Image from usatoday.com
The migration of the American population from rural areas to urban areas has occurred for many years. Both US census data, and the USA today chart that is derived from that data demonstrate this decline in percentage of the total population.The USA Today graphic is called a snapshot, and is one of many featured in every issue of the newspaper. These snapshots are, according to usatoday.com, "easy-to-read statistical graphics that present information on various issues and trends in a visually appealing way." The example above is clearly easy to read and visually appealing. There are only a few numbers on the graph; just enough to assure the reader that this chart is legitimate. The lines are bold and straight making them easy to follow. Even the colors are appropriate, with green representing the natural, rural farmers, and the gray representing the city and its dark streets and buildings. These colors contrast each other well, which makes it easy to see the increasing gap between the urban line and the rural line. After looking at this chart for a few minutes, which is probably longer than any USA Today reader would, I noticed that the graph was not mathematically accurate. The slopes of the two lines should not be constant, but should change to become steeper after 1960. The general idea is still expressed with the lines, but more could be gained if the graph was more accurate, particularly if there was more data.
Data from census.gov
The US census data, from the 1990 US census, shows percentages and total populations living in rural and urban areas. The full data set also tracks rural and urban population shifts in every state. This information can be found here. Urban and rural population shift is an interesting phenomenon that can be examined easily with the census data. The representation of the data in this chart form is not exactly how it appears on the website. I omitted the data for the individual states in order to make it easier to read and fit on the page better. Even with all of the extra information removed, it is not that easy to parse the data. The formatting of the chart, with a separate label for every single number on the chart makes it difficult to draw connections between different pieces of data in order to recognize trends in the data. I reformatted the data into a new chart. The major change I made was removing the single row method from before. All of the data is from the United States, so there is no need to label that it is. With the data formatted this way it is much easier to see how the data trends over time. After looking at the data in this format I noticed that the rural population was increasing even though the percentage is decreasing. The USA Today graphic only discussed the changes in the percentage of people living in rural and urban areas and stated that "a steady decline in rural populations was first noted in the 1920 census."
The formatted data makes it easy to see the deceptiveness of the graphic. Because we are given an incomplete picture USA Today is able to manipulate us to project the knowledge that they want to impart on us. I'm not sure why they would want us to believe that urban migration is so big, but it definitely is deceiving. It seems as if USA Today wants us to think that the rural population is fleeing to the cities at a rapid rate.
Confused by the snapshot, I decided to plot the data using Matlab, in order to see if there was any intentional deception from USA Today.
The new graphs shed even more light on urban migration than the charts. The graphs make it easier to understand trends than with the charts. You do lose a little using the graphs because it is difficult to know exact values just by looking at the graphs. Even though you have more accurate data with the charts pattern recognition is done by keeping track of values and trends mentally rather than visually, which can be difficult to do. In this example I think it is more important to understand the trends and patterns of the situation rather than actual data values, which leads me to believe that graphs are more useful in this situation. I do hesitate to say that, because we have seen how deceptive these charts can be when we looked at the snapshot from USA Today.
I like this analysis -- it's thorough and persuasive. I'm struck (as I have been many times before) by how easy it is to "fudge" numbers to create/shift perceptions. Makes me feel uneasy. I begin to wonder if USA TODAY's graphics are "innocently" fudging the numbers, or if there is some other agenda at work here.
ReplyDeleteBy the way: have you read any of Tufte's work on the visual representation of quantitative information? I think you'd appreciate it.
I am consistently amazed by the shameless manipulation of data into misleading graphs. Just recently I went to my local town meeting (yes, I am such a New Englander). As I came in the door, someone handed me a sheet of paper with a graph showing the comparison of compensation for my town’s schoolteachers and other towns’ teachers. The problem was, they put the baseline y-axis number at 40,000, making it look as if other towns’ teachers were paid something like 5 times as much as ours, when really it was a difference of a few thousand dollars.
ReplyDeleteCynic that I am, I expect nothing less from politicians and lobbyists. But when the press engages in this sort of mischaracterization of data, I find myself asking, what is the goal of journalism? Shouldn’t it cut through the politically-charged fog of misleading info and at least attempt to tell us the truth?
I am curious to see the accompanying article. Knowing what the article says could help us understand whether the chart is purposeful deception or simply sloppy journalism.
ReplyDeleteI suspect the latter. USA Today is known as a second rate news company which generally gives only the simplest and most cursory view of the issues. I would not be surprised if this graphic is just a continuation of the trend.
Even if it is "innocent," though, don't news organizations have some obligation to take a deeper look and make a more thorough analysis?
That's really interesting jwolf, I just wonder what the numbers would look like if you added in migration by age groups. I feel like it is generally the case that families in rural areas have a larger birth rate than urban centers meaning that they should be growing faster than the cities, but as this the city population is growing equal to that of the general population, many individuals must actually be moving from the rural areas at a higher proportion than those staying.
ReplyDeletePersonally I think it would be really interesting to see how the age dynamics have been changing. If individuals who have retired are leaving the urban areas for rural or while people in their 20s tend to leave the rural areas for a more urban environment. I think these numbers would be interesting to see and would have a lot of significance as the baby boomers enter retirement.
Very interesting post - you raise an important question that many social scientists are struggling with: do positivist methods of theorizing correctly depict the social reality?
ReplyDeleteI am not sure if USA today is manipulating us, because their data analysis, one could argue, is just as good as yours. One could further argue that actually simpler clear-cut hypothesis are better, so USA Today could have a more viable explanation.
I wouldn't criticize either USA TOday or your data analysis. Positivism itself, as a method of theorizing, needs to be questioned, because it seems that depending on how we code our data, we can always come up with our pre-established answers.
Considering the nature of positivism, I am sure that you already knew that you could prove USA Today's analysis wrong, even before you looked into their data :)
Your post is really cool. When I look at simplified graphs in NYTimes and other news sources, I always wonder about their manipulation of data. It is really easy to and manipulate graphs to make them fit the article and prove a particular point. In this example however, I agree with Gent. I don't think that USToday intentionally manipulated us. While I agree that your graphs reveal more information about the rural/urban migration; I think that USToday's graph gives the same basic idea.
ReplyDeleteI think USA Today realizes that its readers do not read it for a Pulitzer-worthy, detailed, and captivating story that will change their lives. In fact, I can guarantee that USA Today knows that. People love simple to understand generalizations (here’s one for you right now); the less you have to think, the better.
ReplyDeleteJust to add to your analysis of the snapshot: it is interesting how urban and rural populations are represented. The rural is personified by a white male farmer, who is ready for some good old-fashioned work, as shown by the shovel. The urban on the other hand, is represented by a dark-skinned female businesswoman. Is this how Americans view the urban or rural, or is this how USA Today wants us to imagine it?
I don't think this is evidence of media manipulation--I can't really think of any ulterior motives that would cause them to deliberately distort this kind of data.
ReplyDeleteNevertheless, I think the choice to simplify the data was intentional. Given the audience for USA Today, the goal is to graphically portray the trend in as simple a format as they can. Ultimately, this is illustrative of the importance of audience in crafting any story.
I'm wondering if the USA Today graph is associated with any articles it reported. What's the purpose for the newspaper to plot the graph in that way?
ReplyDeleteI know that numbers can lie but it's my first time to see a real case. Good job in proving your point. Normally we assume that graph is an honest reflection of its data. After reading your post, I just think when we draw a graph, we have added so much information to the original numbers, which could be very misleading sometimes.
I do not think that averaging the data= manipulating it. If they only used 3 points to make a straight line trend, it will represent the average. I think that most people will not assume that the migration rate was constant for the past 80 years- an absurd assumption.
ReplyDeleteHowever, the data you provided for 1960 is different from the USA Today Data by 7%. Did they publish incorrect data or just used a different source/measurement technique? Maybe they accounted for population of certain age, etc.