Measuring Terrorism

• Author: Brian Forst,Suat Cubukcu
• Published date: 01 February 2018
• DOI: 10.1177/1088767917737808
• Date: 01 February 2018
• Subject Matter: Articles

https://doi.org/10.1177/1088767917737808

Homicide Studies

2018, Vol. 22(1) 94 –116

© 2017 SAGE Publications

Reprints and permissions:

sagepub.com/journalsPermissions.nav

DOI: 10.1177/1088767917737808

journals.sagepub.com/home/hsx

Article

Measuring Terrorism

Suat Cubukcu1 and Brian Forst1

Abstract

This study investigates the extent of reporting and nature of biases in open-source (OS)

terrorism databases. We compare OS accounts with official accounts on terrorism

events in Turkey (1996-2012). Results indicate (a) substantial systematic discrepancy

between OS and official accounts, which we attribute primarily to underreporting

in OS accounts; (b) the discrepancy is not random—incident characteristics (victim/

target, offender, and incident types, temporal and spatial factors) and rational factors

(especially newsworthiness) matter; and (c) severity is the strongest predictor of the

probability of OS coverage.

Keywords

definition of terrorism, measurement, media reporting of terrorism, official accounts

on terrorism, open-source data collection, terrorism

Introduction

A fundamental problem confronting the measurement of both crime and terrorism is

that of accuracy: Any system for measuring violence is bound to have both systematic

and random errors. Most of what we know today about terrorism—how often, how

serious, by whom, against whom, and where and when it occurs—derives from a fun-

damentally different source than for crime: accounts by reporters. This is not necessar-

ily a bad thing, when those accounts are accurate, especially in light of limited

alternatives. But researchers and policy makers alike have tended to accept those

accounts as both valid and reliable, providing suitable bases to draw inferences about

terrorism and to develop counterterrorism strategies. It is, of course, important to

develop an empirically based body of knowledge on terrorism derived from the best

data available, both for scholarly and policy purposes, and accounts by reporters in

1American University, Washington, DC, USA

Corresponding Author:

Suat Cubukcu, Department of Justice, Law and Criminology, American University, 4400 Massachusetts

Avenue, NW, Washington, DC 20016, USA.

Email: suat@american.edu

737808HSXXXX10.1177/1088767917737808Homicide StudiesCubukcu and Forst

research-article2017

Cubukcu and Forst 95

open sources are currently the best available. But questions remain: Have we done

enough to think through the consequences of errors, both systematic and random, in

the media’s manner of collecting and reporting terrorist events, errors that we know

exist? What is the nature of these errors? How serious are they? How do they vary

systematically by circumstance? Which types of events are most and least likely to be

reported? Are the validity and reliability of open-source data constant over time? How

should our answers to these questions influence our understanding of terrorism, schol-

arly and otherwise, and our policies for responding to it?

On the Difficulty of Measuring Terrorism

Open-source databases have provided extraordinary opportunities for scholars who

are interested in describing the pervasiveness of terrorism over time and across space.

These databases are useful in finding correlates with terrorism—to understand its

nature, sources, impacts on targeted populations, and the effects of counterterrorism

interventions and policies aimed at preventing it (e.g., Abadie, 2005; Enders & Sandler,

2000; Eubank & Weinberg, 2001; Eyerman, 1998; Gessebner, Jong-A-Pin, & Mierau,

2011; LaFree, Dugan, & Miller, 2015; LaFree, Morris, & Dugan, 2010; Morris &

LaFree, 2016; Mullins & Young, 2012; Piazza, 2007, 2011; Pierskalla & Hollenbach,

2013). First, the events must be observed and reported to a data collector. Then, the

accounts should be organized in accordance with acceptable standards of data validity

and reliability. It is useful for generating descriptive counts and conducting analyses to

gain an understanding of the causes of terrorism and for the purposes of accountability

in counterterror operations.

Although open-source databases differ substantially in their data collection meth-

odologies, selection criteria, and time spans (Drakos, 2011; Schmid, 2004; Sheehan,

2012), they are subjected to selectivity bias at some level, especially, deriving from dif-

ferences in newsworthiness level of incidents and their spatial and situational attributes

(Behlendorf, Belur, & Kumar, 2016; Woolley, 2000). As with crime, terrorism is subject

to media presentations that feed the same voracious public appetite for information

about predatory violence, and to do so more effectively in qualitative terms than by way

of dry statistical exposition. Media accounts of terrorism are designed to capture audi-

ence attention, not to serve the interests of social welfare accounting. The more news-

worthy an event, the greater the chance it will reach wider audiences and draw more

public attention. Not all incidents receive same level of media attention (Behlendorf

et al., 2016; Nacos, 2003, 2003b; Norris, Kern, & Just, 2003).

The terrorist events that exhibit substantive dramatic intensity, higher number of

casualties, and greater damages tend to receive a greater level of media attention

(Chermak, Freilich, Parkin, & Lynch, 2012; Chermak & Gruenewald, 2006; Walsh,

2010). Most incidents are undercovered when they do not exceed a threshold of

emotional significance (Galtung & Ruge, 1965; Weimann & Winn, 1994). Besides

severity, the proximity and relevance of an incident to the audience make an event

more meaningful to the audience and make it more likely to find space in the news

(Golan, 2008; Grundmann, Smith, & Wright, 2000; Ruigrok & Van Atteveldt, 2007).