This study identifies the most efficient methodology for sampling from a population of New York Times articles related to terrorism, which were generated through keyword searching. Efficiency was based on which sample statistic was closest to the population parameters of interest. The smallest sample size, where 68 percent of the sample statistics were within one standard deviation of the population mean and 95 percent of the sample statistics were within two standard deviations of the population mean, were identified as the most efficient. In addition, we determine whether the frequency of news articles is correlated to the temporal distribution of terrorist incidents found in the Global Terrorism Database, which could possibly be utilized to more efficiently sample from the population. Our findings confirm prior research that shows that sampling efficiency is related to the weekly news cycle and, contrary to prior research, the sample must include between 20 to 29 constructed weeks to achieve representativeness of an entire year of coverage for a population generated through keyword searches. In addition, the study also found that there was a limited relationship between the frequency of terrorist incidents and the amount of terrorism coverage in the news.
Parkin, William S. and David A. Green. 2016. "Terrorism in the News: The Efficiency and Impact of Sampling Methods on Data Collection and Content Analysis." Studies in Conflict & Terrorism 39 (April): 668-686. http://www.tandfonline.com/doi/abs/10.1080/1057610X.2016.1141019