Statistics for Health Care Research: The T-Values Essay

1. The t-values in table two reflect the effects of the three months health promotion program on cardiovascular risk factors (Kim, Junes &Song, 2003, 375). The first column of t-ratios compares the pre-test and post-test results while the second column of t-ratios compares the pre-test and the tests took three months after the program completion.
2. From table 2, the t-ratio of 4.14 for the total risk score represents the greatest difference between the pre-test and post-test results. The degrees of freedom n-1 is 21–1=20 and α=0.05. Therefore, the p-value=0.0005 (from the t-table). The difference is statistically significant because the p-value=0.0005˂0.05= significance level (Hazewinkel, 2001, 6).
3. The t-value of -7.75 for program effects on exercise reflects the least difference in the effects on behavior. With df=20 and α=0.05, the p-value=1(t-table). This difference is not statistically significant because the p-value=1 ˃ 0.05=level of significance. These results mean that the program did not statistically affect the exercising behaviors of the elders (Hazewinkel, 2001).
4. Research qualifies for a t-test under the following assumptions:
5. The entries are assumed to follow a normal distribution.
6. For the dependent variables, the entries are recorded at ratio levels. The interval recording also meets this assumption.
7. The variables whose differences are under study are related. In this case, the relationship can be attained by the subjects acting as their own control groups or through matching (Asghar & Saleh, 2012).

In this study, requirement “b” is met through ratio level recording. Requirement “c” is also met through matching after the elimination of the participants who did not complete the program and by the subjects acting as their own control groups (Asghar & Saleh, 2012).

1. For the exercise variable in table 3, the t-ratio of -3.93 for 3 months after the program is larger than the t-ratio of -7.75 immediately after the program. Since the larger the t-value the larger the difference, it means that the results for 3 months after the program were more different from the pre-test results as compared to the results immediately after the program (Hazewinkel, 2001). This indicates that the program has a larger long-term effect on the exercising behaviors of the elders as compared to the short-term effect.
2. To determine the smallest significant t-ratio in Table 2, we determine p-values starting from the smallest t-ratio. The t-ratio of 2.03 which has a p-value of 0.056 and all other t-ratios below it is insignificant because they have p-values greater than 0.05. Therefore, the smallest significant t-ratio is 2.20 with a p-value of 0.04 ˂ 0.05 (Venables & Ripley, 2002).
3. Larget ratios are more likely to be significant because they yield smaller p-values and the smaller the p-value the more likely the difference is significant (Venables & Ripley, 2002).
4. As the analysis shows, the blood pressure was not significantly affected by this promotion. This is because it yielded smaller t-ratios with p-values˃0.05.
5. For systolic blood pressure, the means indicate that there were improvements from pre-test through to 6 months. The standard deviation reduces from pre-test to post-test. It again increases from post-test to three months after post-test. Clinically, while the means show improvements, the increase in deviations indicates relapse cases.
6. This design is not strong. It is not randomized and does not control extraneous factors. Therefore, it is weak.
7. I would not recommend the implementation of this program even though the risk factor outcomes showed a significant difference. This is because while the p-value of 0.005 after completion shows a strong difference, it increases to a p-value of 0.019 after another 3 months. This is a trend that may negate the gains after some time. In addition, this could be due to other uncontrolled factors (Hazewinkel, 2001).

References

Asghar, G. & Saleh Z. (2012). Normality tests for statistical analysis: A Guide for Non-Statisticians, 10(2). 12-13.

Hazewinkel, M. (2001), “Student distribution”, Journal of Mathematics, 4(2): 6-12

Kim, C., Junes, K., & Song, R. (2003). Effects of a health-promotion program on cardiovascular risk factors, health behaviors, and life satisfaction in institutionalized elderly women. International Journal of Nursing Studies, 40(4), 375–81.

Venables, W.N. & B.D. Ripley, B.D. (2002). Modern applied statistics with S. Statistics and Computing, 4(1), 87-93