American Statistical Association
New York City
Metropolitan Area Chapter
2010 Statistical Poster Competition
GUIDANCE ON POSTER CONSTRUCTION
A poster must be able to stand alone, without a narrator to tell the story or a report to discuss the data. Not only must viewers understand the individual graphics, but they must also understand the relationships among the graphs, and how the graphs address the central question.
Focus - the central idea or conclusion should be the most prominent feature of the poster.
Be consistent - use graphs consistently to present the same types of data. Consistent use of colors and patterns is particularly important (e.g., data for girls is green in every graph where it appears). Use colors with restraint; colors should enhance recognition of the conclusion. If two graphs will be compared, make the size and axis labels consistent.
Each graph should have its own title, labels, and legend. Units should be correctly marked on axes and axes should be scaled appropriately.
Use an informative title. The title may convey the major conclusion to be drawn from the data or ask a question which draws the viewer in.
The poster is not a report and not an art project. The central focus should be on the graphs, and the graphs should convey the central message. Excessive artwork should be avoided. Data tables and detailed written explanations should not appear on the front of the poster.
Do not use too many graphs. If the poster contains more than five graphs, then it almost certainly has too many. Usually three to five graphs suffice - some show what data were collected and others summarize the conclusions. Each graph on the poster should convey new information about the data that cannot be seen in the other graphs. It is generally not a good idea to plot the exact same data two different ways (i.e., a pie chart and a bar chart).
Denominators - percentages are often an appropriate tool for comparing different groups of data, however, it is important to always report group sizes. For example, percentages would be misleading if a poster says that 75% of girls and 25% of boys like pokemon, and omits to report that 150 girls and 50 boys were surveyed.
Use pie-charts cautiously. It is hard to compare one pie chart to another, or to compare little slices to big slices. Often the data that is shown in pie charts is meaningless. For example, a pie chart of the number of kids that won a contest in each grade does not take into account the number of kids in each grade (see the above comment on denominators).