Review of “Less of the Same: The Lack of Local News on the Internet” This study, based on comScore panel data, examined the consumption of local news sites in 100 local markets. The analysis revealed that most local news sites (that reach at least 1 percent of the local audience) are affiliated with traditional media outlets and very few are Web-only operations. Additionally, traffic to these local news sites accounts for only about half a percent of total monthly page views and minutes in these markets. Market concentration (measured by HHI) is high. These findings suggest that the “audiences are small—and proportionally much smaller than even many publishers and journalists seem to realize.” The researcher concludes that online news has a readership problem, which leads to its revenue problem. This is indeed a thorough, comprehensive study (perhaps not the first one though) on the state of local news on the Internet. Data were obtained from a major online research firm (comScore), the analysis was systematically carried out, and the conclusions corrected some strongly and wrongly held assumptions that have blocked a reality-based view on the economics of online news for quite some time. The researcher should be commended for such a solid and timely study that generates much-needed findings at a time when the state of news media is receiving unprecedented attention. The researcher did a great job addressing key issues and problems associated with online audience measurement in general and comScore’s panel approach in specific. Still, some of the limitations inherent in this methodology should be reiterated as they may influence how certain results are interpreted. First, as stated by the researcher on p. 9, most Internet research firms do not fully disclose details about their recruiting efforts and thus the representativeness of their panel cannot be evaluated independently. The researcher then reasoned that because avid users are more likely to be recruited, “audience reach numbers would be biased upwards rather than downward.” This may not always be the case. Other factors such as the size of the home sample over work sample (noted on p. 10) may result in an underestimation of news site traffic because online news usage usually peaks during work hours. Using the comScore data as is, the researcher should recognize such problems without suggesting that such data represent “a favorable portrait of online news audiences.” Second, a local news site must reach 1 percent of the local audience to be included in the analysis. As noted by the researcher throughout the study and illustrated by the “census” (actually follow-up examinations) of news sites in five specific local markets, the 1 percent threshold eliminated a number of small news sites, while the study aims at examining local news sites. Given the subject of this study, this is probably not the best decision. Why 1 percent as opposed to 0.5 or 0.3 percent? In addition, market share (in percentages) is a relative measure. A local (especially hyper-local) news site in a large market such as New York may effectively serve its community but fall below the 1 percent threshold. Furthermore, what if aggregate usage of many such hyper-local sites constitute substantial market share as illustrated by a long-tail distribution? The elimination of local news sites due to the 1 percent threshold leaves room for speculation, especially when coupled with the potential representativeness problem of the panel approach. Having said that, I don’t think there is a perfect solution for this problem. One remedy would be to also examine data collected by Nielsen/Netratings or Hitwise, compare notes, and check if the overall picture -- that local sites account for a relatively small market share -- remains the same. (I believe it does). The regression analysis does not add much information to the findings. It is not clear why some predictors are considered relevant in the first place and why some interaction effects are examined but not others. Without reporting the p value, what is the significance level used in Tables 3 and 4? It is also necessary to specify what coefficients are reported in these two tables. It is not clear that the “per person” measures (those in Table 1) are per resident or per Internet user. It is more reasonable to use the latter because not everyone is online. Some minor issues: The title of the study (“the lack of local news”) is misleading. This study examines the level of consumption, not the amount of content. On p.1., this citation in the opening paragraph -- “recent Pew studies have found that more Americans now get their news from the Web than get it from print newspapers (Purcell et al. 2010)” -- may not be totally accurate. (Direct quotes from the source: “61% say they get some kind of news online; 50% say they read news in a local newspaper; 17% say they read news in a national newspaper). In addition, according to the researcher’s reasoning, the number of online news visitors is not a good measure of online news consumption. Typos: p. 11, line 2 of the last complete paragraph, delete “that breaks.” p. 12, line 12 of the first paragraph, “our” should be “are.” p. 17, The Tucson Sentinel (“Tuscon”) -- should be “Tucson.” H. Iris Chyi, Ph.D. Assistant Professor Phone: (512) 471-0553 chyi@mail.utexas.edu http://newmediaresearch.org