Amazon calculates a product’s star ratings using a machine learned model instead of a raw data average. The machine learned model takes into account factors including: the age of a review, helpfulness votes by customers and whether the reviews are from verified purchases.
This is a book about how to win an election in the age of high tech and big data. Campaigning is no longer about broadcasting a candidateÃ¢(tm)s message to the widest possible audience. It is not even about distinguishing between a candidateÃ¢(tm)s supporters, opponents, and the undecided to send a different message to each group. It is about tapping vast quantities of data available about most American adults and targeting each person with the most effective personalized strategy. And it is about finding the best messages, issues, even rumors, using experimental comparisons. It is about doing this below the radar, out of the awareness of the public. This is the nature of the modern political battlefield as seen by both parties. The strategy, tactics, and personalities who make this possible are the subject of this book.
The bookÃ¢(tm)s narrative is a collection of smaller stories. There are brief biographies of those who developed and refined new approaches to collecting and using voter data. There are success and failure stories of various campaigns and of major battles within those campaigns. And there are the specific tactics these people deploy. It is impossible to list them all, so here are a few:
- Using public records, pollsters mailed each person in a precinct a list of who had voted and who had not voted in the last electionÃ¢"along with an announcement of their plan to mail out updated lists after the next election. This increased voter turnout by 20%. When this strategy as put into practice, the mailers went only to voters likely to support the pollsterÃ¢(tm)s candidate, resulting in a selective increase in voter turnout.
- It is very difficult to sell new tactics. Ã¢aeIf you do something different, everyone will point at the thing you did different and say thatÃ¢(tm)s why you lost. So if youÃ¢(tm)re the campaign manager you donÃ¢(tm)t do anything different. If you follow the rule book strictly they canÃ¢(tm)t blame anything on you.Ã¢
- Researchers do not find sufficient similarity among Republicans, Democrats, and Independents to use these broad categories. Instead they develop more specific typologies, such as the one from Times Mirror: Ã¢aeTwo of the clusters were distinctly Republican (Enterprisers and Moralists), four Democratic (New Dealers, Sixties Democrats, the Partisan Poor, and the Passive Poor), and two leaning in each direction (Upbeats and Disaffecteds towards the Republicans, Seculars and Followers towards the Democrats). Eleven percent of American adults were found to be fully, and seemingly permanently, detached from politics; Times Mirror called them Bystanders.Ã¢ (See Pew ResearchÃ¢(tm)s web site for a more current example of such a political typology.)
- ItÃ¢(tm)s no surprise that telling people they Ã¢aeshouldÃ¢ do something produces defensiveness and resistance to change. But telling them that a large number of other people are doing it increases their chances of doing the same. This approach, developed by social psychologists to encourage general prosocial behavior, translated well to get-out-the-vote programs.
- Prospective voters asked if they would vote for an African-American often answer positively when they have privately decided they are not comfortable doing so. This has led to vote overestimates for African-American candidates. Researchers found they could get more accurate estimates by asking prospective voters if they thought their neighbors would vote for an African-American. This approach worked as Ã¢aeÃ¢Â¦a way of correcting for the inability of voters to be as honest and self-aware as pollsters like to pretend they are.Ã¢
The book provides a readable and seemingly thorough account of how campaign tactics have developed in the age of big data. I would be a more valuable book if some of the biographical information were removed in favor of more detailed description of campaign tactics and statistical procedures. As written, it gives a sense of these techniques and how they are used. More detail is needed, if not in this book, then in a companion volume that is more methods-oriented. IÃ¢(tm)d like to see Sasha Isenberg write something like what the forensic linguist John Olsson has produced, both a serious text, Forensic Linguistics and a popular audience collection of interesting cases, Wordcrime.