Class label | Class 1 | Class 2 | Class 3 | Class 4 |
Content Caseys (%) | Infrequent Indigos (%) | Low-Functioning Lees (%) | Multiple-Partner Morgans (%) | |
(n=913) | (n=583) | (n=218) | (n=173) | |
Overall class size | 48.86 | 30.94 | 11.65 | 8.62 |
No sex in last 4 weeks* | 16.63 | 99.39 | 42.48 | 25.51 |
Had unsafe sex in past year* | 0.01 | 0.01 | 0.12 | 68.60 |
≥2 sex partners in past year* | 1.24 | 0.02 | 5.93 | 99.89 |
Diagnosed with any STI in the last 5 years | 0.78 | 0.02 | 2.34 | 4.30 |
Paid for sex in past year* | 0.14 | 0.51 | 1.71 | 6.43 |
Low sexual function score | 4.55 | 5.63 | 98.81 | 17.17 |
Dissatisfied with sex life | 4.81 | 32.16 | 53.78 | 9.34 |
Distressed or worried about sex life | 1.13 | 12.19 | 57.40 | 7.42 |
All estimates are unweighted.
Probabilities >50% are shown in bold type to indicate items that members of a given class were more likely to report. In naming and characterising the classes, dimensions in which classes differed strongly and a probability of >50% for a certain item were considered an indication that members of a given class were more likely to report that risk factor.
The methodology we employed presupposes the existence of discrete latent classes. However, there may be individuals who better fit one or more continuous latent dimensions. The model output expresses the probability of class membership for all individuals for each class as well as the probability of each characteristic given membership of each class. Given each individual had a particular probability of class membership for each class, individuals were assigned to the latent class for which they had the greatest probability of membership. Each class was named after the item that was most prevalent in members of that class.
*Same and opposite sex.