Introduction

Can strength training 30 minutes everyday improve performance on an attention test? The population parameter of interest for this study was the true difference in population proportions of peoples’ improved attention after doing exercise versus people who did not do exercise.

Existing literature shows mixed findings for the effects of strength training specifically on attention. Hawkins et al. reported that exercise was associated with improvement in attention proccesses of older adults (1992). Multiple studies using aerobic exercise as the treatment condition detected positive effects on attention, while only some studies using resistance training as the treatment detected effects (Fernandes M. de Sousa et al., 2019). However, one study in particular found that a session of resistance exercise increased both attention and executive function (another major factor in ADHD) test scores, while an aerobic exercise session improved only the executive function scores (Tsuk et al., 2019). I believe that it is worth investigating whether 2 days of consecutive strength training can positively impact attention.

Data Collection Methods

The observational units in this study were 30 participating Islanders from across the Islands (n=30, ages 20-91), randomly selected and randomly assigned into 2 conditions (random number generators allowed for random selection and random assignment). The variables were strength training exercise (explanatory, binary categorical) and improved attention (response, quantitative).

On Day 1, both groups 1 & 2 completed Attention Test 10 mins task. Consecutively, group 1 (treatment group) completed Strength Training 30 mins task. On Day 2, group 1 (treatment group) repeated Strength Training 30 mins task. Consecutively, both groups repeated the Attention Test 10 mins task. I then subtracted the 2nd scores from the 1st scores for an improved attention score.

Descriptive Statistics

library(readr)
STN<-
read_csv("~/MATH 247/MP3 Clare Data - Sheet1-2.csv")
head(STN, n=2)
data(STN)
favstats(Improvedattention ~ Exercise, data = STN)
bwplot(Exercise ~ Improvedattention, horizontal = TRUE,
main="Side-by-side boxplots", data = STN)

Just based on the box plot, I could observe that I would probably retain my null hypothesis. Contrary to initial expectations, the no exercise (control) group had higher mean improvement as well as more variability (longer box and whiskers).

Analysis of Results

This study’s sample was made up of 30 Islanders aged 20-91 who were willing to particpate in a research study. The population of interest is the general population across all islands above the age of 18 who are willing to participate in a research study. The parameter of interest for this study was the true difference in population proportions of peoples’ improved attention after doing exercise versus people who did not do exercise.

Null hypothesis: There is no significant difference in peoples’ improved attention after doing exercise versus people who did not do exercise.

Alternative hypothesis: There is a significant popsitive difference in peoples’ improved attention after doing exercise versus people who did not do exercise.

H0: \(\mu_{ST}\)-\(\mu_{N}\) = 0

H1: \(\mu_{ST}\)-\(\mu_{N}\) > 0

In this study, a type I error would represent rejecting the null hypothesis and falsely stating that the study found a significant positive difference when it does not exist. A type II error would represent retaining the null hypothesis and falsely stating that the study did not find a significant positive difference when it did exist.

As the sample was randomly selected and assigned, it can reasonably be considered a representative sample from the population of interest.

n.st<- 15
n.n<- 15
x.bar.st<- -2.93333333
x.bar.n<- 0.06666667
SD.st<- 5.049281
SD.n<- 9.437514
# difference between the sample proportions of exercisers vs resters
x.bar.diff<-x.bar.st-x.bar.n
cat("the difference in sample means is",x.bar.diff)
## the difference in sample means is -3
set.rseed(888) 
Att.null <- do(1000) * diffmean(shuffle(Improvedattention) ~ Exercise, data = STN)
head(Att.null,2)
dotPlot(~ diffmean,
        data = Att.null, 
        main="Simulated null distribution of the difference in sample means",
        xlab="difference in sample means",
        width = 2,
        cex = 1)

favstats(~ diffmean, data = Att.null)
dotPlot(~ diffmean,
        data = Att.null,
        main="Simulated null distribution of the difference in sample means",
        xlab="difference in sample means",
        width = 2, cex = 1,
        groups = (diffmean >= -3))

SE.x.bar.diff<-sqrt(SD.st^2/n.st + SD.n^2/n.n)
cat("The standard error of the difference in sample means is",round(SE.x.bar.diff,2))
## The standard error of the difference in sample means is 2.76
confint(t.test(Improvedattention ~ Exercise, data = STN))
stat(t.test(Improvedattention ~ Exercise, data = STN))
##        t 
## 1.085542
two.sided.p.value<-pval(t.test(Improvedattention ~ Exercise, data = STN))
cat("the two-sided p-value is",two.sided.p.value)
## the two-sided p-value is 0.2897548

The difference in sample means is -3, the standard error of the difference in sample means is 2.76, and the t-statistic is 1.086. This can be interpreted as the sample difference being 1.086 standard deviations above the mean. The two-sided p-value is 0.29, which is not significant (neither is .0.14). There is a 14% chance that this finding could occur by chance alone. I am 95% confident that the difference in mean improvement score achieved by exercisers and non exercisers is between -2.74 and 8.74. As this confidence interval includes 0, I can say that it is possible that there is no difference in mean improvement score between groups. With this approach, my analysis would lead me to retain the null hypothesis. I would conclude that 2 sessions of strength training do not significantly improve scores in attention tests achieved by islanders aged 20-91.

However, I would not be able to accurately interpret my data using that theory based approach, as it did not satisfy validity conditions for a two-sample t-test. Those conditions called for 20 data values for each group. I continued by testing with the simulated null distribution above.

p_value<-prop(~(diffmean >= -3), data = Att.null)
cat("the right-tailed p-value is", p_value)
## the right-tailed p-value is 0.878

Earlier on I simulated a null distribution of 1,000 possible values of difference in sample means, assuming that exercise condition has no effect on attention improvement. R gave me a simulated right-tailed p-value of 0.878, which is non significant. The chance of observing a difference in sample means of -3 or more extreme assuming the null hypothesis is correct is 87.8%, which is very high.

x.bar.diff <- 3

# simulated standard deviation
SE.x.bar.diff <- sd(~ diffmean, data = Att.null)

# margin of error for 95% CI
MoE <- 2 * SE.x.bar.diff

LB<-x.bar.diff - MoE # lower limit of 95% CI
UB<-x.bar.diff + MoE # upper limit of 95% CI
round(cbind(LB,UB),3)
##          LB    UB
## [1,] -2.281 8.281

I am 95% sure that the long-run mean improvement score is -2.281 to 8.281 higher with the strength training exercise treatment as opposed to the no exercise treatment. This interval is extremely close to that of the t-test approach. As this confidence interval includes 0, I can say that it is possible that there is no difference in mean improvement score between groups. With this approach, my analysis would lead me to retain the null hypothesis. I would conclude that 2 sessions of strength training do not significantly improve scores in attention tests achieved by islanders aged 20-91.

Conclusion

This study’s findings are in line with those from existing literature, of which the effects of resistance training on attention have been inconsistent. This study suggests that participants on the Islands similar to these individuals who were willing to volunteer as participants will not see more improved attention if they complete 2 sessions of strength training exercise compared to if they did not exercise at all.

If I were to conduct this study again, I would use a larger sample size to meet validity conditions for a theory based approach. I would consider achieving this through use of a within-subjects design. Future directions for this research could be explored through a longer treatment period, restricted age group, or with an added 3rd condition of aerobic exercise.

Works Cited

Fernandes M. de Sousa, A., Medeiros, A. R., Del Rosso, S., Stults-Kolehmainen, M., & Boullosa, D. A. (2019). The influence of exercise and physical fitness status on attention: A systematic review. International Review of Sport and Exercise Psychology, 12(1), 202–234. https://doi.org/10.1080/1750984X.2018.1455889

Hawkins, H. L., Kramer, A. F., & Capaldi, D. (1992). Aging, exercise, and attention. Psychology and Aging, 7(4), 643–653. https://doi.org/10.1037/0882-7974.7.4.643

Tsuk, S., Netz, Y., Dunsky, A., Zeev, A., Carasso, R., Dwolatzky, T., Salem, R., Behar, S., & Rotstein, A. (2019). The acute effect of exercise on executive function and attention: Resistance versus aerobic exercise. Advances in Cognitive Psychology, 15(3), 208–215. https://doi.org/10.5709/acp-0269-7