rating. Cume ratings are rarely used. Cume Persons listening to a station Survey Area Population X 100 = cume rating**
* for a specific demographic group and time period Audience Duplication Another concept of listening/viewing behavior is audience duplication. Studying patterns of audience duplication is an important technique of analysis for inheritance effects, audience flow, station/channel loyalty, and repeat viewing. However, to study audience duplication, one needs access to raw data. When the data are available, the researcher can examine of the people who were watching/listening at different points in time. Audience duplication in radio across broadly-defined dayparts is called recycling. Audience duplication across several points in time produces reach and frequency data. Time Spent Listening A measure of how well radio maintains its audience is that of time spent listening (TSL). It is the amount of time the average person in the cume audience spends listening to a specific radio station during a particular daypart. It combines gross and cume measures. A station likes to have larger TSL estimates. It is possible that a station with a high TSL has a relatively small number of heavy listeners whereas a station with low TSL may have a larger number of listeners. In this case, advertisers will favor gross audience size. Number of quarter hours in daypart X AQH (00) Cume Audience (00) for daypart = TSL
For example, the table below compares the TSL for two stations. It focuses on the following: Metro area, evening drive time (3-7 pm), and the Women 18-34 demographic. Keep in mind that evening drive consists of 80 quarter hours (5 days X 4 hours X 4 quarter hours). AQH W 18-34 WBZE WABI 700 700 Cume W 18-34 7600 3900 (QHs) TSL 7.37 14.36
For both stations, the numerator in the formula will be 80 X 700 or 56000. Dividing 56,000 by each station's cume yields the TSL in terms of quarter hours. To convert to hours, divide by four:
QHs WBZE WABI 7.37 14.36
Hours & minutes 2 hr. & 7 min. 3 hr. & 51 min.
What does this tell us? Women 18-34 who listen to WABI spend nearly twice as much time listening to the station than women who listen to WBZE. WABI is able to hold its female audience longer. A look at the formula reveals that, the larger the cume (AQH (00) being equal), the lower the TSL. In this case, a larger number of different W 18-34 make up WBZEs AQH audience. Turnover Another indicator of audience maintenance is turnover: how many times the audience turns over or how rapidly different listeners cycle through the station's audience. Turnover can be thought of as the number of new listeners a station must attract in a time period in order to replace those who are tuning out. The lower the turnover factor, the more loyal the stations audience (the more the station maintains audience). The higher the turnover factor, the faster the rate of cume growth (the more different quarter-hour audiences are included in the cume). A turnover ratio of one would mean that the same people are in the audience hour after hour. It also would mean that during the daypart under consideration, all of the different persons who listen to the station throughout the week are tuned in during an AQH. Turnover can tell us about audience maintenance but not about audience size. [Source: J. Fletcher, Profiting from Radio Ratings, p. 49] Turnover is calculated using both gross and cume measures. CUME (00) in a daypart AQH (00) in a daypart 7 pm): AQH W 18-34 WBZE WAIB 700 700 Cume W 18-34 7600 3900 Turnover 10.9 5.6 = Turnover factor
For example, consider these two stations' audiences in the Metro during evening drive (M-F 3-
WBZEs audience turns over more often (nearly 11 times across the daypart or 11 audiences of 700) than does WAIBs audience. WAIBs audience is more loyal. A station would prefer higher AQH and lower cume. Comparisons Using Cumulative Measures Cumulative measures are not used as extensively in buying and selling. They are used more to get a deeper understanding of audience behavior. Programmers find cume measures useful for programming analysis. One can compare the TSL and exclusive cumes of stations with similar formats (stations with similar formats are likely to "share" their audiences). TSL can be computed to compare men and women in various age categories to show who are the heaviest and lightest radio users in a specific market. A radio station changing format could monitor its cume audiences
(exclusive and duplicated cumes). Successful repositioning should yield increased duplication of audience with another station in the format. The competitor station's exclusive cume should be reduced as well. Adding Estimates Together Care should be taken when using audience estimates (such as those derived from the Arbitron report). Only certain estimates can be added together. The general rules are: 1. Do not add together daypart estimates (dayparts may contain different number of quarter hours) 2. Do not add ratings and shares for different demographic groups because the base may be different. 3. Cumes cannot be added together for stations (and dayparts) due to duplication you cannot count the same person twice. The following table summarize what can or cannot be added: Estimate AQH (00) AQH Rating AQH Share CUME (00) CUME Rating Station+ Station YES YES YES YES NO Demo+ Demo YES NO NO YES NO Daypart+ Daypart NO NO NO NO NO
In general: ratings may be added together as long as they have the same base. Do not add ratings/GRPs from different, individual markets (e.g., cities). Do not add GRPs for different demographic groups. Reach and Frequency The cume measure used by advertisers and media planners is referred to as reach. It is the estimated number of different people reached by an ad schedule. In other words, it is the number or percentage of different people or homes exposed at least once to an advertising schedule over a specified period of time (excludes duplication). It is the same as cume the unduplicated audience a station delivers. Reach can be expressed as percentage of the total possible audience (a rating) or as different persons. Reach is generally estimated for a four week period of time. Thus, it is easier to compare TV network schedules to magazines. Most ad agencies have computer programs available to predict or project the reach of a media schedule based on input like GRPs [see Chapter 10]. We will not attempt to calculate reach in this course. A companion estimate to reach is that of frequency. It is the average number of exposures among those who were reached. That is, it is the average number of times the unduplicated listener will hear an advertising message. We might say that an ad campaign reached 80% of the population, but that it did it with an average frequency of 2.5. Such a campaign would generate 80 X 2.5 or 200 GRP's. Reach X Average frequency = GRP's
Usually agencies and audience researchers only know the GRPs for an ad schedule. They will use mathematical models to estimate reach and frequency. We typically calculate frequency using this formula: Gross Impressions (GI) for a spot schedule Cume = Frequency
Based on this formula, we can see that, the lower the cume, the higher the frequency (assuming GI is equal for two stations). In advertising effective frequency or effective exposure refers to how many times a commercial must be seen or heard before it is effective. An assumption of media planners is a minimum of three exposures is necessary for an ad to be effective. The Arbitron Report Now we will examine the Arbitron radio report for our market. Here we will have an opportunity to apply many of the terms and formulas we have learned. The photocopy set contains only a portion of the full report. Through 1995, Tallahassee only had one report a year, done during the Spring sweep. Beginning in 1996, Tallahassee had both a Spring and Fall book. We will move through the various sections of the report beginning with the map showing the counties making up the Tallahassee Metro, TSA, and DMA. To most effectively use this chapter, try to answer the questions that are asked [try to locate the appropriate numbers from the Spring, 2002 Tallahassee Arbitron report. Then, check your answers against the answers that are supplied. Below we grow through the contents of the local radio report, section by section. Market Segments (Map p. 2) The Tallahassee Metro area (consisting of Leon and Wakulla Counties) was ranked #162 in Fall, 2000 [out of 286 markets in the U.S.]. The Metro generally corresponds to the federal governments Metropolitan Statistical Area (MSA). It is the central city together with county or counties in which it is located. The Metro is also thought of as the county or counties containing the cities in which most of the local radio stations are situated/licensed. The Total Survey Area (TSA) is defined as the Metro area plus additional, surrounding counties. The TSA is based on historical radio listening patterns. A county outside the Metro is assigned to the TSA if it accounts for a minimum number of in-tab diaries (10) that ascribe listening to the Metro stations. For the Tallahassee market, the TSA consists of nine counties from Florida (including the two Metro counties) and eight counties from Georgia. A county can be assigned to more than one TSA. That is, a county on the fringe of Tallahassee's TSA may also be assigned to a neighboring market's TSA as well. The Designated Market Area (DMA) is a term derived from television. Nielsen has assigned every county in the U.S. to a market area referred to as a DMA. Each U.S. county is assigned to one and only one DMA. In the map at the front of the Arbitron report you can see a thick black line that encircles a number of North Florida and South Georgia counties. Some of the counties in the DMA are outside the Tallahassee TSA. The DMA is irrelevant for Tallahassee radio because Arbitron only reports radio listening estimates for the top 50 DMA's (we are well below that). Population Estimates (p. 3)
This table shows the estimated P 12+ population and in-tab diaries for each county in the TSA. For example: Leon (a Metro county) has a population of 210,800 P 12+ and 589 in-tab diaries (its 589 in-tab diaries represent 94.1% of the total of the Metro (n= 626 diaries) and 54.7% of the total intab diaries for the TSA (n=1076). Population Estimates and In-tab Diaries by Demographic Group (p. 4) This table provides population estimate for various demographic groups in the Metro and TSA. This is an important table that contains population estimates required for calculating various estimates. For example, read the line for M 18-24 (Metro): 25,500 males 18-24 in the Metro they represent 11.0% of the 231,100 P 12+ in the Metro 44 in-tab diaries were collected from this demographic group the 44 diaries represent 7.0% of the 626 in-tab diaries in the Metro the 44 diaries represent 11.0% of all Metro diaries when the data are weighted Sample balancing or weighting is done, in part, to compensate for too many or too few diaries from a particular demographic group. For instance, M 35-44 are estimated to comprise 7.8% of the Metro population. However, only 6.5% of the in-tab Metro diaries came from M 35-44. Therefore, weighting will be applied to bring up the contribution of M 35-44 to 7.8% of the Metro. Diary Placement/Return (p. 4) This table shows the record of diary placement in the Metro and TSA for the Spring, 2002 sweep. As you can see, Arbitron lists unlisted and listed telephone households/persons separately. We will look at the combined total ("Tot") for the Metro: 911 estimated usable households containing 1,908 estimated usable persons were in the sample for the 12 week sweep 853 of the households were contacted 583 households accepted diaries (thus 64.0% of sample households) 1,230 persons were sent diaries (thus 64.5% of sample persons) 626 in-tab diaries (out of 1,230 persons) or 50.9% 610 diaries had been the target for the sweep; Arbitron achieved 103% of its target NOTE: The Spring, 2002 survey ran from 3/28/02 - 6/19/02. Station Information (p. 5-6) PLEASE NOTE: THE TALLAHASSEE RADIO MARKET MAY HAVE CHANGED SINCE THE SPRING, 2002 REPORT (E.G., FORMAT AND CALL LETTER CHANGES). These pages provide information (particularly for use by those outside the market) on the format, sales representative, etc. of stations in the market (that meet minimum reporting standards). You will notice that Arbitron is limited to commercial radio stations. Therefore, WVFS-FM ("V-89"), WAMF-FM, WFSU-FM and WFSQ-FM do not appear in the report. Note that eight formats are represented by the 13 commercial FM stations in the market:
Station format Adult contemporary (AC)/Hot AC Album-oriented rock (AOR) Alternative Christian/Gospel/Religious Country Rhythmic Oldies/Classic Hits Pop contemporary hits Urban contemporary (UC) Metro Market Profile
Number of stations 2 1 1 1 2 2 1 3
Arbitron also provides economic and demographic information about the market. This information can be of use to local, regional and national sales people who need to know about the market or wish to be able to profile the market to advertisers. One can use these pages to determine our market potential as well as to educate those outside the market to the qualities of the listeners/ consumers. Sales people who wish to trumpet the strengths of radio as a medium can use the information (e.g., W 18-34 spend 15.75 hours a week listening to radio and 97.1% of these women are reached by radio over 24/7; see p. 9). The metropolitan area data are updated from the 1990 Census, or are from the Audit Bureau of Circulation (newspapers/magazines), from PRIZM, from Arbitron itself, and from other sources. This section gives you some indication of the variety of data from secondary sources that can be used to profile a market. The Area Lifestyle Profile (p. 7) is derived from a geodemographic segmentation system. All zip codes in the U.S. have been grouped into 62 clusters (zip codes with many similarities of lifestyle) by Claritas, Inc. To make it more manageable, the PRIZM system further organizes these 62 clusters into 15 broader social groups that are summarized in the table. The chart shows the national composition for each group as well as the percentage of persons in the Tallahassee Metro within each group. Example: C2 2nd City Centers (mid-leve, white collar couples, mid-income families and college town singles). 9.03% of the U.S. population 12+ into fall this group, but in the Tallahassee Metro 40.69% fall into this group. Thus, about four and one-half times the national figure occurs in our area, giving Tallahassee an index of 442. Study over the tables in the Metro Market Profile sections of the report. Ask questions of your instructor if you are not clear on their contents -- you are responsible for knowing how to read and interpret these (as well as all other tables) in the report. Target Listener Trends (p 14+) Beginning with this trends section, the Arbitron report provides gross and cumulative measures of the audience in the Metro. Target Listener Trends shows how stations are doing over time. It can indicate how programming and format changes (by a station or its competitors) impact a station's audience size. It can also show how consistent a station is in drawing an audience. If there is a change in a station's audience, it will show the direction of change. Changes in a station's performance during different dayparts can be detected (e.g., is it improving in the a.m. but declining in the p.m.?). The bulk of the Arbitron report is made up of the Trends and its companion section, Target Listener Estimates. The SP' 02 book compares: SP '00, FA '00, SP '01, FA' 01 and SP '02. It also gives a two book average for FA '01 and SP '02.
NOTE: With the exception of the last section of the Arbitron report, all sections report data for the Metro area only. Also, in all discussions and examples, we exclude stations outside the Metro area (WGWD-FM). The trends section of the report provides: 20 of the most requested demographic groups: P 12+, P/M/F 12-24, P/M/F 18-34, P/M/F 1849, P/M/F 25-49, P/M/F 25-54, P/M/F 35-64, Teens 12-17 Five dayparts: M-S 6am-mid, M-F 6-10, M-F 10-3, M-F 3-7, M-F 7-Mid Four estimates: AQH (00), Cume (00), AQH rating, AQH share NOTE: The totals on the bottom of the last page of each demographic group are for all stations in the market. Q: Which FM stations show the largest increase and largest decrease in P 12+ AQH share during Mon-Sun 6am - mid. between SP '01 and SP '02? What is the percentage increase from SP '01 to SP '02 for each? A: FM WHBX WGLF WBWT WBZE WAIB WHTF WXSR WWLD WTNT WTLY WEGT WVHT WAKU SP '01 12.6 6.5 4.1 7.2 3.4 6.5 4.8 4.4 5.5 4.4 3.8 2.0 1.0 SP '02 % Change Frequenc 12.7 0.79 96.1 7.3 12.31 104.1 6.7 63.41 100.7 5.7 -20.83 98.9 5.7 67.65 103.1 5.1 -21.54 104.9 5.1 6.25 101.5 4.4 0.00 106.1 4.1 -25.45 94.9 3.8 -13.64 107.1 3.2 -15.79 99.9 1.3 -35.00 105.7 1.3 30.00 94.1 Format
UC AOR UC AC Country
Pop. Con. Hit
Alternative Rhythmic Old. Country AC Classic Hits UC Christian
To calculate the difference between SP '01 and SP '02, compute the following: AQH share SP '02 - AQH share SP '01 AQH share SP '01 X 100
We compare spring-to-spring in part because we assume that conditions that might affect radio listening are more alike. Thus, differences are more likely to be due to station programming and personalities. Keep in mind that few stations would be concerned about a broadly-defined audience such as P 12+. Stations have specific target audiences (e.g., P 18 - 49). Q: For SP '02, what is the AQH (00) for all stations in the Metro (for M-S 6am-mid, P 12+)? A: 31,500 note that Arbitron truncates it as 315 you must add two zeros; bottom p. 16]
Q: For SP '02, how many more persons were listening in an AQH during the morning drive period (610 am) compared to the evening drive period? A: 46,000 - 41,300 = 4,700; bottom p. 16 Q: Again, for SP '02, for M-S 6am-mid, how many different P 12+ listened to at least one station at least once (5 minutes in a quarter hour) A: 216,500; Cume (00); bottom p. 16 Q: What proportion of P 12+ who live in the Metro listened to radio at least once across the week? A: 216,500 / 231,100 = 93.7% that is, 94% of the Metro P 12+ population listened to radio at least once a week. This compares to 92.4% in FA '00. The Metro P 12+ population figure derives from the table on page 4 of the Arbitron report. Q: The AQH rating for M-S 6am-midnight (SP '02) is 13.6. How was the estimate derived? That is, what figures were used to calculate the rating? A: 31,500 = AQH (00) 231,100 = population P 12+ (from table on page 4)
Reliability of Estimates Estimates are just that - estimates. We can never really know the actual number of people listening to a particular station at a particular time. The Arbitron diary, like all sample surveys, is subject to statistical variation or sampling error. If we did a survey 10 times, we might obtain 10 different figures for the number of listeners to a station. We try to gauge how close our survey estimate is to the population value if that could be known (e.g., if we did a census). The statistical measure of sampling error is called standard error. The standard error of any estimate (applied to the estimate; plus or minus) produces a range called a confidence interval. That is, if the estimated value is 6, knowing the standard error, we can place an upper and lower limit around the value and say that the population value falls somewhere in that range or confidence interval. We also can state how confident we are that a survey of the total population would yield an estimate within that confidence interval. When a confidence interval is equal to plus or minus one standard error, we can say we are 68% confident that a survey of the total population would yield an estimate within that confidence interval. The level of confidence increases as the we increase the number of standard errors. In general, the smaller the standard error, the more reliable the estimate. Standard Error can be calculated as follows: square root of p X q n p = the rating; q = 100-p n = sample size
However, because Arbitron uses a complicated sampling procedure instead of a simple random sample (SRS), "n" is replaced with ESB (Effective Sample Base).
square root of p X q ESB
p = the rating; q = 100-p ESB = Effective Sample Base
Arbitron conducted extensive research to be able to provide Arbitron users with an easy method of estimating sampling error in the market report. To calculate the standard error of an Arbitron estimate, we use Tables A and B in the rear of the market report (pages M1 and M2). Table A: This is the numerator (p X q) in the standard error calculation. It contains the square roots for ratings ranging from 0.1 to 50.0. The table is the same in all reports/markets. Example: M 18-34 M-F 6am-10 am (SP '02) AQH rating for WGLF is 3.1 (p. 54). According to Table A, the value is 17.33. This is equal to "p X q". Table B: This is the denominator of the standard error calculation (ESB). These values are customderived for each market and survey. It is customized because each report has different in-tab totals for various demographic groups and different weighting is applied to the age/sex/survey area. We use Table B to find the appropriate demographic group, daypart, and estimate. Example: Continuing with the above example, we find the figure that represents M 18-34 and morning drive (note that nearly all figures are for AQH estimates). The figure is 17.48. Now, we do the calculation: se = 17.33 ------- = 0.9914 or 1.0 which is one s.e. 17.48 Thus, we are 68% confidence that the "real" rating for WGLF (M 18-34, morning drive) is 3.1 1.0 or falls within the range of 2.1 to 4.1. If we want to be 95% confident (2 se), the confidence interval would be 3.1 2.0 or 1.1 to 5.1. Thus we are 95% certain that between 1.1% and 5.1% of all M 18-34 in the Metro were listening to WGLF during morning drive. Notice that, as the confidence level increases, the error band also increases. Q: Now, what if the rating is smaller? WAIB's AQH rating for M 18-34 for 6am-10 am is 1.1 (p. 54), about one-third that of WGLF (WAIB's AQH (00) is 500 compared to WGLF's 1,400). What is the confidence interval for WAIB's rating? A: se = 10.43 ------= .5967 = 0.6 which is one s.e. 17.48 (note: this is the same as above)
We are 68% confident that the rating for WAIB (M 18-34, morning drive) is 1.1 0.6 or falls within the range of 0.5 to 1.7. At the 95% confidence level (2 se), the confidence interval would be 1.1 1.2 or 0.0 to 2.3. Target Listener Estimates (p. 17+) This section of the Arbitron report eliminates the data from previous reports (SP '00, etc.). It does, however, show a 2-book average. The Estimates section provides the same 20 demographic breakouts and the same four estimates as the Trends section. The Estimates portion replaces the five
major dayparts of the Trends section with 10 new dayparts which largely cover the weekend (the only nonweekend daypart is M-F 6am-7pm). For the first questions, we'll look at W 18-34 (we'll refer to both the Trends and Estimates pages in the following questions - p. 84+). Q: How many different women 18-34 listen M-F 6am-mid but not on the weekend (SP '02)? A: From the Trends section we find that the SP '02 M-S 6am-mid Cume is 41,500. The weekend 6am-mid Cume is 34,400; thus 41,500 - 34,400 = 7,100; 7,100 different W 18-34 listen during the week but not on the weekend Q: What station had the highest AQH share of W 18-34 (SP '02) during the M-F 6am-7pm period? A: WBWT-FM with 13.8 Q: How was that figure calculated? A: AQH Share = WBWT AQH (00) = 1,100 8,000 get this figure from bottom of page 87
Q: A station has the same number of people listening during midday and evening drive, and the station has the same rating for both dayparts, but its share decreases from midday to evening drive. What could explain this? A. PUR (persons using radio) increased; with more listening in the later daypart, thus, the station's proportion of all listeners is smaller (see table below) Daypart ` midday p.m. drive AQH 2,000 2,000 Rating 3.5 3.5 Share 10 5 Listeners 20,000 40,000
Q: A station (WTNT) has a M-F a.m. drive-time AQH rating for W 18-34 (SP '02) of 0.6 (p. 85). Knowing that there are 48,100 W 18-34 in the Metro, (1) how many women were listening to WTNT station in an AQH (calculate this)? (2) How is the population figure arrived at? A. (1) 289. This is an example of projected audience - we project the audience by the formula: Rating X Population. Look at WTNT's AQH (00) figure -- it is 3 (00). (2) The population figure is sum of the women's age subgroups 18-24 and 25-34 (found on p. 4). Q: Which station had the highest unduplicated audience for W 18-34 in SP '02 across the week (Mon-Sun, 6am-mid)?
A: Cume (00); WBWT with 20,900. This is its unduplicated audience or reach within this demographic group.
Q: What is the Cume Rating for WBZE during evening drive in SP '02for W 18-34? A: Cume (00) = 7,800 Population = 48,100 = 16.2
Thus, about one out of six W 18-34 in the Metro area listened to WBZE at least once during evening drive. Q: Next, let's consider the concept of turnover -- how many different groups of persons make up a station's audience. Look at WXSR for the entire week (Mon-Sun, 6am-mid) for W 18-34. What is WXSR's turnover factor in SP '02? A: CUME (00) for week AQH (00) for the week 11,200 = 500 = 22.4
Q: Now, let's look at the turnover for WBZE in SP '02? A: CUME (00) for week 12,500 AQH (00) for the week 200 = 62.5
WBZE turns over its audience almost three times more often than WXSR. By comparison, WXSR does a better job of "maintaining" its audience. In other words, more listeners "cycle through" WBZE during the week. Q: Can we add the AQH (00) for M-F morning drive and the AQH (00) for M-F midday for a station to get a total AQH (00) for the two dayparts? Yes or no and why? A. NO. Morning drive has 80 QH's, but midday has 100 QH's. Since they have a different number of QH's, we cannot add the estimates together. Listener Composition This section of the Arbitron report shows how much each demographic group contributes to a station's AQH and CUME 12+ audiences. We use it to find which stations reach the largest share of a particular demographic group and to determine if a station's audience is highly concentrated in a demographic group or is spread across a broader demographic target. As noted earlier, this section, like all sections except the final portion of the report, only provides figures for the Metro. Listener Composition includes:
One daypart: Mon-Sun 6am - mid Estimates: For AQH: AQH (00), AQH rating, AQH share, and percent that demographic group's AQH (00) is of total 12+ audience; For CUME: Cume (00), Cume rating, and percent that demographic group's Cume (00) is of total 12+ audience Demographic groups: 16: P 12+, Teens 12-17; M/F 18+, 18-24, 25-34, 35-44, 45-54, 55-64, 65+ Q: What percentage of WGLF's AQH audience is made up of M 35-44 (p. 114)? A: 20% (500 M 35-44 listening in an AQH out of a P12+ AQH audience of 2,300) Q: What two FM stations have the highest concentration of W 25-34 in their AQH audiences? A: WHTF (18%); WXSR (17%) Q: What percentage of WWLD's weekly Cume is made up of M 35-44 (p. 117)? A: 14% Q: WHBX's Cume (00) for M 35-44 is 4,700. State what this number means. A: 4,700 (different) M 35-44 tuned into WHBX at least once during the week. OR: WHBX reached 4,700 different M 35-44 during the week. Both AQH and Cume can be added across cells in this section to form custom demographics. Look at TAE: Are the highest demographic cells consistent with a station's stated target audience? TAE = Target Audience Efficiency Example: WTNT target audience P 25-54: WTNT's AQH (00) for its target audience = [0+200]+[0+100]+[200+100] (for M/F 25-34 + M/F 35-44 + M/F 44-54) = Station's AQH (00) (target) Station's P 12+ AQH (00) X 100
WTNT's TAE = 600 1300
= 46.2
A little less than half (46%) of WTNT's AQH (00) P 12+ is comprised of its target audience. Listening Locations
This section shows how many persons 18+ are listening to a station in one of four locations: home, car, work, or other. Where a person listens may affect the type of message an advertiser wishes to employ (e.g., for heavy in-car listening: new car dealers, auto parts, oil/lube services; for heavy at-work listening: office equipment, office supplies, restaurants). Knowing where the listening is occurring is helpful in determining programming elements such as traffic reports, contests, newscasts, etc. This section contains: Dayparts: M-F comb...