Predicting the Winner: The Untold Story of Election Night 1952 and the Dawn of Computer Forecasting

Chinoy, a professor of journalism at the University of Maryland, debuts with an incisive and entertaining look at the origins of computerized election forecasting. Rather than a product of the digital age, machine-generated election night projections date back to the early 1950s, Chinoy explains. He focuses in on the 1952 election cycle, when efforts to use machine technology to predict winners became a prominent feature of newly televised election returns. After the three major pollsters (Gallup, Roper, and Crossley) flunked election night 1948 with their inaccurate projections of Harry S. Truman’s defeat, election-watchers turned to new computer technologies. However, many in the cognoscenti (including CBS anchor Edward R. Murrow) doubted the accuracy of these machines, which were designed to quickly transform incoming returns into predicted outcomes. As Chinoy reveals, the networks’ race to predict the winner became a man vs. machine face-off, pitting seasoned broadcasters against UNIVAC (the Universal Automatic Computer) at CBS and the Monrobot at NBC. (The UNIVAC, housed in a warehouse, was represented by a prop console, while the desk-sized Monrobot was actually present on set.) Chinoy’s easygoing narrative sheds light on a fascinating web of characters who brought about a shift in election reporting norms, including CBS’s Walter Cronkite and UNIVAC co-inventor John Mauchly. It’s an appealing deep dive into the intersecting history of journalism, technology, and electoral politics. (May)

Correction: A previous version of this review incorrectly stated that the Monrobot was an IBM machine and that it was represented on set by a prop console.