Yesterday’s Papers and Today’s Technology

Abstract

A B S T R A C TThis article considers the methodological implications of using digital newspaper archives for analysis of media content. The discussion identifies a range of validity and reliability concerns about this increasingly prevalent mode of analysis, which have been under-appreciated to date. Although these questions do not deny a role for the use of proxy data in media analysis, they do highlight the need for caution when researchers rely on text-based, digitalized archives.

Reason Over Politics

“…I conducted a search of The Economist Historical Archive (see Gale 2017) to find out how often the word occurred in articles between 1834 and 2013. 1 Deacon (2007), with good reason, acknowledges the pitfalls of using digital databases to gather samples. It is certainly possible that, for example, relevant articles are overlooked when using databases instead of original hard copies.…”

Ample Sample? An Examination of the Representativeness of Themes Between Sampling Durations Generated From Keyword Searches for 12 Months of Immigration News From LexisNexis and Newspaper Websites

“…LexisNexis has taken a leading position among these online databases as “the media archive of choice for many academic and political sources across North America and Europe” (Deacon, 2007, p. 7). The service’s academic version offers information from more than 15,000 sources to more than 2,000 U.S. universities (LexisNexis, 2015).…”

“…Despite the widespread use of LexisNexis, previous studies have raised concerns about the methodological implications of using this database and similar services to generate a sampling frame for content analysis (see Deacon, 2007; Weaver & Bimber, 2008). Samples may not be representative of larger bodies of news accounts when researchers treat LexisNexis like a print product by creating a sampling frame from the database using the same procedures for a print product.…”

Nooit meer bladeren? Digitale krantenarchieven als bron

“…more abstract, subtler and multifaceted concepts).' 46 Zoeken naar achternamen is -OCR-fouten daargelaten -bijvoorbeeld betrekkelijk eenduidig, hoewel misschien minder eenduidig dan op het eerste gezicht lijkt, zoals bovenstaand voorbeeld heeft laten zien. Maar wanneer men op zoek is naar abstractere onderwerpen, zoals "popmuziek", "euthanasie" of "homoseksualiteit' of "journalistieke ethiek" wordt het allengs ingewikkelder.…”

“…Bovendien kan men zich met recht en reden afvragen of het überhaupt mogelijk is om 'the subtleties and complexity of meaning making' in dagbladen te onderzoeken louter met behulp van trefwoorden. 47 Gelukkig bieden databases de mogelijkheid om ook digitaal te bladeren in hele nummers. Dat biedt ten eerste de mogelijkheid om berichten te bestuderen in de context van een complete krant.…”