In this blog post, we’ll examine various examples of how statistical data and graphs can be distorted, and explore how to read statistics critically.
Headlines such as “60% of Gold Transactions Take Place on the Black Market… Scale of Tax Evasion Unfathomable,” “Thyroid Cancer Increases by 24% Annually… Controversy Over Overdiagnosis,” and “22 Years Ago, 75% of the Population Belonged to the Middle Class… Where Do You Stand Now?” are all phrases actually used in real news reports. We live in a flood of numbers and statistical data. New statistics are constantly being generated, and they are processed and presented in various forms to make them easily understandable to readers. While these numbers are highly persuasive, they are not always objective or accurate in and of themselves. Misused statistics can cloud readers’ judgment and lead them to accept false information as fact. Let’s now examine two examples of how statistical data has been misused.
Looking at the proportion of the government budget allocated to social security spending, as of 2005, welfare states in continental Europe recorded 42%, and those in the Anglo-American bloc recorded 29%, while South Korea’s figure stood at just 12%. Furthermore, while past governments were unable to allocate sufficient funds to social welfare due to their focus on economic development, the report explains that this ratio has been set at 28% in the following year’s budget and is expected to reach Anglo-American levels in the near future.
The most questionable aspect of this data is the claim that the share of the welfare budget within total government spending increased significantly from 12% to 28% in just a few years. While the development of the national welfare system is a positive trend, given that the national budget is formulated through a careful process, the assertion that the share of the welfare budget more than doubled in just five years is difficult to accept. Such figures could give the impression that government policy has changed very abruptly.
The author’s statement that the ratio of social security expenditures to the 2005 government budget was 12% appears to cite the social security expenditure ratio calculated by the Ministry of Planning and Budget at the time, based on the general budget. It is true that, under the general account standards at that time, the social development sector accounted for approximately 12% of the total budget. On the other hand, the figure stating that 28% was allocated in the 2010 budget is cited from the Ministry of Strategy and Finance’s National Fiscal Management Plan. In that plan, the share of the health, welfare, and labor sectors is presented as approximately 28%, and the same document indicates that the welfare sector’s share of the budget in 2005 was also approximately 25%.
In fact, a review of the National Fiscal Management Plan reveals that the share of the welfare sector in the government budget has shown a steady upward trend over the long term. However, the increase—from approximately 25% in 2005 to about 28% in 2010—does not represent the dramatic change claimed by the author. This discrepancy arises from a direct comparison of two sets of data compiled using different criteria. Comparing statistics with different sources and calculation standards as-is can lead to erroneous conclusions and raises questions about whether the ratios for European welfare states and Anglo-American countries were calculated using the same standards.
The following examples illustrate the distortion of statistics through the use of graphs. All three graphs were used to effectively convey specific arguments. The first is data presented to explain the safety of U.S. beef; the second is a graph showing changes in the approval ratings of the two presidential candidates during the election; and the third is data used to illustrate the growth rate of the service industry.
The first graph explains that most beef from cattle aged 30 months or older is consumed domestically in the U.S. as ground beef, while safe beef is primarily exported overseas. However, this graph merely presents that fact without providing additional information on how cooking beef and processed meat are distinguished, or what types of beef are actually exported. Consequently, readers may end up with even more questions, and the graph may cause confusion rather than resolving their curiosity.
The second graph shows the difference in approval ratings between the two candidates during the presidential election. In actual figures, the gap between the two candidates narrowed significantly from 6.8 percentage points to 5.1 percentage points and then to 2.2 percentage points. This trend suggests the possibility of a reversal in the next poll. However, in the graph, the bar lengths for these three figures are depicted as nearly identical, failing to adequately reflect the actual magnitude of the change. Viewers of the graph may get the impression that the gap in support has not narrowed significantly. Since opinion polls can influence voters’ judgments, the graph must accurately reflect the actual figures.
The third graph emphasizes that the growth rate of the service industry has fallen significantly compared to the past. While it is true that the growth rate since 1999 has been considerably lower than that of the 1970s and 1980s, the graph portrays this difference as much larger than it actually is. Furthermore, this data fails to take into account the process by which the South Korean economy, after experiencing rapid growth in the 1970s and 1980s, shifted to a more stable growth rate following the foreign exchange crisis. Since changes in the service sector’s growth rate are closely linked to changes in the overall economic growth rate, it is difficult to conclude solely based on this graph that the service sector is losing its competitiveness.
As seen in the examples discussed earlier, statistical figures and graphs can serve as powerful tools to strongly support specific arguments. Because numbers appear to possess objectivity and reliability, readers are easily persuaded. However, even objective data can lead to a misunderstanding of the facts if the statistics are compiled incorrectly, data based on different criteria are compared, or the graph is intentionally distorted. Therefore, rather than unquestioningly accepting statistical data as fact, readers should exercise critical judgment by examining the source of the data, the criteria used for calculation, and the method of presentation. Furthermore, when using statistics to persuade others or support one’s own arguments, one must carefully verify that the data effectively conveys one’s opinion without distorting the facts.
