Issue link: https://www.massageandbodyworkdigital.com/i/68005
SOMATIC RESEARCH The mean duration of backaches was compared between the pretreatment period and the treatment period, as shown in Image 2. The group showed a decrease in mean backache duration from 5.88 days to 2.88 days during the treatment period; each study participant also experienced a decrease in mean backache duration. However, the p-value of this measurement did not achieve statistical significance (p = 0.089), due to the small number of participants in the study. In this hypothetical study, all of the participants individually experienced shorter backaches during the treatment period. In addition, the group average length of backache duration decreased as well. However, the study had a small sample size, and therefore insufficient power to reliably detect a treatment effect. So, even though the researchers observed a trend, with a p-value of .089 or 8.9% chance of this being a false positive result—greater than the 5%, in other words—that trend was not statistically significant. For each participant in the study, and for the group mean, there are two geometrical shapes connected by a line. The first shape indicates the average baseline value of backache duration before the treatment phase of the study began. The second shape indicates the average duration of backaches during the treatment phase of the study for that participant. All of the lines move downward as we read the graph from left to right, indicating that—even though it turned out not to be statistically significant—there was a definite trend for all individuals, and for the group as a whole, to experience shorter backaches in the treatment phase. Sometimes, such trends in the data, even if not statistically significant in themselves, can serve as the basis for deciding whether it makes sense to study the question again with a larger sample size, to see whether it is a real effect on outcomes. 8 7 6 5 4 3 2 1 0 PRE-TREATMENT PERIOD Image 2. Duration of backaches, in days. Finally, the team examined the effects of their massage protocol (procedure) on the intensity of backaches (how much pain) the study participants experienced. Study participants were asked to fill out questionnaires about the pain they experienced, including a visual analog scale ranging from 0 to 10, where 10 indicated the most severe pain. Image 3 presents the results of those questionnaires on reported pain intensity. A wide range of variation in treatment effect was reported by participants, with two participants reporting major reduction in pain, one participant reporting some reduction in pain, and one participant indicating that pain had actually increased during the treatment period. The group average intensity declined, but not significantly (p = 0.23 [23%, or slightly more than 1 time out of 5 that you would see this result by chance as a false positive]). Participants 1 and 2 showed a steep drop in the line as you read the graph from left to right—those are the two who showed the "major reduction in pain" described in the preceding paragraph. Participant 3 also showed an improvement—the line 130 massage & bodywork january/february 2009 dropped, but not as much as the other two—and shows the participant who showed "some reduction in pain." Interestingly, Image 3 shows that Participant 4 had backaches that got more intense over the course of the study. The Results section is not where you expect the researcher to address what something like this means, only that it happened. If interpretation of why this person experienced more intense backaches is addressed, it would fall under the Discussion section. As you would expect for a group average, the results lie somewhere in the middle—as a group, the backache intensity decreased, but the mild improvement and the increase averaged out with the two marked improvements to make a group average between "marked" and "mild." The effect of small sample size is also demonstrated dramatically here—when 1 participant out of 4 has an increase in backache intensity, that represents 25% of this study population. One participant out of 100 in a different study, however, would only represent 1%, and the implications of those percentages for the larger population of massage clients TREATMENT PERIOD Participant 1 Participant 2 Participant 3 Participant 4 Group mean