Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please contact us to request a format other than those available.
Misinterpretation is a common problem when using statistical information. It may be caused by a number of factors.
This section explains how statistics can be misused by
Each day, we are bombarded by numbers in the media. Many of the facts and figures quoted in the news, such as unemployment and divorce rates, originate from The Daily, Statistics Canada's first and official release of statistical data and publications produced by Statistics Canada. It presents analysis of newly released data with source information for more detailed facts. Although the information from The Daily is objective, it is interesting to analyse how newspapers from different regions can put a different spin on the same facts.
Sometimes, data are misunderstood by the media. Example 1 shows how crime patterns can be oversimplified and misinterpreted.
The Daily published data on crime statistics with the following subheadings in the release:
The Daily. Tuesday, July 18, 2000
Crime statistics
Based on the information in the release from The Daily, Canadian newspapers headlines read:
The newspaper headlines above strayed from the content of the main feature. Focusing on one aspect of the data, the newspapers ignored the main finding—that the national crime rate had declined for the eighth consecutive year. Secondly, the data revealed that the rate for violent crime fell 2.4% in 1999, the seventh consecutive decrease after 15 years of increases. All major categories of violent crime declined in 1999, including homicide (-4.7%), attempted murder (-8.8%), assault (-2.0%), sexual assault (-7.3%), and robbery (-1.5%).
The release also reported that the national homicide rate has generally been falling since the mid-1970s. That trend continued in 1999; 536 homicides were reported by police, 22 fewer than the previous year. The 1999 homicide rate—1.76 homicides for every 100,000 people—was the lowest since 1967.
Although it is possible to speculate that rising prosperity may have contributed to a lower crime rate, the data provided by Statistics Canada did not measure the relationship between the economy and crime statistics. And if it had, the observation would appear to be that the crime rate had been decreasing since 1991, the year that the Canadian recession began.
The Daily did mention that numerous factors contribute to changes in the crime rate. However, it looks as if some journalists took this statement to be a fact. The connections found in the newspaper headlines are purely speculation and were not revealed in the objective data. The journalists misinterpreted the release most likely because they misunderstood the underlying causes and effects of crime.
Statistics Canada representatives spend much time reviewing the media use of release data each day. These representatives also answer media questions regarding the data and make certain that the data are properly understood. If a misunderstanding has occurred, then the representatives try their best to correct it.
It is important to understand the statistical definitions and concepts behind the information that you are using. If you are examining labour force issues, you should become familiar with the definitions for terms such as unemployment, employment, and participation rate. If you are looking at data on environmental issues, you will need to consider the definition and concepts associated with words such as forest, woodland, extinct, endangered species, and national park.
A great advantage of statistical information is that it can be compared, allowing trends and characteristics to be revealed. For example, one can compare the weather of Vancouver with that of Halifax, past sporting results with the present, or the academic performance of men with that of women.
However, problems can arise in the comparison of statistics when the underlying definitions, classification or methods of data collection are different. This is especially true for statistics from different sources. Nowhere is this more apparent than with vital statistics. Consider Table 1 below.
Number of persons | Married | ||||
---|---|---|---|---|---|
1998 | 1999 | 2000 | 2001 | 2002 | |
Both sexes | 14,630,173 | 14,711,793 | 14,806,694 | 14,913,766 | 15,018,130 |
Males | 7,299,132 | 7,337,226 | 7,381,266 | 7,431,522 | 7,476,537 |
Females | 7,331,041 | 7,374,567 | 7,425,428 | 7,482,244 | 7,541,593 |
Source: CANSIM table 051-0010. |
According to Statistics Canada, the definition of married includes people who are legally married and living together, those who are legally married and separated, and those who are living in common-law unions. If you were to compare the numbers in the table with numbers from a survey that did not include 'common-law' in the 'married' category, the results would be very different since the definitions are not the same. Look at the data in the above table. Why do they indicate that there are more married women than men? Logically, the two numbers should be the same. What appears to be happening here is that individuals are attaching their own definitions to the term married, and this causes the numbers to be slightly different.
Factual information must have integrity, objectivity and accuracy. Yet it is important to recognize that information can be misinterpreted by personal bias, inaccurate statistics, and even by the addition of fictional data.
Consider this quotation below:
Toffler may be overly cynical in his point of view, but in reality, people and organizations do manipulate information for their own uses. For this reason, you should always be critical about the information that is provided to you. Make certain you know where the information is coming from and find out whether or not the source is credible. Also, try to find out what methodologies were used to collect and process the data.
Finally, it is important to know how accurate the statistics are because surveys are subject to two types of errors: sampling and non-sampling errors. Sampling errors occur in sample surveys because only a portion of the population is studied, not the entire population. Non-sampling errors are present in both sample surveys and censuses.