|
| |
 |
|
| STATISTICAL RESOURCE |
|
| Year : 2021 | Volume
: 4
| Issue : 1 | Page : 139-140 |
|
Algorithm to choose the appropriate test for comparison between two groups
Sampada B Dessai1, Vijay M Patil2
1 Department of Surgical Oncology, Sir HN Reliance Hospital, Mumbai, Maharashtra, India
2 Department of Medical Oncolocgy, Tata Memorial Hospital, Mumbai, Maharashtra, India
| Date of Submission | 22-Feb-2021 |
| Date of Decision | 01-Mar-2021 |
| Date of Acceptance | 05-Mar-2021 |
| Date of Web Publication | 26-Mar-2021 |
Correspondence Address:
Vijay M Patil
Department of Medical Oncology, Tata Memorial Hospital, Mumbai, Maharashtra
India
 Source of Support: None, Conflict of Interest: None
DOI: 10.4103/crst.crst_45_21
The current manuscript describes the method of selection of statistical tests for comparing two groups. In this article, we discuss the factors on which the selection of tests depends and provide an algorithm for the selection of the inferential statistical test.
Keywords: Chi-square, comparison, median test, statistics, t-test
How to cite this article:
Dessai SB, Patil VM. Algorithm to choose the appropriate test for comparison between two groups. Cancer Res Stat Treat 2021;4:139-40 |
| Background | |  |
Research is the most essential and yet one of the most neglected aspects of advances in oncology.[1],[2],[3],[4] Selection of appropriate statistical tests, testing of assumptions, and interpretation of the results are rarely taught even in the premier medical and academic institutions of the country.[5] To bridge this gap in the knowledge of the budding oncologists and to improve their understanding about the basics of statistics, Cancer Research, Statistics, and Treatment in every issue publishes an article related to statistics.[6],[7],[8],[9],[10],[11] In continuation with the earlier articles published as a part of this series, the current article deals with one of the most common analyses reported in oncology – comparison of dependent values of interest between two groups, cohorts, or arms. The current article helps in the selection of an appropriate test for comparison.
| Steps in The Selection of Tests | | |
Hypothesis
A hypothesis or supposition could be that there is a difference in the values of a variable between the two groups. An example of the hypothesis could be that there is a difference in the age distribution or albumin levels between two groups. The below-mentioned tests are of value if the idea is to see a difference between the two groups, classically known as the “difference” in inferential statistics. These tests are of no value if an association or relation between two variables is being sought – for example, determining whether a response seen on magnetic resonance imaging correlates with the pathological response.
Type of variable
The type of variable is the next important aspect of selecting a test to assess the difference between two groups using inferential statistics. Variables can be of three types:
- Normal or scale data - e.g., height, weight, biochemical levels, and blood count values
- Ordinal data - e.g., response, categories of education, or income
- Dichotomous or nominal data - e.g., presence or absence of abnormal values.
Descriptive to be compared
The choice of test depends upon whether the comparison is for the mean, median, or proportion of normal distribution. This applies to continuous data such as normal or scale data. The selection of a test depends on whether the data are normally distributed. Normal distribution is tested using a P-P or Q-Q plot, subjectively. A plot with a straight diagonal line with the observations falling on the line suggests normal distribution. However, the interpretation of this plot is subjective. Objective evaluation of the normal distribution can be done using the Shapiro–Wilk test. A P < 0.05 suggests that the variable is not normally distributed.
Relation between the independent groups
Groups to be compared can be independent or non-independent. In independent groups, the subjects in one group cannot also be a part of the other group, whereas in non-independent groups, the same subjects are assessed before and after an intervention. For example, a study assessing the adverse event of a drug will comprise two independent groups, whereas a study assessing the effect of a hematinic on the hemoglobin levels will comprise non-independent groups in which repeated measures are taken on the same subjects.
| Selection of The Appropriate Test | | |
The algorithm outlined in [Table 1] can be followed for the selection of an appropriate test.
| Conclusion | | |
This article provides a simplistic overview of which test to select when two groups are to be compared.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
| References | | |
| 1. | Noronha V. Making a case for cancer research in India. Cancer Res Stat Treat 2018;1:71-4.  [Full text] |
| 2. | Maruti Patil V, Singhai P, Noronha V, Bhattacharjee A, Deodhar J, Salins N, et al. Effect of early palliative care on quality of life of advanced head and neck cancer patients: A phase III trial. J Natl Cancer Inst 2021;doi: 10.1093/jnci/djab020. Online ahead of print). |
| 3. | Patil V, Noronha V, Dhumal SB, Joshi A, Menon N, Bhattacharjee A, et al. Low-cost oral metronomic chemotherapy versus intravenous cisplatin in patients with recurrent, metastatic, inoperable head and neck carcinoma: An open-label, parallel-group, non-inferiority, randomised, phase 3 trial. Lancet Glob Health 2020;8:e1213-22. |
| 4. | Institute of Medicine (US) Committee on Health Research and the Privacy of Health Information: The HIPAA Privacy Rule; Nass SJ, Levit LA, Gostin LO, editors. Beyond the HIPAA Privacy Rule: Enhancing Privacy, Improving Health Through Research. Washington (DC): National Academies Press (US); 2009. 3, The Value, Importance, and Oversight of Health Research. https://www.ncbi.nlm.nih.gov/books/NBK9571/. [Last accessed on 2021 Mar 05]. |
| 5. | Zodpey SP, Negandhi HN. Training in clinical research in India: Potential and challenges. Indian J Community Med 2009;34:173-4. [ PUBMED] [Full text] |
| 6. | Darling HS. Basics of statistics – 4: Sample size calculation (ii): A narrative review. Cancer Res Stat Treat 2020;3:817-28. [Full text] |
| 7. | Manjali JJ, Gupta T. Critical appraisal of a clinical research paper: What one needs to know. Cancer Res Stat Treat 2020;3:545-51. [Full text] |
| 8. | Darling HS. Basics of statistics-3: Sample size calculation – (i). Cancer Res Stat Treat 2020;3:317-22. [Full text] |
| 9. | Darling HS. Basics of statistics – 2: Types of clinical studies. Cancer Res Stat Treat 2020;3:100-9. [Full text] |
| 10. | Darling HS. Basics of statistics-1. Cancer Res Stat Treat 2019;2:163-8. [Full text] |
| 11. | Dessai S, Patil V. Testing and interpreting assumptions of COX regression analysis. Cancer Res Stat Treat 2019;2:108-11. [Full text] |
[Table 1]
|
Related articles