Purpose. There is a chronic shortage of foster families. Compounding this problem, many licensed foster families are unwilling to accept the growing number of harder to place children (e.g., teenagers, children with emotional and behavioral problems, children with special needs). Using Latent Cluster Analysis (LCA), different subgroups of foster families are identified based on their willingness to accept different types of children.

Methods. The Willingness to Foster Scale (WFS) was administered as part of a national survey of 304 licensed foster mothers from 35 states who were recruited from foster parent associations. The WFS has four subscales, measuring willingness to foster: (1) children with emotional and behavioral problems (40 items, a = .96); (2) children with special needs (19 items, a = .90); (3) children who are older (4 items, a = .77); and (4) children from a different race, religion, culture, or sexual orientation (4 items, a = .66). Each item is rated on a 4-point scale ranging from not willing to foster this child under any circumstances (1) to willing to foster this child without any extra help or support (4). For each subscale, higher scores indicate a greater willingness to foster.

Results. Latent Cluster Analysis (LCA) was used to classify foster mothers based on the four WFS subscales. Latent Gold was used to conduct this analysis. Models with one through six latent clusters were estimated, and numerous indices of model fit compared. A four-cluster model provided the best fit. Mothers in Clusters 1 (21%) and 2 (35%) were characterized by relatively high scores on all subscales relative to mothers in Clusters 3 (35%) and 4 (9%). However, mothers in Cluster 1 were distinguished from those in Cluster 2 by an almost complete and unanimous willingness to foster children from a different race, religion, culture, or sexual orientation. Also, mothers in Cluster 4 were distinguished from those in Cluster 3 by an almost complete and unanimous unwillingness to foster older children.

Implications. Latent Cluster Analysis is well suited for developing and describing different subgroups of foster families, but LCA has rarely been used in child welfare research. By identifying and understanding different subgroups of foster families, agencies could increase placement stability by better matching foster families and children. Also, with this information, agencies could improve foster parent recruitment, training, support, and retention.