Understanding and Using Descriptive Statistics

Abstract

There is sometimes confusion over the correct statistical techniques that should be used in a particular setting. This paper outlines some methods that can be applied both to display information graphically and to summarise data numerically. The relative strengths and weaknesses of a number of diagrams and summary measures are examined. The particular method(s) that is used depends upon a number of factors. These are, firstly, the level of measurement of the data that have been collected; secondly, the overall distribution of the data being investigated; and, finally, the overall strategy of the plan of investigation. The reasons as to why a particular technique should be used are not always apparent and some methods for making a decision are discussed. Investigators should consider their choice carefully and be in a position to justify that choice.

Effective Reporting of Quantitative Data

“…These encompass measures such as frequencies, percentages, measures of central tendency (for example, mean, median and mode) and measures of spread (for example, standard deviation, variance, range and interquartile range). Gissane (1998) provided more detail on how these are calculated and used. Although this form of quantitative reporting seems straightforward, there are hidden dangers.…”

“…A pie chart or stacked bar chart can be used for all types of data where it is intended to demonstrate proportions of a total. A number of examples of these can be seen in Gissane (1998). In addition, Denscombe (2003) provided a checklist (Table 2) to assist in ensuring that all the necessary considerations are made when presenting quantitative data.…”

“…Many useful resources already exist for this, including a number published in the British Journal of Occupational Therapy (Kirkman 1992, Kelly 1994, Drummond 1994, Hall 1994, Hayes 1996, Huff 1999, Thomas 2000. This paper complements that of Gissane (1998), which introduced descriptive statistics and included some examples of how to present these. A forthcoming paper in this research series, on inferential statistics, will provide a greater depth on what these are and on important considerations surrounding their use.…”

“…Fig. 1 is an additional example of a visual representation of results to those presented by Gissane (1998). It is a scatter plot from a study that investigated the test-retest reliability and discriminate validity of the COPM in stroke patients (Cup et al 2003).…”

How to present your data I: Graphs

“…They are effective because readers can compare categories, assessing the differences between them [12]. The length of each bar represents either the count in a specific category, or the percentage of the total number of observations [13,14]. An example of a bar chart is shown in Fig.…”