Chapter 12 Mass Audiences Looking at Polls Through Different Glass Nate Silver changed the history of the election season post 2008 analyzed all the presidential polling data back to 1952 and ran 10 000 computer simulations of the election every day based on historic data and current tracking polls o Correctly predicted the winners of every U S Senate Contest and the winner of the presidential election in 49 states Chapter Insights Sophisticated methods are used to measure and assess audiences Statistics are a foundation for measuring mass audiences Mass audience size is measured by pressruns sales and surveys Audience measuring companies have gone digital to track media habits Reactions of mass audiences are part of media content decision making Audience analysis includes demographics geo demographics and psychographics Discovering Mass Audiences Research is sophisticated Media research techniques overlap into many aspects of modern life including marketing politics and governance 1 Audience Research Evolution a Joseph Pulitzer s 1880s style of research today s research with vast resources i There are data from research that can be sliced and diced into useful truths about audience ii Research is census like income levels educational attainment religious and other affiliations other research focuses on attitudes and values or habits and motivations and reactions to new things iii Used well the new data gathered largely by surveys can lead to successful decision making in television programming and advertising campaigns and be a useful factor in choosing what makes news 2 Survey Industry a Public opinion surveying is a multi billion dollar a year business i Clients include major corporations political candidates the mass media and major news organizations b George Gallup introduced probability sampling c Institute of American Public Opinion Gallup polling organization cranks out regular surveys for clients d Major companies whose work is often cited as measures of public opinion include i Nielson network television ratings ii Arbitron mostly radio audiences in local markets iii Gallup human nature and behavior specializes in management economics psychology and sociology iv Pew independent opinion research group attitudes toward press politics public policy issues v Harris best known for the Harris Poll for pioneering and engineering internet based research methods 1 Industry totals for polling companies are hard to come by because some companies are privately held and don t disclose their revenues Audience Measurement Principles Different techniques are used to measure the effectiveness of mass media messages some like public opinion polling are incredibly accurate others are less reliable and sully the reputation of serious sampling George Gallup Media People his mother in law s campaign for Secretary of State gave him an opportunity to combine his three primary intellectual interests survey research public opinion and politics and eventually come up with statistical exploration o Statistical exploration drawing conclusions from a segment of the whole Some critics say that polls influence undecided voters toward the front runner with a bandwagon effect others say it makes elected officials too responsive to the momentary whims of the electorate and therefore discourages courageous leadership Gallup was convinced that public opinion surveys help to make democracy work and fiercely defended them arguing that good surveys give voice to the inarticulate minority that legislators otherwise might not hear o 1 Probability Sampling everyone in the population being surveyed has an equal chance to be sampled a In the probability sampling method pioneered by George Gallup in the 1940s 4 factors figured into accurate surveying i Sample size ii Sample selection iii Margin of error iv Confidence Level b Sample size number of people surveyed i How do you choose a sample size Statisticians have found that 384 is a magic number for many surveys 1 384 Number of people in a properly selected sample for results to provide 95 confidence that results have less than a 5 percent margin of error for a population size of 500 000 or more 2 Population group of people being studied c Sample Selection Process for choosing individuals to be interviewed i A good sample gives every member of the population being sampled an equal chance to be interviewed divide by magic number 382 e g at Layne College 2000 divided by 332 would mean an interval of 6 2 every sixth member of the student body would need to be polled d Margin of Error Percentage that a survey may be off mark i For absolute precision every person would have to be polled in the population but that precision is hardly ever needed and would be extremely expensive and impracticable ii Pollsters must therefore decide what is an acceptable margin of error for every survey they conduct 1 Increasing the sample size will reduce the margin of error 2 Professional polling organizations that sample U S voters typically use sample sizes between 1 500 and 3 000 to increase accuracy a Also measuring subgroups within the population being sampled requires that each subgroup be represented by 384 properly selected people e Confidence Level degree of certainty that a survey is accurate i The level of confidence and margin of error are inversely related ii A larger sample can improve confidence just as it also can reduce the margin of error 2 Quota Sampling Demographics of the sample coincide with those of the whole population a Besides probability sampling pollsters survey cross sections of the whole population b Both quota sampling and probability sampling are valid if done correctly but Gallup abandoned quota sampling because he could not pinpoint public opinion more closely than within 4 percentage points on average 3 Evaluating Surveys a Sidewalk interviews cannot be expected to reflect the views of the population same goes for call in polls b Journalists run the risk of being duped when special interest groups suggest that news stories be written based on their privately conducted surveys some organizations selectively release selfserving conclusions c To guard against this careful newsrooms insist on knowing methodology details before running poll stories these are questions that reporters are instructed to ask i How many people were interviewed and how were they selected ii When was the poll taken Opinions shift over time and with elections even over night iii