Presenting Data on a Graph

¶ … Misleading Graph Statistics are often represented in a graphical format to increase the ease and speed of comprehension. A graphical image, such as a graph, is able to convey a significant amount of data in an effective manner and is often much easier to read compared with a paragraph of text that presents the same data.

The benefit of the visual representations is their ability to show patterns or trends in data, for example changes over time, or differences between different categories, allowing for comparison and contextual assimilation of the data. Graphs may frequently be used with an accurate presentation of data, but it is also possible for graphs to present data that is technically accurate in a manner that may still be misleading, especially if readers assume that good practices are being applied.

A good example of this is the data that was presented a graph depicting the increase in the number of welfare claimants between 2009 and 2011. The graph was produced by the Senate Budget Committee Republican staff and reproduced by Daniel Halper in a blog on the Weekly Standard website (Halper, 2012). The graph is shown below in figure 1. Figure 1; Graph showing number of welfare claimants in the U.S. 2009-2011 The graph is titled "Over 100 Million People in the U.S.

Now receiving Some Form of Federal Welfare." The question is whether or not this is presenting a fair picture of the data. The title is accurate the graph does appear to who that there has been an increase from under 100 million to more than 100 million. The graph is also labeled with the way that the data has been calculated, counting only people who are receiving means tested welfare, specifying that it does not include social security or Medicare.

For a graph to be meaningful it should be ladled accurately and it should also clearly define the data that it is presenting; if the data itself is not clearly defined than the contents have the potential to be misleading (Smith, 2014). If data is inaccurate there is also the potential for representation. However, in this instance the source of the data is given as the U.S. Census survey of income and program participation, which is a credible source, unlikely to have provided accurate information.

Therefore, in terms of fairness, it appears that the information provided on the graph itself is likely to be accurate. However, this does not mean it is not misleading. The reason for using a graphical representation is to create a fast and easy understanding of the data, making it easier to read than the equivalent text. There are several assumptions associated with the way graphs are presented.

Halper's blog article, which is using the graph prepared by the Senate Budget Committee Republican staff, is not complying with those expectations, and as such the representation of the data may be seen as potentially misleading, due to the way in which the readers may interpret it. There are several ways in which graphs may be misleading. One of the most common practices when misrepresenting data is to alter one or both of the axes to minimize or exaggerate differences between the different bars or data points.

In figure 1, the y-axis does not start from zero, which is the common practice (Smith, 2014). Instead it starts from 94-000,000 and then goes up in 2 million increments. It may also be argued that because the data represented in the first bar codes partway up the graph, then maybe the interpretation that this was a true starting point, rather than just starting point.

The effect of this reduced Y axis is to lessen the total scale that access, a strategy which emphasizes the difference between the bars representing the data for each period represented along the x-axis. The impact of this is to increase the gradient of the trend, so it looks like there has been a significant increase.

Where there is a graph that doe not begin at 0, or the scale is not equal along the axis most statistics book will advocate putting a break in the line so show the axis is not continuous. This impact of this may be that the data will be misinterpreted by the reader to indicate a much sharper increase; an interpretation that may be of some political or other benefit to the creators of the graph.

However, just because the graph may be misinterpreted duties presentation, does not necessarily mean there's been a purposeful misrepresentation; it may simply be poor grass growing. However, in this case it is certainly clear, that had a full y-axis been presented, that the difference between the figures at the beginning of the x-axis for 2009 Q1, and the end of the x-axis for 2011 Q2, would be visible by a much shallower gradient across the bars. Another consideration, which may support the potential.