BASED ON LAST 3 LECTURES

COMMENT ON THESE 3 ANALYSES, SAME QUESTIONS AS YOU ANSWERED.

1)

BERNIE

Part One - Multiple Testing

In many ways, comparing multiple sample means is simply an extension of what we covered last week.Just as we had 3 versions of the t-test (1 sample, 2 sample (with and without equal variance), and paired; we have several versions of ANOVA - single factor, factorial (called 2-factor with replication in Excel), and within-subjects (2-factor without replication in Excel). What examples (professional, personal, social) can you provide on when we might use each type?What would be the appropriate hypotheses statements for each example?

Part Two - ANOVA

NOVA is the acronym for analysis of variance (Tanner & Youssef-Morgan, 2013). There are three types of ANOVA and they are single factor, factorial, and within subjects. "ANOVA and the t-test answer the same questionare differences between groups statistically significant?" (Tanner & Youssef-Morgan, 2013, p. 105). An example for the single factor ANOVA is the hardness of water. During one of my science courses we tested the hardness of water by using tap water and two different bottled water samples. All three samples where tested using chemical test strips. Before the tests were conducted I made an alternative hypothesis, but after the test I realized I should have made a null hypothesis. There was not a significant difference of water hardness in the results of each sample.

The factorial ANOVA is an independent variable with multiple groupings within the factor with a single outcome. Women voters is an example of this ANOVA. Specifically, how many women voters were single mothers. This would be a null hypothesis. Regardless of how many women were single mother the outcome of women who voted would not change.

Part Three - Effect Size

The within subject's ANOVA is the same subject under different conditions with repeated tests. For an example of this a heart stress test for patients over 50. If four 50-year-old men have their heart rates monitored for 10 days while all conducting the exact same activities and being able to compare that data. The heart rate would have to be checked after each event by all four men. These tests would lead me to an alternate hypothesis. I think based on other factors of each of the men's health would make a significant difference in the results of heart rate.

2)

SHERRI

Analysis of variance (or ANOVA) is a test of significant differences among two or more independent groups, when the IV is nominal and the DV is interval or ratio" (Tanner & Youssef-Morgan, 2013)."Factorial ANOVA is ANOVA with more than one independent variable. One-way ANOVA involves just one IV" (Tanner & Youssef-Morgan, 2013).Doing multiple tests gives one a broader and different view of how something can be calculated.

For my example of a one-way ANOVA I will use the retail business.If you want to compare each transaction average that each employee sells and the only independent variable is based on how much is sold. The null hypothesis would be that the average transaction would be the same across the board for all employees. The alternate hypothesis would be that if you add the independent variable of more sales per employee and what two or more employees were selling, you would find that individually, their checks would be significantly different.

A factorial ANOVA would add factors like what time of day does sales occur most.Is it a holiday or weekend?What is the personality of the employee?How experienced is the employee at sales?The Null hypothesis: each check average would be the same.Alternative hypothesis: each check average would change based on the additional variables I just mentioned.

From my readings of the effect size, say on someone that is ill, I gather that the more the medicine taken, the better the results given.I am having some trouble understanding this.But then it depends if they are male or female, young or older, already sickly or completely healthy.There are so many questions involved when performing hypothesis tests.

3)

BRIAN

Part One - Multiple Testing

We have learned in the past few weeks that we cannot simply use one statistical test and expect suitable answers. There seems to always be a variance that we must take into consideration. When we try to find the mean equality of three or more groups we can use something called ANOVA or Analysis of Variance. By using ANOVA we streamline group comparison, this way we do not have to compare each group against one another. By using ANOVA we can keep confidence in our statistic results. When we run multiple tests, our confidence drops. When we analyze the variances, we are able to come to a more satisfying conclusion.

Part Two - ANOVA

This test analyzes the differences in results. As we saw in lecture 8 this week, groups A - F were analyzed. One difference we see is that the count of group A and E are significantly more than the other groups. This is a notable variance for this statistic. The means vary from 23.5 to 62.5. This ANOVA test would be beneficial in measuring my department and if everyone is paid equally. My department has 10 "steps" with each comes a 5% raise. This goes for every year of service as long as standard expectations are met. Policing is heavily populated with males rather than females, therefore this test would examine the differences in grades as we saw in lecture 8.

Part Three - Effect Size

Effect size is used to measure the strength of an effect. This is when you take a sample of a population and it's results differ from the population results. This difference is considered the effect size. We want to understand what the variables are that causes such a significant difference. The simplest way to explain this is if we had a group of 100 people trying to lose weight and we took 10 and used a diet pill, then we compared those 10 to the 100. The effect size measures this to determine if the hypothesis is null. If the difference is over our set value.