Page 23 - PATIENT REGISTRY DATA FOR RESEARCH: A Basic Practical Guide
P. 23

2.4 Prepare a Dummy Table for Statistical Analysis

                       After having obtained consent from all the relevant authorities, and also assessed the


               feasibility of establishing the registry (i.e. by ensuring a sufficiently high quality of data

               collected by an adequate sample size of the registry data), then the subsequent step is to

               prepare a dummy table which displays the overall format and structure of layout of the results


               to be presented. Dummy tables are actually empty tables containing only the variable items

               along with their statistical measures, and which will only be filled by actual data after data


               analysis has been performed (see Figure 2.1).

                       Using the analogy of building a house – constructing a dummy table for a data set is


               similar to drawing a schematic plan for the house. In other words, the dummy table shall

               display the probable final output of a data set which is designed to answer the objective(s) of

               the study. A dummy table can be designed as long as all the variable items and their units of


               measurement, along with the specific statistical analyses to be conducted on the variables are

               known (for example: to measure age in years and then to calculate its mean and its standard


               deviation).

                       Researchers will first have to decide the appropriate statistical measures to be reported


               and then to select the best format to present the results in a most informative way. An

               important way to get some idea of how to prepare a dummy table is to adopt the overall


               structure and format of dummy tables which commonly appear in previous research

               publications, which are often found in standard research reports or scientific journals. In


               addition, there are also some standard ways of designing a dummy table for presenting both

               descriptive and inferential statistics which have already been proposed by scholars (Lang &

               Secic, 1997).


                       A good dummy table serves several important functions, namely : (i) to confirm that

               the stated hypothesis/objective is doable and achievable, (ii) to provide a template for
   18   19   20   21   22   23   24   25   26   27   28