Unleashing the Power of Content Analysis
During World War II, intelligence experts developed skills a discipline they called “Content Analysis.” In this case, Content Analysis means the ability to draw insights not necessarily directly from information itself, but rather in the frequency in which the information is presented.
Quantifying the number of mentions a particular topic receives, and connecting them to other topics, can provide analysts a look at what topics are of the greatest significance during a particular era, or within a particular context. It converts qualitative information into quantifiable data.
To understand this concept, consider the following example from Hal Gieseking:
During WW2, “German newspapers would not run articles about a shortage of food in particular cities, even if (the city) was near starvation. American analysts instead counted the number of times the word ‘food’ was used in the media. The theory: if is there were shortages, German propaganda would carry more articles proclaiming no shortages or that supplies were on their way. A sudden increase in the word ‘food’ word would probably mean that German propaganda was working overtime on a problem, a need to calm a hungry population. Very real food shortages were developing.
This type of “Content Analysis” quantifies information from any and all sources and draws conclusions about trends. The key is analyzing the number of mentions of a particular topic – not necessarily what is present in the content itself. This definition, provided in Audience Dialogue, sums it up best:
Content analysis, though it often analyses written words, is a quantitative method. The results of content analysis are numbers and percentages. After doing a content analysis, you might make a statement such as ‘27% of programs on Radio Lukole in April 2003 mentioned at least one aspect of peace-building, compared with only 3% of the programs in 2001.’
Content analysis now has interesting implications in the Big Data world. Many of you may have viewed the Google Trends site, which tracks the popularity of searches by key word, by timeline or by city. This is raw content analysis. Enter the term “big data,” and you’ll see that keyword searches began to surge at the end of 2011. Twitter’s top trending topics also reflect a raw form of quantified content analysis.
Recently, a team of researchers from the University of Bristol in the UK, led by Thomas Lansdall-Welfare conducted a content analysis across 150 years of British newspapers, spanning the time period between 1800 and 1950. This represented “millions of articles, representing 14% of all British regional outlets of the period.”
In one example cited, the researchers looked at the use of the terms “English” or “British” as national identifiers. “We observe that the terms British and English were reasonably widespread in use across most of the United Kingdom in 1854. By 1940, the use of English had dwindled, with British becoming the default national identifier,” Lansdall-Welfare and his co-authors write. In another example, the authors tracked advances in technology across the UK landscape during the 150-year time period.
Through this method, they were able to document the “transition from steam to electricity and from horses to trains. For steam, we can see that mentions during its highest use year in 1854 are widespread, with concentrations focused around major ports. However, the adoption of electricity replaces steam by 1947, with electricity being mentioned particularly in reference to London, Leeds, and areas of the Southwest. During the earliest peak of attention to horse in 1823, we see that mentions are mainly diffused across the country without a distinctive pattern, indicative of their use in rural communities, and there is only the odd mention of train, which on closer reading, was revealed to be generally in a different content (referring to animal training or processions). By 1948, the decline of horse has clearly taken effect, all but disappearing from that map, whereas train is heavily mentioned, particularly around major cities, displaying a similar pattern to that of electricity.”
The scholarly applications for content analysis are evident, and the discipline is primarily academic. Enterprises could also benefit as well. Again, content analysis as a discipline has been around for some time, but with the abundance of big data available, there are new possibilities. Not just for historical trend analysis of 150 years of British history, but also looking at current market conditions. It can play a role in mapping trends that aren’t readily explained within available information. Social media mentions are one place to start – sentiment analysis is a prime example of the ability to quantify mentions of a product or service. The abundance of text data now available to enterprises also opens up possibilities for quantifying trends shaping businesses and industries.