Introduction to line graphs

Line graphs are used to communicate the relationship of or changes in one measure such as frequency or probability of an event over a range of values in another variable such as time and dose levels. A line chart is a very common type of visual display many people come across in various media such as in the newspaper or on television (e.g. stock values line graph, trends in historical weather or the forecast). Although general awareness may not necessarily be the best measure of broad applicability of visual understanding in benefit-risk assessment, such exposure to line graphs may make them suitable for communication of benefit-risk information to most people including the general public through mass media, patients, physicians, regulators, and other experts.

A line graph is a two-way graph where each point on the line represents two values ¨C a value on the vertical axis and a value of the horizontal axis (Cartesian graph). The values on the line may correspond to the actual raw data as in a scatter plot, or may represent a ¡°best fit¡± summary data. The interpretation of the values on a line graph are very dependent on the units (e.g. probability, weight in kilograms, and weight in pounds) and scales (e.g. log scale and per 1000 patients scale), so it is also important that users understand the units and scales used on each axis since they may not always be the same and directly affect interpretation. However, users need to understand that some measures may also be unitless, for example utility (as used in many benefit-risk assessment methodologies) has more abstract interpretation.

Line graphs can be misleading when they are used to represent ranks, nominal or ordinal measures because the size of the intervals may not be equally spaced but are commonly presented as being equally spaced (see Kosslyn 2006).
Line graphs can be produced in many software packages such as Stata, R, SAS, SPSS, Tableau, Spotfire, QlikView, IBM Many Eyes, Google Drive, and Microsoft Excel?. The differences may only be in the flexibility, ease of use and aesthetic value. Line graph can also be used interactively such as allowing users to modify input parameters or through the use of 'drill-in' function to investigate underlying data.