A set of training materials for professionals working in intervention epidemiology, public health microbiology and infection control and hospital hygiene.
Need help with your investigation or report writing? Ask the Expert. Free advice from the professional community.
You can't make decissions on this page's approval status because you have not the owner or an admin on this page's Group.
To use information for decision making epidemiologists need to organise the data collected in a standard format that allows summarising many observations. Data organisation and graphical presentation also serve the purpose of communicating results to various audiences. As the saying goes, a picture is worth a thousand words. On the other hand, this is true only if the illustration in question is thought through and constructed properly, otherwise the graphic may just hide a thousand words. So presenting epidemiological data is a rigorous task and, even if there are no fixed rules, some guiding principles can be defined.
The task may start from any type of dataset. For example a line listing of data collected by means of a questionnaire during an outbreak investigation or from a national surveillance dataset. As a first step in examining the data consider the type of variables in the dataset. The type of variables, and what you would like to show (e.g. distribution or comparison) would guide the choice of method of data display. Tables and graphs, charts or diagrams facilitate description and interpretation of distributions, trends as well as relationships in the data and complement arguments presented in the text. Maps are used to display geographical or place data. Today's powerful computer tools simplify data display and analysis to a great degree.
Join the discussion about this article in the forum!
Webster posted on 8/8/2010 9:24:14 PM:
Hey Alain and Ágnes,
Congratulations to a clear and concise delineation on how to present data.
I have only one general issue, which I like to bring to your attention in short. The web-based format offers some advantages over the classical book form, and I am unsure whether so far the medium has been suitably exploited/explored. Namely, the "introduction" quite rightly states that the way data are most appropriately displayed depends on the scale of the variable of interest. Well, imagine a new intrepid field epidemiologist (NIFE) is eager to display his freshly collected data and has appropriately started by determining the scale of his/her variables. What now?(S)he knows that the variable is measured on, say, a nominal scale, but is unsure as to how best summarise the information of the variable. (S)he needs to read the entire chapter to find out which of all the different possibilities applies to a particular scale. An alternative would be to have a page where to every variable scale (type) one would find the tables, graphs, etc. that are an appropriate display.For example:Nominal variable: Tables: Frequency table,. Bars: simple bar chart (depending on how many groups) And so on...One could also display it in a table with two columns (gridlines invisible). On the left are the scales of the variables (nominal, etc.), and on the right the different presentation formats, ideally blockwise (according to tables, graphs, etc.) If you then click of the scale of a variable, the Apropriate Presentation Formats" (APF's) are highlighted in bold, or arrows would point from the scale to the different APF's, or otherwise. By clicking on the single APFs, one would jump to the appropriate text.Even without such a (new) "decision tree", for each display-tool (eg, stacked bar chart) for each APF it should be stated for which variable type it is suitable (ideally at the beginning or end). For example, it doesn't tell you for which variable type line graphs are appropriate.The rest of my few comments are not nitty-gritty but rather picky: - The order of the Headings (links) of the chapter on the bottom is not the same as on the right hand side.- it would be good to have a button at the end of each page that jumps you back to the beginning rather than having to scroll up there.- "case-control study" should be hyphenated.Subchapter "Types of variables":- Numerical variable is introduced in plural; sometimes an "A" precedes the type of variable, sometimes not.- the point "organisation of data" is rather slim. Consider deleting it from the title of the subheading chapter and place the text (even with a line list example) before the introduction of variable types.Subchapter "Types of variables":- there is a table-heading called "two-by-two tables". Consider using the generic term "contingency tables"- dummy tables: I tend to put the column "cases" before "total", anyway.Subchapter "Other types .."- it should be explained what a box-and-whisker plot displays.
Hope this mail finds you well and you will find some of it helpful.
Agnes Hajdu replied on 8/12/2010 11:17:34 PM:
Thank you for the thorough review of the chapter and for the valuable suggestions! (and sorry for the late reply...)
Currently I am exploring possible formats for the guide you recommended - summarising appropriate displays for each type of data (with conditions that may apply). More challenging than I thought. :-) Your minor comments are also appreciated, will do the modifications.
Alain is on holidays now, he promised to reply after the 16th. I hope you are available to discuss any issues remaining.
You need to be logged in to post comments.
You can log in here. You can register here if you haven't done so yet.