1-2 Paragraph Response to Classmate’s Post – Statistical Significance and Meaningfulness. SPSS

By Day 5

Respond to at least one of your colleagues? posts and explain the benefits and consequences of the ?relaxed? level of significance.

Classmate’s Post:

“The true power of statistics lay in how the information is described and understood. Significance does not always mean meaningful (Laureate Education, 2016). Significance is based on the null hypothesis and has a ?low probability? that differences between sample means are due to random error (Cosby, p. 265, para. 1, 2012). Meaningfulness is defined by the results? ?applicability? to real situations (Laureate Education, 2016). We conduct a null hypothesis to determine significance. Probability needed to determine significance is also called the ?alpha level?, with the most common probability is .05, meaning out of 100, there are 5 chances the results contain random error in the sample (Cosby, p. 266, para. 3, 2012). The word ?significance? should only be used in context of probability after a test has been conducted (Bangdiwala, 2016).

We can either accept or reject the null hypothesis. When we reject the null hypothesis, we are saying we have determined the research hypothesis is true to the population because the population means are not equal and there is an effect between variables. When we accept the null hypothesis, we are rejecting the research hypothesis in that the population means are equal (Cosby, 2012).

If the null hypothesis is true and we reject it, we have committed a type l error. If the null hypothesis is false but we accept it, we have committed a type II error (Frankfort-Nachimas & Leon-Guerrero, 2018). While the most common alpha is .05, we can eliminate the type l error risk by changing the alpha to .01 (Frankfort-Nachimas & Leon-Guerrero, 2018), (Cosby, 2012). It must be noted that when we lower the alpha we increase the odds of a type II error (Frankfort-Nachimas & Leon-Guerrero, 2018). The probability of making a type II error is based on three factors. 1. The alpha level. By lowering the level to .01 we make it difficult to reject the null hypothesis, thus increasing the chance of its acceptance (Cozby, 2012). 2. Sample size. A larger sample size is always best to determine true differences. 3. Effect size. A type II error is unlikely when there is a large effect size.

A type I error is less likely to cause harm (Frankfort-Nachimas & Leon-Guerrero, 2018). A type II error can hold more detrimental effects to participants (Cozby, 2012). The effect of the errors also depends on situational contexts. For example, a type I error in the context of criminology would mean finding a defendant guilty when they are innocent (Cozby, 2012). In this example we can easily see how bad that could be. In the medical field a type II error of deciding to not operate when the patient needs it to survive, would be more detrimental (Cozby, 2012). As far as research goes, researchers are generally more careful to avoid type I errors, especially when their results may be published (Cozby, 2012).

A researcher will ?relax? the alpha to a .01 from a .05 to eliminate the odds of a type I error (Frankfort-Nachimas & Leon-Guerrero, 2018). In terms of contextual research, a type I error may be more harmful (Cozby, 2012). Thus, eliminating error in a type I, is more acceptable even though it widens the possibility of a type II error. All published research is not quality. A scientific or statistically significant fact, does not necessarily mean truth. It simply means it is true to the best of our knowledge thus far. Remember, we once thought the world was flat (and apparently some are starting to do so again).”