Refer to \"The Use of Electro-Acupuncture in Conjunction with Exercise for the T
ID: 3359975 • Letter: R
Question
Refer to "The Use of Electro-Acupuncture in Conjunction with Exercise for the Treatment of Chronic Low-Back Pain," by Yeung, Leung, and Chow.Examine the question the researchers were trying to answer and discuss/explain why you feel the t-test was chosen. Choose one of the other tools you have learned in school or work related and explain why it would not provide relevant findings.
(Read "The Use of Electro-Acupuncture in Conjunction With Exercise for the Treatment of Chronic Low-Back Pain," by Yeung, Leung, and Chow, from Journal of Alternative and Complementary Medicine (2003).)
Explanation / Answer
Tthe question the researchers were trying to answer: Whether a series of electro-acupuncture (EA) treatment in conjunction with exercise on the pain, disability, and functional improvement scores of patients with chronic low-back pain (LBP) is more effective than the conventional exercise only treatment or not.
To answer this question the researchers categorized patients randomly into the control (26 members) and treatment group (26 members). And performed several statistical testing procedures (t-test, chi-sq test, repeated measures analysis of variance) to find out the answer to the research question.
Here the researcher has chosen the t-test due to appropriateness. The objective was to see the mean difference between the groups and there are several cases where the t-test is most appropriate.
Let’s consider the pre-treatment data for the two groups for a quantitative variable (say NRS score), to see whether there is a significant difference between the means for groups or not. Here the test is a two sample test for population mean, the sample sizes are small (less than 30 so we can’t apply the Z test) and the population standard deviations are unknown. Clearly the most appropriate parametric test is the t-test. Any other statistical techniques like correlation analysis, regression analysis or any other tests are used in situations (with certain objective) which are completely different than this one. Thus the t-test is used to check between the significant differences between the groups for particular variable on particular time period.
Next let’s consider the test within the group for two different time periods (say pre-treatment and 1-month follow up) to see whether the follow up has significantly lower mean or not. Here as the individuals are same so the data is the dependent and the comparison in mean of these dependent samples should be done using a paired or dependent t-test. None of the other statistical tool considers the dependency between the samples to test the hypothesis.
Thus the t-tests were appropriate in situations for this research.