In one of his autobiographic writings, Mark Twain confessed: “figures often beguile me.” He then went on to quote and misattribute (fittingly, it turns out) the now-famous sentence: “There are three kinds of lies: lies, damned lies, and statistics.”
It is sometimes close to impossible not to share Twain’s mood when reading the notes coming with the official statistics. Not because they are necessarily untrue, but because, deliberately or not, they may be misleading or beside the point.
Sometimes, we stumble on press releases that drown the reader with data of uncertain relevance. The outcome is seldom an augmented understanding of whatever issue is at stake. The press release associated with the “Business Climate Survey on Restaurants & Similar Establishments and Retail Trade for May 2020” is a good example.
Every month, a sample of businesses answer a simple question about changes in their revenue when compared to the same period in the previous year. Did it grow, or decrease, or stayed about the same? It is possibly overwhelming for the reader that in 500 words and 45 lines the document includes over 30 assorted percentage-related quotes. A careful selection of the more significant figures and trends might advantageously replace such a torrent.
Further, the first highlight of the text is the percentage of restaurants (17 percent) that reported an increase in revenue relative to May 2019. In April, the corresponding value was lower (12 percent). The press release concludes the change was justified by the “launch of electronic consumption cards.”
It is a plausible explanation. But a plausible explanation is not necessarily the main one or even a true one. As the statistical staff is no doubt aware, coincidence or simple correlation does not prove causality. Other reasons may be equally or more relevant and deserve examination. Note, for example, that the number of visitors rose in that month, and similar variations were observed in earlier periods.
Further, and still sticking with the restaurants, isn’t possibly more meaningful to realize that almost three-quarters of restaurants declare revenue drops? Or, over half of the Chinese restaurants report that income dwindled by more than 50 percent? Not to mention that in other retail activities, with no exception, all businesses report losses.
From the statistical services, we need reliable data, not hand-picked figures meant to prove that the government policies are right. To make that case is incumbent on other departments and institutions.