HOW TO WRITE A SCIENTIFIC ARTICLE : PART IV: PRESENTATION OF DATA
Rohini M. Belsare, M.A., M.Ed 1*, Devangini R. Broker, MD (Hom)2
1 Advisor, Publications, Dr. M. L. Dhawale Memorial Trust
2 Assistant Professor, Department of Repertory & Senior Research Officer, Department of Research and Medical Informatics, Dr. M. L. Dhawale Memorial Homoeopathic Institute (MLDMHI), Palghar
*Address of correspondence: Mrs Rohini Belsare
How to cite this article:
Belsare RM, Broker DR. How to write a Scientific Article: Part IV: Presentation of Data. Journal of Integrated Standardized Homoeopathy (JISH) 2019; 02(02)
Received on: June 10, 2019
Accepted for Publication: June 12, 2019
This article provides introductory guidelines for data presentation, which is a necessary part of any scientific writing. It shows that the mode of presentation depends on the types and measurement of data. A graph is an effective tool for data visualization and creating impact, but is often misused due to wrong selection of graphs. Choosing appropriate graphs is important to promote accurate and effective communication of results in the scientific world. Additionally, this paper provides a summary flow chart to aid a quick selection of the appropriate graph as per the guidelines.
Keywords: Graphical presentation, Data Presentation
Collecting raw data per the objectives is a crucial activity in any research. Raw data must be rendered meaningful through summarizing, processing, and analysing or synthesizing.1 Sir John Richard Hicks, a British economist, explained the importance of accurate data presentation by saying, “classified and arranged facts speak for themselves; narrated, they are as dead as mutton.”
How do we present data collected so carefully? It can be presented as a text, a table, or a graph. The format will be determined by the purpose and objectives of the research and the nature of the data collected.
The data presentation should:2
- be concise without losing the details
- arouse the reader’s interest
- be simple and meaningful, allowing the researcher to form conclusions and derive the discussion themes
- act as a link between the objectives and the conclusions
The above guidelines, though simple to understand, are complex to apply while formulating plans to present data. A few studies on the quality of data presentation in medical speciality journals have indicated many deficiencies.3,4 This may be due to a variety of reasons like a lack of graphic formats in author guidelines, less attention paid by the evaluator who fails to provide critical feedback on data interpretation and less attention paid by the author as well.5 There is plenty of theoretical literature on data presentation; unfortunately, data on the appropriate use of tools of graphical presentation are scattered.
This article focuses on the concepts and tools of data presentation, which will help the authors and researchers to render data meaningful to the scientific world as well as to lay persons. It emphasizes the concepts of graphical presentation and the utility of each variety of graphs. The article also equips the reader to read the graphs, which helps in the critical evaluation of scientific writing.
Steps of Data Presentation:1
TYPES OF DATA: QUALITATIVE & QUANTITATIVE
Broadly, data is classified into Qualitative and Quantitative. As the name suggests, Qualitative data signifies quality and is discrete in nature, whereas Quantitative data has magnitude and is continuous in nature.2,10 To simplify further with an example, different remedies used in the management of rheumatoid arthritis and its progress, the proportion of remedies used represents quality and hence is Qualitative data; whereas, the C-reactive protein level measured every 3 months to understand the effect of Homoeopathic management represents Quantitative data as it has magnitude and is measured. Qualitative data is measured in Nominal (Names – Different remedy names, outcome of the disease into cured – not cured, etc) or Ordinal (severity of rheumatoid arthritis as classified into mild, moderate, and severe). On the other hand, Quantitative data is measured in either scale or interval data. Hence, representation of Quantitative data in graphs will be in a continuous series, whereas in Qualitative readings, the units will be in discrete series, each separated by some distance. Therefore, one criterion for selecting the method of presentation will be whether the data is Qualitative or Quantitative.
Data can be presented in ONE of three ways:1,6
- As text
- In tabular format, or
- In graphical format
In textual presentation, the numerical data is presented in a descriptive format i.e. paragraphs or sentences.1, However, if the data is in a large quantity, it can be more usefully presented in a graph or a table, as descriptive text takes longer to read.1 Hence, a text format is useful when we are giving small information in 2 – 3 sentences.
In tabular presentation, numerical data has been summarized for either statistical processing or presentation. Tabulation involves the systemic presentation of data in rows and columns to facilitate comparison between the figures.6 Usually, the dependent variables are presented in columns whereas independent variables in rows.
One needs to follow the guidelines for representation of Qualitative and Quantitative data2,6.
The proverb “one picture is worth a thousand words” is apt to explain the purpose of the graphical presentation which is to demonstrate the relationship between two variables in a more appealing way for easy understanding of complex information. Large data can be presented very simply through graphs.1 The basic components of a graph are graph no., title of graph, axis names, and legends. Now-a-days, Microsoft Excel software is used widely for formulating graphs. We can have 2-D or 3-D graphs for greater impact, depending on the need of data visualization.
The selection of appropriate graphs, like that of tables, depends on the type of data2 and measurement of data.10 In this article, we explain the frequently used graphs in medical articles and posters.
The Graphical Presentation of Qualitative Data:
- Bar Graph:
The bar graph, used to present Qualitative – nominal data, demonstrates the comparison between the different groups at a glance2,10. Here, we take the example of a simple bar chart (Fig. 1) where the horizontal bar graph is used to represent the commonly prescribed organ remedies among Type 2 DM. The vertical bar graph has limited space for mentioning the text at the x axis6. The horizontal bar graph overcomes the limitation of the vertical bar graph. Here, in Fig. 1, horizontal bars are used since the remedy names are long.
The multiple bar chart (Fig. 2) contains two or more bars that represent two or more series of values at a time.2,6 In Figure 2, for example, the response to treatment across the groups as per the changes is presented in multiple bars to focus on the comparison between and within the groups.
Component or stacked bar graph (Fig. 3) where each bar is sub-divided into components or bars. It represents a situation where there are multiple nominal variables in one category. These graphs are not commonly used, as too many variables may confuse the reader, but a correct understanding can be achieved through a careful reading.
2. Pie chart or sector diagram:
A pie or circle chart is a circular chart divided into sectors to present data into relative frequencies or percentages. Its purpose is to focus on comparing the size of a specific category (a slice of the pie) with the entire pie.6 For comparing each category, the bar graph is the best. The perspective 3D pie chart, polar area diagram and the doughnut charts are an extension of the pie or sector diagram. Pie charts are useful if there are limited, i.e. less than 6 categories and their proportions are distinctly different. Otherwise, each category’s slice will look the same and will not create the desired impact. Due to these limitations, statisticians avoid using pie charts. Figure 4 illustrates a pie chart. The visual difference in each slice, however, is insufficient, making it difficult to appreciate the comparison through comparing the dimension of each slice.
Graphical Presentation of Quantitative Data:
The histogram is most commonly used to represent Quantitative – interval grouped or ungrouped data2,6,10. In the case of grouped data, the class interval must be the same. A histogram conveys the comparison between each class at a glance at any given point of time. In addition, the shape of distribution in the histogram conveys the approximation of probability distribution i.e. normal distribution or skewed distribution patterns. Another use of this graph is to highlight extreme observations or any gaps in the data.6 There are many patterns of the histogram.
The similarity between the vertical bar chart and histogram is “vertical rectangular bars”. Due to this reason, they are often interchanged and falsely used to present data. We can see that in Figure 5, there are no gaps between the bars of the histogram representing the variables. They are in a continuum and hence there is no space between the class boundaries whereas in Figures 1, 2, and 3 one can see the gaps between the two bars as the data is discrete. Due to this, a histogram is always used to present Quantitative (continuous) data whereas bars are used to present Qualitative (discrete) data. Moreover, Figure 5 presents the age distribution of geriatric patients visiting the Homoeopathic OPD. It has become a skewed distribution as the maximum numbers of patients are between the ages of 60 to 65 years, the skewed and extreme values perhaps being due to mortality and morbidity.
2. Frequency polygon:
This is most commonly used for presenting Quantitative ratio data. It has a polygon (a figure with many angles).2,6 It is an area diagram of frequency distribution developed by joining the mid–points of class interval at the height of frequencies by straight lines.2,6,10
Figure 6 illustrates the prevalence of scabies in the school over a period of 3 years after Homoeopathic treatment representing the trend of the rate. Here at a glance, we can see the progress made in the disease condition. It is not useful if the class intervals are unequal or open-ended, a histogram being more useful in such a case. If the data is large with small class intervals, the line takes the shape of a curve and this graph is known as frequency curve, which helps to understand the probability distribution pattern2,6,10. When the cumulative frequency is presented against the class intervals and successive points are joined in line, the “Ogive” chart is constructed, which is useful to demonstrate the number of proportions of person above or below a given value2,6 .
3. Line graph:
It is a graphical presentation of Quantitative data – both interval and ratio data – which is useful to present the trends over a period of time2,6,10. It is frequently used in statistics and health sciences as it is visually appealing and easy to understand. However, it is to be used cautiously as inconsistent data may give a distorted image of the trend and create a wrong impression. The line can be single or multiple for the purposes of comparison.
Figure 7 illustrates the effect on overall expressions of Homoeopathic treatment, comparing it with the placebo group. It is evident that if the similimum is correct, the improvement occurs within the first 3 months as compared to placebo. Once improvement is initiated, both the groups have followed a gradual pace towards improvement. The difference between the two groups has persisted for the whole year.
4. Area chart:
It is the graphical presentation of Quantitative data – both interval and ratio – which is useful to present the magnitude of a specific sub-category in relation to other sub-categories10. For example, Figure 8 shows the magnitude of geriatric patients visiting the Homoeopathic OPD for complaints related to the musculoskeletal system (MSS) and respiratory system (RS). The diagram depicts the magnitude or higher proportion of geriatric patients suffering from MSS as compared to RS complaints as well as the trend, which is gradually decreasing with age.
5. The scatter – plot diagram or correlation diagram:
A scatter plot is used to show the relationship between two variables, interval & ratio, and ordinal measurement of the data10. One cannot formulate the scatter plot using categorical nominal variables. It is the most accurate way to display correlations as illustrated in the example below Figure 9. It is extremely effective to visually focus on how one variable changes in relation to a change in the other variable(s). Here in Figure 9 a positive correlation is seen as, the innovative teaching learning method score is increasing and simultaneously the perception of the application this knowledge in future is increasing. This is evident as the observations are clustered on the right side of the graph. Like in the line diagram, if the data is inconsistent, a distorted image of the trend may be obtained.
6. Whisker box plot diagram:
This type of graph is being increasingly used to show the shape of the distribution, its central value, and its variability. It is used to present non-parametric distribution patterns. Hence the interval, ordinal measurements can be presented with the box plot though only a few authors use it for the interval or ratio measurements.6 Compared to the histogram, they occupy less space and help to compare the two-distribution pattern.
Figure 11: Flow chart representing the selection of appropriate graphical presentation:
The first step is to classify the data in Qualitative and Quantitative type. The second step is to understand the type of measurement used for that data. The last step is then to select appropriate graphs that will represent it in a best way.
Data can be presented in any one of these forms – textual, tabular, or graphic. Depending on the purpose and objectives, one can choose any of the formats to present the data. The types of data, measurement of data, and the purpose of data presentation are the basic foundations to choose the appropriate graphs.
- Junyong I, Sangseok L. Statistical Data Presentation, Korean journal of anesthesiology 2017; 70(3): 267 – 276.
- Mahajan BK, Methods in Biostatistics for Medical Students and Research workers. 7th Edition, New Delhi: Jaypee Brothers Medical Publishers (P) Ltd., 2010: 12
- Cooper R.J., Schriger D.L., Tashman D.A. An evaluation of the graphical literacy of the Annals of Emergency Medicine. Annals of emergency medicine 2001; 37: 13 – 19
- De Amici A., Klersy C., Tinelli C. Graphic data representation in anaesthesiological journals: a proposed methodology for assessment of appropriateness. Anaesthesia and intensive care 1997; 25: 659 –664.
- Cooper R.J., Schriger D.L., Close R.J. Graphical Literacy: The Quality of Graphs in a Large – Circulation journal., Annals of emergency medicine 2002; 40: 317 – 322.
- Kartikeyan S., Chaturvedi R.M., Bhosale R.M., Comprehensive Textbook of Biostatistics & Research Methodology. Chapter 4: Classification and Tabulation of Data, India: Bhalani Publication House., 2016: 75.
- Patel M.K., Broker D.R., Tamboli P.P. Using multimedia for enhancing the observational skills of homoeopathic post-graduate students during case taking. Journal of Integrated Standardized Homoeopathy (JISH) 2018;01(01):1-11.
- Shah P.M., Pethani V.R., Nandha D.G., Tamboli P.P., Broker D.R. Proportion of anxiety and depression prevalent in patients visiting homoeopathic hospital OPD in rural area of Miyagam – Karjan using Hospital Anxiety and Depression Scale (HADS): a cross sectional survey. Journal of Integrated Standardized Homoeopathy (JISH) 2018; 01(02):
- Goda SC, Ambekar NB, Tamboli PP, Broker DR. Exploring the role of Homoeopathy in the management of malnutrition in children in the age group of 1 to 4 years: A pilot study. Journal of Integrated Standardized Homoeopathy (JISH) 2018; 01(01):12-20.
- Kumar R., Research Methodology a step by step guide for beginners, Chapter 16 Displaying Data, 3rd edition, Sage Publication, 2011, 264 – 274 p. available from: https://www.pdfdrive.com/research-methodology-a-step-by-step-guide-for-beginners-e18463258.html
- Tiwari N.L., Tamboli P., Diabetes Mellitus – Defining scope and clinical approach for homoeopathic management. Indian Journal of Research in Homoeopathy 2008: 2: 28 – 36
- Broker D.R., Tamboli P.P. A Need – based analysis to develop a curriculum for the subject of Research Methodology and Biostatistics for postgraduate Homoeopathy students. Journal of Integrated Standardized Homoeopathy (JISH)2018; 01(02)
- Jani N., Nigwekar A., Clinical Analysis of Geriatric Patients in the light of Homoeopathy. Indian Journal of Gerontology 2016: 30(2): 276 – 283
- Goda C., Tamboli P., Patil S., Valimbe S., Mhatre K., Role of Homoeopathic treatment in scabies infection in Adivasi children attending ashram shalas (resident schools) Indian Journal of Research in Homoeopathy 2010: 4 (2): 33 – 40
- Dhawale K.M., Tamboli M.P., Katawala M.Y., Tambitkar N.N., Tamboli P.P. Use of homoeopathic remedies in the management of learning disabilities. Indian Journal of Research in Homoeopathy 2014;8:87-94.