RESEARCH NOTES
HOW TO ANALYSE YOUR RESEARCH DATA? ILLUSTRATIONS WITH HANDS-ON EXERCISES USING SPSS.
Loh Keng Yin1 MMed (FamMed, UKM), Teng Cheong Lieng1 MMed (FamMed, UM) FRACGP, Wong Kam Cheong2 MBBS, MSc.
1International Medical University, Malaysia; 2University of Queensland, Australia
Address for correspondence: Dr Loh Keng Yin, Senior Lecturer, International Medical University, Jalan Rasah, 70300 Seremban, Malaysia. Tel: 06-7677798, Fax: 06-7677709, Email: kengyin_loh@imu.edu.my
Loh KY, Teng CL, Wong KC. How to analyse your research data? Illustrations with hands-on exercises using SPSS. Malaysian Family Physician. 2006;1(2&3):77-81
INTRODUCTION
Statistical analysis for a quantitative study is often perceived to be the most difficult step by a novice researcher. On the other hand, some researchers tend to over-analyse their research data in search of the illusive “significant” p-value. Some of these problems and pitfalls can be reduced if the researchers give some thoughts to their research objectives.1
Another issue that trouble the novice is how much statistical knowledge one needs to have. There is no straight answer to this question; we feel that the information provided in this article is probably the bare minimum needed by most, if not all, researchers embarking on a research project. What about performing your own statistical analysis using statistical software? Although ability to handle statistical software is desirable, it is not mandatory as it is now possible to outsource to people who can do this properly. The researcher should, however, be able to tell the statistician what analysis is needed and to interpret statistical results. Take note that the statistician cannot undo the errors in the data (e.g. inadequate research design, inappropriate definition of research variables, inaccurate measurement during data collection, or data entry errors) – hence great care must be exercised during these earlier steps of research process.
There are several statistical packages available to assist you in data analysis. SPSS software is applied in the following example. You can download a free trial version from www.spss.com (prior registration necessary)
DESCRIPTIVE STATISTICS
We shall start by this example: You have conducted a survey of 160 diabetic patients in your clinic. The mean HbA1c of these patients was 8.9% (SD=2.2, range 5.2-15.7). The gender breakdown is males 44.4%, and females 55.6%. The ethnicity breakdown is Malays 28.1%, Chinese 41.3%, and Indians 30.6%. Other summarised data are given in Table 1. You may download this dataset (in SPSS format) from the online journal (hba1c.sav, link) (please request the hba1c.sav file from the Editor) or download a zip file here.
The SPSS commands for obtaining the above statistics for ‘HbA1c’ are as follows:
From the menus choose:
- Analyze
- Descriptive Statistics
- Descriptive
- Select the variable ‘hba1c’
- Click ‘OK’
The SPSS commands for obtaining the above statistics for ‘gender’ and ‘race’ are as follows:
From the menus choose:
- Analyze
- Descriptive Statistics
- Frequencies
- Select the variable ‘sex’ and ‘race’
- Click ‘OK’
Table 1: Mean HbA1c by gender and ethnicity
| Characteristics | HbA1c, % (SD) |
| Gender | |
| Male (n=71) | 8.7 (1.9) |
| Female (n=89) | 8.9 (2.3) |
| Ethnic group | |
| Malays (n=45) | 9.6 (2.5) |
| Chinese (n=66) | 8.4 (2.0) |
| Indians (n=49) | 8.7 (1.9) |
| All patients (n=160) | 8.9 (2.2) |
The SPSS commands for obtaining the statistics ‘HbA1c’ breakdown by ‘Gender’ in Table 1 are as follows:
From the menus choose:
- Analyze
- Reports
- Case Summaries
- Select ‘hba1c’ into the ‘Variable’
- Select ‘sex’ into the ‘Grouping Variable’
- Click on the icon ‘Statistics’ and select ‘Number of Cases’, ‘Mean’, and ‘Standard Deviation’
- Uncheck these two boxes ‘display cases’ and ‘Limit cases to first 100’
- Click ‘OK’