Many excellent statistical software packages available commercially (e.g. SAS®, SPSS®, S-Plus®) can perform the statistical procedures outlined in this chapter. Some of these software packages, such as SPSS, are very user friendly. In addition to procedures for data analysis, some software, such as SAS, can be used to assist in developing a study design. With the help of such software, we can perform statistical tasks much more efficiently and with greater power than ever before. Indeed, statistical procedures that once took months to calculate can now be performed in less than 5 seconds. Although this is a blessing, it can also become problematic. That is, in the wrong hands these powerful statistical software packages can produce misleading if not downright false results.
One must exercise great care and use caution when conducting complex statistical analyses. As such, to increase the chances of producing a valid analysis, we offer the following recommendations: (1) Whenever possible, and ideally at the study design stage, consult with a statistician regarding the best statistical approach to addressing a given research question. (2) Begin with graphical and/or tabular displays of the data to develop a sense of what the data look like and to assess possible errors in data input and processing. (3) Carefully examine your data to assess whether the assumptions required for the use of parametric methods are adequately met (sometimes a simple data transformation can transform skewed data into a normal distribution, thereby meeting a parametric assumption). (4) Always opt for the simplest statistical analysis possible to test a given hypothesis.
Was this article helpful?
Your heart pumps blood throughout your body using a network of tubing called arteries and capillaries which return the blood back to your heart via your veins. Blood pressure is the force of the blood pushing against the walls of your arteries as your heart beats.Learn more...