Hi. Similar to last week, take a quick look over of this website: http://www.statisticssolutions.com/ordinal-association/Links to an external site.. If you cannot directly open the link then copy and paste the url into the address bar.

Gamma is used for testing the association between ordinal variables.

1. Why is it that gamma can show direction of an association and Lambda does not? Hint: think about the types of measures that each statistic is used for (i.e., nominal, ordinal, interval/ratio).
2. Looking at the results for Somers'd on the website, how come you would want to analyze the Somers'd for the father's education level as the independent variable and not child's level of education?

3.The gamma value is.484. What do you make of the statistic that knowing a father's education level can improve our prediction of the child's education level by 48 percent?

1. What happens if there is a zero in the crosstabulation when you want to analyze gamma? Please reference the article you just read to answer why you cannot analyze gamma if there is a zero in the crosstab. Note: I am not saying that gamma equals zero.

must cite and reference