episode to characterize the sample. To test the factor
structure of the Coping Scale, Exploratory Factor Analysis
(EFA) was conducted on the original 14 items using
principal components analysis. EFA provides preliminary
information as to whether there are underlying factors that
are present in a scale specific to a particular sample
(DeVellis 1991; Guada et al. 2011). The scree test was used
to determine the number of factors to retain (DeVellis
1991). The scree test allows for visual examination of a
graphical illustration of the eigenvalues. The eigenvalues
are illustrated by dots and presented in descending order
(Ledesma and Valero-Mora 2007). These dots are linked
with a line and the point at which the line levels off
establishes the cutoff point (Ledesma and Valero-Mora
2007). As such, the number of factors that should be re-
tained is determined by the eigenvalues that fall above this
cutoff point. With regard to missing data, only one case had
missing data on the variables examined in the current
analysis and was therefore excluded by listwise deletion.
Hierarchical logistic regression analyses were then
conducted to examine how identified coping factors were
associated with major depressive episode, controlling for
demographic and homelessness experience variables.
Variables were entered sequentially in 3 steps. First, de-
mographics were entered, particularly focusing on age,
gender, and ethnicity. Second, homelessness experiences
including primary residence, transience, and length of time
homeless were entered into the model. Finally, coping style
factors were entered into the third block of the model. Our
primary interest was to determine if types of coping ac-
counted for a significant amount of variance over and
above demographic and homelessness experience vari-
ables, and these coping variables were therefore entered
together in the third block of the hierarchical regression
model.