Question

Please answer using your own words (i.e., don't quote directly from the text).: Describe and contrast the three ways of describing results: comparing group percentages, correlating scores, and comparing group means. For each, give two examples (other than those given in the text) of their appropriate use.

Answer 1

comparing group percentages: classic case for a χ2(chi-Square ) test. This is a basic test of association, and will give you a p-value to assess statistical significance. Just beware that χ2 tests are sensitive to sample size. This test is about “Consistency”. In certain cases average or mean is misleading .

Example 1: Bulls-I toy example to hitting target is not the only criteria and also maintain consistency while hitting target.

Example 2:Suppose 100 organizations are given and few are service based and other is product based .If you would like to know which one is making profits you need association and need to compare their group percentages .

correlating scores: Correlation is a measure to find the relationship between two variables like positive, negative, or no relationship.

Example1: Height of a person from the age between 1-19 years .it is positive relationship because with the age height also increases.

Example 2: Car resale value over the period of time. Which means with the usage of car resale value will decrease. Hence negative relationship.

Example3: Height growth of a person from age group 25-35 will have no relationship which means even though age increases height will not have any impact.

comparing group means: We use Z and T test to compare mean. In statistics 3 is not equal to 3 unlike mathematics in stats you need to conduct test and then say equal or not. Suppose student1 got marks in two subject’s 1 and 5 and mean is 3 and student2 got marks in two subjects 3 and 3 and average is 3 so which one would you prefer is the challenge here. So mean groups also compared by statistical tests.

Example 1: two compare mean income of two different organizations

Example 2: Compare the quality of originality of two different pearl sets based on their mean scores by set of attributes .