From College To Media
Sac State vs. The Suck State of Mainstream Media
Chapter 7: How Stats Influence Perceptions

Introduction: A Journey Through Mass Media Education
Chapter 1: Why I Majored in Communication
Chapter 2: The Art of Radio
Chapter 3: Writing TV News
Chapter 4: Studying To Be a Rock Star
Chapter 5: Exploring the Age of Incoherence
Chapter 6: The Science of Media Persuasion
Chapter 7: How Stats Influence Perceptions
Chapter 8: Strategies for Winning a Debate
Chapter 9: Programming Music Instead of Computers
Chapter 10: My Career in Media

by Alex Cosper, November 13, 2015

My Statistics class was one of the most interesting courses that directly prepared me for my radio career, which turned out to be all about numbers, ratings, percentages, rankings and projections.

I began developing a sense of stats at a young age. In third grade I was in a bowling league, in which I tracked my weekly average, highest score, total pins and other numbers, as bowling is a game of exact numbers. When I realized I was probably never going to be a champion bowler I shifted my focus on politics, dreaming that someday I could be President of the United States, so I started studying the history of presidential elections and began following political polls, which are based on estimates.

After politics seemed too corrupt of a career goal I started dreaming of being a rock star, studying the music charts, which are more about rankings, based on estimates. Eventually I learned Casey Kasem's American Top 40 radio show was merely counting down Billboard's published chart that ranked commercial products based on a mix of reported record sales and radio airplay. I went on to learn some of this information was distorted based on how much record labels bribed record stores and radio stations to distort data, until the data samples went electronic in 1991.

Settling Arguments on Greatness .. With Stats

I went on to follow the stats of football, which embodies more numbers than any other sport I can think of. Football gave me a more objective perspective on stats since I simply wasn't big enough in height, weight and strength - three huge biases that favor the highest paid pro football athletes. If pro football were a non-contact sport and more about speed and strategy, I would probably be able to compete, but I simply didn't have the physique to make the cut. The mental and mathematical aspects of the game are much more appealing to me than the rough body contact.

Football isn't just a sport for thugs knocking each other around. I began watching it in the late seventies, right when it was developing as America's favorite sport. My favorite team at the time was the Pittsburgh Steelers, because they won four Super Bowls in six years. In 1980 they won their fourth championship, which made me want to analyze why they were consistent winners.

Each game delivers enough statistics to write an encyclopedia. Even the field is marked with numbers, which makes it a unique game that's more about measurements than violence. In fact, the evolution of NFL rules since the eighties has consistently moved toward more safety to reduce injuries. Rules of a game play a big role in shaping stats. Baseball has 9 innings, which begs for a chart that lists runs for each inning. Football and basketball games are divided into four quarters, which demand charts that show how many points were scored per quarter.

Speaking of four quarters, that reminds me of how a corporation reports its earnings every quarter. Even if a company earns limited revenue all year, you can still create a table showing earnings in each quarter, making it look like a report of serious financial activity. Even if the numbers are negative, as when a shell corporation only exists to borrow money to feed other properties with revenue in a ponzi scheme, scammers can still construct a report out of it that might fool investors into thinking the firm actually has legitimate income. If data is presented in an authoritative style, inexperienced researchers are quick to believe the hype.

Football is a game of respect, and while there may be aggression involved, most of the respect comes from the numbers. Usually the players with the best stats get the most respect from the media and sports fans. Players with the worst stats tend to be the ones that get cut. Consider how the team that loses most of their games by a few points doesn't command as much respect as the team that barely wins most of their games, even though the two teams may be otherwise evenly matched.

I have always found it interesting that society is full of sports fanatics who can memorize game and player stats, but I've never known that many people who memorize political poll numbers, election stats or music chart positions the way I did, at least in my younger days. Sports are different because they allow you to verify the stats for yourself by watching games as they unfold.

For those who believe sports are rigged so that casinos can cash in, all it would hypothetically take to rig a game is bribing the head official, who has the power to nullify any given play. You can also maybe bribe a quarterback to throw a victory away with an interception at the end of a close game, but these are just hypothetical scenarios, not serious conspiracy theories to start researching for clues.

Perceptions of Popularity

Most people I've known don't question polls or charts very much. It's rare when I come across someone who questions if the number one recording in the nation is really popular or just hype. In the early eighties a lot of records that I thought were artistically weak became chart toppers, which made me suspicious. Why did all the Beatles' hits sound artistic and deserving of intense popularity a few decades earlier, whereas a bunch of bland wallpaper songs in the early eighties were beating out the records that seemed to be the most creative with the most buzz among music fans?

The Beatles are an example of a band that made it to the top due to the quality of their music. No matter how they were initially marketed, they set statistical records that no other recording artist has come close to achieving as far as most units sold, critical acclaim and social impact. As of 2015 they are still the only artist listed in the Guinness Book of World Records to have sold a billion units. There's not much mystery to their popularity once you listen to their entire catalogue. It's simply a collection of brilliant entertaining musical art that can be explained many ways.

When something bland or ordinary makes it to the top, then I begin to ask questions and see if other sources portray the product as a winner. There are other metrics to assess a band's popularity besides record sales. Do they perform shows at stadiums or small venues most of the time? That question, though, presumes that big artists prefer big money from big shows, but some of them don't. Another metric is how many hits they've had on the radio, although that presumes everything radio plays is a reflection of audience taste, which isn't always true, as new music takes several weeks to generate stats that confirm popularity.

I later learned in my radio career that there are several ways marketers can drive their records up the charts without necessarily getting any sales or phone reaction from radio airplay. Many radio station programmers play follow the leader when it comes to adding new music. In other words, they follow the national charts instead of testing new records for themselves. Whatever the charts say is "what's happening," that's what they automatically believe. Once people start seeing a video a lot on TV and hearing it a lot on the radio, they will likely perceive the track as a "hit" even if it's something they don't necessarily like.

Playing the Numbers Game

It's not just a theory that high numbers influence society. Many music fans don't have time to keep up with the thousands of releases every week, so they listen to the radio, glance at charts, read reviews or talk with friends about "what's happening." You'll get a different perspective from everyone you talk with about "what's happening" when it comes to music. The only thing all these perceptions have in common is that they represent a tiny fraction of the total music available.

Since no one has time to listen to everything that comes out, people turn to sources they trust that give them short lists of songs or artists to consider. The "top 40" is what many mainstream music fans turn to as their limited universe of musical knowledge. The people who have more time to listen to music, such as college students who don't have jobs, might impress you with a much wider list of favorites than the average 9 to 5 worker.

Since time constraints limit the amount of music people listen to, it creates scarcity of music. Last century people listened all day to radio stations, which typically shaped their playlists from a scarcity of time slots based on the most requested or highest testing results from research. This century, due to competing technology and longer work schedules, the average radio listener only listens 20 minutes per day or less and it's usually in the car. This fact has influenced radio stations to program tighter playlists of higher testing songs rotated more often, in an attempt to hang on to listeners who want to hear their favorite songs in short time frames.

Radio stations care about audience stats because that's the data they use to convince advertisers to buy advertising time. That means radio doesn't allow itself much time to experiment with untested music anymore. College or public radio, which aren't bound by commercialism, don't have to worry about ratings and can get away with as much experimentation as they want. Consequently, they end up playing less familiar music, which usually doesn't draw as much audience.

Chart placement is very important to major label artists trying to develop a story with radio. If the first record is a hit and the next few follow-ups are not hits, radio tends to shy away from adding more records by that artist. Programmers don't want to fill their limited slots with artists that "aren't happening." When it comes to new music to test, stations prefer music by familiar proven artists, unless some kind of interesting set of stats is developing with a new artist, such as selling lots of albums or concert tickets without airplay.

The fact that stats influence both the media and the masses keeps the door wide openen for marketers to manipulate data as a way to increase media exposure. Media loves to report on polls and surveys because stats help simplify stories as to "what's happening," even if the numbers are secretly based on fabrication.

Last century the record industry was able to create illusions of popularity based on how many units an album shipped to retail. There were many cases of albums being certified gold or platinum by the RIAA simply on retail shipments alone, regardless of consumer sales. The purpose of heavy shipments was to get the attention of media journalists to review such products, leading to exposure that potentially led to real sales.

The Nature of Shifting Numbers

The media loves political polls because they simplify whether a candidate or issue is winning or losing, creating a soap opera effect with audiences. Election polls are particularly interesting how they can shift rather quickly. In 1979, for example, it appeared that Ted Kennedy would take the democratic nomination away from incumbent President Jimmy Carter the following year, but by primary season Carter was able to take the lead and eventually win the nomination. He even led in general election polls until a month before losing the election to Ronald Reagan.

So why do polls change a lot? Does it mean voters keep changing their minds? In reality, polls can only be evaluated for accuracy when they are compared with election results or other tangible evidence that confirms the data. Election polls may have a wider margin of error until it gets close to the election, since there's no way of proving poll accuracy until after the votes are counted.

The direction numbers are moving sometimes tells a deeper story than actual percentages. In this month the lowest ranked candidate jumped ten percentage points while the top ranked candidate fell ten percentage points, the headline might favor the lower ranked candidate as "surging." Keep in mind candidates pay for their own polling to compare with media polls. They also spend campaign money on the media outlets that publish polls, not that there's any correlation. Just keep in mind anything's possible behind closed doors when it comes to the blending of bribery with media distortion.

The media certainly does play a role in influencing public opinion. Polls have a way of creating a bandwagon effect for some individuals, such as undecided "swing voters." Throwing around stats in commercials that 9 out of 10 doctors recommend a certain drug, can actually boost sales for the product if promoted to a large enough audience. When the public is constantly reminded through media that a song or artist is "hot," it can lead to mass acceptance of popularity as fact, even if it's all based on hype. Marketers can only hope to influence a fraction of the market, not every consumer, since too many psychological and lifestyle factors make social conditioning a complex puzzle.

Mixing Up the Margin of Error

Statistics mainly come from events, surveys and experiments. What journalists tend to downplay is that stats don't really mean anything if there the researcher, funder, sample size and methodology are not disclosed. Simply saying "55% of Americans think the economy is getting worse" without disclosing details about the source is pure persuasion rather than informative reporting. Here are notes I've taken from my Sac State Statistics class and from other sources since then:

01. You must raise a question for a specific population before you create a survey sample.
02. An effective study starts with a well defined population while a poor study reflects an ambiguous population.
03. Any selected sample must be a microcosm of the specific population to study.
04. Every survey has some type of built in bias based on its methodology.
05. Bias needs to be minimalized as much as possible to generate an accurate survey.
06. Samples should be taken as randomly as possible within parameters that reflect the population.
07. In order to limit bias, make sure that no subgroup of the population is over-representated or under-represented.
08. Phone surveys exclude people who don't regularly answer their phones.
09. Door-to-door surveys are biased against people who don't regularly answer the door.
10. Mall surveys do not capture opinions of people who avoid shopping at malls.
11. Online surveys exclude people who do not spend much time on the internet.
12. Mail surveys are biased against people who do not participate much in the postal system.
12. Large random samples (1,000+ for national surveys) are more accurate than small (200 or less) samples.
13. The way a survey question is phrased can inject bias into the equation (such as: do you agree that ... ?)
14. An average can be expressed as a mean (total divided by frequency), median (midpoint value) or mode (most occurring).
15. A frequency distribution is a table or chart that summarizes data for variables generated from surveys.
16. A bell shaped curve is a frequency distribution graph of organized data that emphasizes the mean average.
17. Experimental studies involve comparing a treatment group with a control (non-treatment) group.
18. Placebos are fake treatments give to control groups to compare with treatment groups.
19. A blind experiment is when participants in a study do not know if they are in a treatment or control group.
20. Double blind experiments are when neither participants nor researchers know who is in which group (monitored by third party).
19. In order to draw conclusions from a scientific experiment, consistent results must be repeated with different participants.
20. Correlation and causation are often mistaken by analysts who jump to conclusions about data.
21. When two variables appear to affect each other, they may be considered correlated, but that does not prove causation.
22. Only when correlation is consistently replicated should conclusions about cause and effect relationships be drawn.
23. Margin of error is often unqueestioned and may be wider than expected if parameters are not scientific enough.
24. Standard deviation (s) expresses the range of data related to the average (wide spread = high s, tight spread = low s).
25. Surveys and experiments may be influenced by the entity that funds the research.

Saved By Zero

One of the most interesting songs of my college era at Sac State was "Saved By Zero" by The Fixx in 1983. It's an ironic message since most people regard zero as having no value. In a sense the song could have double meaning, as either "saved by nothing" or "saved through elimination of all values." Judging by songwriter Cy Curnin's explanation, the intended meaning zeroes in on the latter. It's about wiping the slate clean back to the starting point, similar to the feeling of liberation from letting go of all possessions to the point of having nothing to lose.

Zero is often considered a "goose egg" instead of a "golden egg" in our society. In a football game zero is never the score a team wants as "shutouts" are always embarrassing ways to lose. But undesired numbers have a pretty wide range and aren't restricted to single digits. In a political poll, single digits are extremely low, but so is 48% in a general election. Yes, 52-48 is indeed considered a landslide in a national election.

In the game of radio ratings, a 10 share is considered awesome since there are so many stations competing for market share. But in radio research if a record doesn't test as high as 60 percent with its audience, it often gets dropped from the playlist. When it comes to a classroom test, 70 pecent is often the requirement for a passing score. In the game of capitalism, the object of a company is to grab as much market share as possible. But when market share exceeds 75 percent, the company may face complaints it's becoming a monopoly and should be broken up.

When it comes to percentages 100 seems like a pretty big number. But when it comes to raw data 10,000 becomes a very low number for attendees if a football stadium holds 60,000 seats. Yet there are still examples of much higher numbers equated with failure.

At one time if a major label album didn't sell at least 1 million units it didn't break even, so it was considered a financial flop, at least by execs and marketers. Artists frequently got dropped from their labels if their first album sold 4 million and the follow-up only sold a million. In the new century, albums that sell at that level are much more scarce, so even a half million seller is considered a victory, even if it generates a net loss.

Perhaps at some point the biggest winners in music will be the ones "saved by zero." That is, the independent minimalist artists who spend zero on recording and marketing somehow make thousands of dollars in profit per song through digital distribution, instead of falling billions in debt like the big labels.

In general, zeroes tend to be most impressive when there's a number from 1 to 9 in front of them, as in 1,000,000 or 100,000,000. Believe it or not, as high as those numbers seem for an individual bank account, they're considered nothing compared with the biggest corporation, which earn billions per year.

So if a million dollars is nothing to the top 1 percent, why can't they share it more with society? Another way of asking that same question is: why is the trickle down economic theory such a big lie? That opens up an interesting debate, which leads into the next chapter about my Argumentation class at Sac State.

Continue to Chapter 8: Strategies for Winning a Debate

© All Rights Reserved. Privacy Statement