Statistical data that is in numerical form can be represented in various graphical forms; as charts or graphs. The choice of the right method of representing statistical data graphically depends on the type of the data and data characteristics like variation and tendency. Graphical representations are appropriate for statistical data sets that need to be summarized into a single measure or unit of data analysis. Representing such data requires the use of graphical methods that fit or highlight the preferred measure of data tendency. Such measures of tendency in statistical data sets include mode, median, or arithmetic average which considers every score in a given set of statistical data. If the graphical representation requires a more accurate picture from the data set, then the graphical method chosen must factor in the variation characteristics of the given data set. These measures of data variation in statistical data include range, variance and the standard deviation (McCluskey et al, pg. 127-130.
Graphical methods for representing statistical data include bar charts, line graphs, pie charts, scatter plots, histograms, cumulative frequency pie charts, and polygons among several others.
In the statistical analyses and graphical illustrations contained in this article, the statistical data set was obtained from an experiment on the type of detergent making the largest bubbles. The data set is as follows: Type A: 44.0 cm, 38.9 cm, 30.8 cm, 29.4 cm Type B: 25.6 cm, 30.2 cm, 23.3 cm, 20.1 cm Type C: 10.0 cm, 15.4 cm, 21.6 cm, 12.9 cm. (N.P.Q.A.L, pg. 2)
Bar charts/ Bar graphs
Bar graphs are ideal for representing categorical data. They are also used in representing ungrouped data with discrete frequency observations (Bachi, pg. 74-90).
- Bar charts represent data in each category for a frequency distribution.
- They are useful in presenting as summary of large data sets in a visual form
- They are easy for estimation of key values at a glance
- They must be accompanied with additional explanations for the data
- They are easy to manipulate to give a false impression
- They involve a lot of assumption of the patterns and effects or causes of statistical data
They represent data as a percentage
- The summarize data in a visual form
- It possible to display relative proportions of multiple data
- Easy to understand hence mostly used in business
- It is not possible to reveal the exact values of the data
- They fail to display data continuity, several pie charts are required to display data change over time
- They are easy to manipulate with intent of showing false impression
Scatter graphs are important when displaying the relationship between two variables is necessary. They are useful in correlation when displaying large data sets (Bachi, pg. 74-90)
- They are Easy to show data correlations
- They represent the minimums and maximums of the data set effectively
- Difficult to analyze the data as it is not represented in any linear form
- It difficult to identify positive and negative correlation
In conclusion, the choice of an ideal graphical representation of statistical data is significantly influenced by the type of the data to be analyzed or represented. However, bar charts are the most used graphical representation for showing comparisons within the data set. In addition, they are more presentable and appealing to the eye across a wide range application.
McCluskey, Anthony, and Abdul Ghaaliq Lalkhen. “Statistics II: Central tendency and spread of data.” Continuing Education in Anaesthesia, Critical Care & Pain 7.4 (2007): 127-130.
Bachi, Roberto. “Graphical methods for presenting statistical data: progress and problems.” AUTO-CARTO II (1975): 74-90.
Nagle Garret & Witherick Michael. “skills and techniques for geography” (2002)