Measure of Association (adolescent sexual behavior)

[ban26ana]

I'll ask my question first, and then show you the problem from the book. Do I use lamda or gamma to calculate this? I can't figure out if the age groups would be considered ordinal or nominal. Or if the frequency would be ordinal or nominal.

Here is the problem from the book:

In a study of adolescent sexual behavior in Ontario, Canada, sexually active teens were asked to report their use of condoms and birth control. The following cross-tabulation of data is presented separately for boys and girls. Results are presented for three age groups.

Males (N=2918)
Frequency of Use of Protection Age of Teen
12-13 14-15 16-17 Total
Always 145 401 532 1078
Intermittent 81 301 213 595
Never 355 557 333 1245
Total 581 1259 1078 2918

Females (N=2824)
Frequency of Use of Protection Age of Teen
12-13 14-15 16-17 Total
Always 75 586 726 1387
Intermittent 48 147 284 479
Never 264 446 248 958
Total 387 1179 1258 2824


Calculate the appropriate measure of association for these tables. Interpret the relationship between age and frequency of use of protection for the sample of sexually active Ontario teens. Is there a difference in the level of association for boys and girls?

 

[royhaas]

Re: Measure of Association

Age groups should be considered ordinal if you want to consider a monotone relationship between age and another variable. Frequency is a ratio variable.