3 Description of the Data
3.1 Food safety indices
We follow the mass-media index approach often pursued in the literature. To explore the news
content of the popular press coverage of food safety, our food safety indices are based on full-
text articles and abstracts from four US-based newspapers—Christian Science Monitor, New York
Times, Wall Street Journal, and Washington Post. These newspapers are selected because they
are available for searching during the entire sample period (1980(3) to 2005(4)). The search engine
used was LexisNexis Academic. Keywords searched were specified in Piggott and Marsh, which
were food safety or contamination or product recall or outbreak or salmonella or listeria or E. Coli
or trichinae or staphylococcus or foodborne. Based on this pool of articles, the search was narrowed
down to individual meats. The keywords beef or hamburger or meat, pork or ham or meat, and
poultry or chicken or turkey or meat were used to locate articles related to beef, pork, and poultry
food safety, respectively. Every article was read to determine its pertinence to food safety with
irrelevant ones dropped from the information base. Articles regarded as being positive about a
meat spcies were separated from negative articles. Depending on our objective judgement of the
likelihood that the article would adversely affect consumer demand for a meat species, negative
articles were assigned a score of 0, 0.25, 0.5 or 0.75, with 0 being the least likely and 0.75 the most
likely. Similarly, all positive articles were also scored. For each meat type, quarterly negative-news
indices (NIDX) were constructed by adding up scores on negative articles, and positive-news indices
(PIDX) by adding up scores on positive articles, related to this meat. Figure 1 plots these data
series for the 1980(3)-2005(4) period.
Over the entire sample, beef on average suffered a higher degree of negativity than pork or
poultry on average. The NIDX takes a mean of 11.24 for beef, 4.03 for pork, and 6.67 for poultry.
But this is not the case during the 1980-1992 subsample period. In fact, from 1980 to 1992, poultry
safety attracted more media attention than beef and pork. The NIDX values for beef, pork and
poultry in the 1980-1992 period are 2.78, 2.38 and 3.52, respectively. In the first quarter of 1993 an
outbreak of E. coli bacteria poisoning traced to hamburgers at a fast-food chain in the Northwest
received intense media coverage and caused beef NIDX to peak at 45.75. This incident ignited heavy
criticism over the soundness of the nation’s meat and poultry inspection system that brought the
pork NIDX to its maximum at 20.75 and poultry NIDX to one of its highest at 23. This event