Examining Clinician Differences that Influence Ratings of the
Probability of Harm and Rehospitalization on a Standard Dangerousness
Assessment
Christina M. Harris
University of Nebraska –
Lincoln
|
|
Results
Analysis of variance
was used to explore the relationship between working in a forensic facility or
not, usefulness ratings of these assessments and the frequency these
assessments are given as they relate to probability of harm ratings and
probability of rehospitalization/ involuntary
outpatient commitment.
Probability of Harm Ratings
Probability of Harm Ratings
As hypothesized,
there was a significant three way interaction of whether or not someone worked
in a forensic facility, how useful they found these types of judgements for
making decisions (usefulness rating), and the frequency of providing formal
assessments (rarely/never, monthly, weekly) of “dangerousness” on total
probability of harm ratings, F (2, 430) = 3.625, MSE= 1,694.271, p=.0270, r=.129,
refer to Table 1 and Figure 1. Also as hypothesized, there was not a lot of
variability in probability of harm ratings. Pairwise comparisons revealed that
for those who work in a forensic facility, harm probability ratings were
equivalent across the three levels of frequency whether the respondent had high
or low usefulness ratings. Among those who do not work in a forensic facility,
probabilities of harm ratings were also equivalent across the three levels of
frequency but only if the respondent had high usefulness ratings. For those who
did not work in a forensic facility and had low usefulness ratings, probability
of harm ratings were equivalent for those who gave dangerousness assessments
never/rarely to monthly and monthly to weekly and harm probability ratings were
greater for those who gave assessments never/rarely than those who did weekly
(r=.273).
Among those who had
low as well as high usefulness ratings, probability of harm ratings were
equivalent among those who worked in a forensic facility and didn’t if they
gave these assessments rarely/never and monthly. However among those who had
low usefulness ratings and gave these assessments weekly, those who work in a
forensic facility had higher probability of harm ratings than those who do not
(r=.355). Results show the opposite effect for those who had high usefulness
ratings such that if the respondent worked in a forensic facility and gave
these assessments weekly, he/she reported lower probability of harm ratings
(r=.262).
Those who provide assessments
of dangerousness never/rarely and monthly gave equivalent ratings of
probability of harm whether they worked in a forensic facility or not and found
these assessments useful or not. Among those that provided assessments of
dangerousness weekly, those who worked in a forensic facility and had low
usefulness ratings gave higher probability of harm ratings that those who had
high usefulness ratings (r=.252). We see the opposite effect for those who give
these assessments weekly and do not work in a forensic facility, with those
with low usefulness ratings giving lower probability of harm ratings than those
with high usefulness ratings (r=.385).
There was not a
significant two way interaction of frequency and whether or not participants
worked in forensic facility as they relate to probability of harm ratings, F
(2, 430) = .373, MSE= 1,694.271, p=.689, r=.042. This interaction was
misleading for those who do not work in a forensic facility and had low
usefulness ratings with those who give assessments never/rarely giving higher
probability of harm ratings than those who give them weekly. This interaction
was also misleading for those who had low usefulness ratings and gave these
assessments weekly with those who work in a forensic lab giving higher probability
of harm ratings than those who do not.
It was misleading for those with high usefulness ratings and gave these
types of assessments weekly such that those who do not work in a forensic
facility gave higher probability of harm ratings than those who do.
There was not a
significant two-way interaction of usefulness ratings and frequency as they
relate to probability of harm ratings, F (2, 430) = .970, MSE= 1,694.271,
p=.380, r=.067. This interaction was misleading for those who do not work in a forensic
facility and had low usefulness ratings with those who give assessments
never/rarely giving higher probability of harm ratings than those who give them
weekly. It was misleading for those who had low usefulness ratings and gave
these assessments weekly with those who work in a forensic lab giving higher
probability of harm ratings than those who do not. It was misleading for those with high
usefulness ratings and gave these types of assessments weekly such that those
who do not work in a forensic facility gave higher probability of harm ratings
than those who do.
There was a
significant two way interaction of whether or not participants worked in a
forensic facility and how useful they found these assessments on probability of
harm ratings (F (1, 430) = 8.746, MSE= 1,694.271, p=.003, r=.141) such that
those who do not work in a forensic facility and had high usefulness ratings
tended to have higher probability of harm ratings than those who had low
usefulness ratings (r=.167). Refer to Table 2 for the means and standard
deviations. Among those who worked in a forensic facility, probability
of harm ratings were equivalent for those with high and low usefulness
ratings. This interaction was misleading
among those who give these assessments very frequently (weekly) and work in a
forensic facility however it was descriptive for those who do not work in a
forensic facility.
There was not a
significant main effect of frequency of providing formal assessments of
“dangerousness” on probability of harm ratings F (2, 430) = .320, MSE=
1,694.271, p=.726, r=.039. This is misleading for those who do not work in a
forensic facility and had low usefulness ratings with those who give
assessments never/rarely giving higher probability of harm ratings than those
who give them weekly.
There was not a
significant main effect of whether a participant worked in a forensic facility
or not, F (1, 430) = .001, MSE= 1,694.271, p=.971, r=.002, on probability of
harm ratings. This effect is misleading
for those who had low usefulness ratings and gave these assessments weekly with
those who work in a forensic lab giving higher probability of harm ratings than
those who do not. It was also misleading
for those with high usefulness ratings and gave these types of assessments
weekly such that those who do not work in a forensic facility gave higher
probability of harm ratings than those who do.
There was not a
significant main effect of usefulness ratings of these types of assessments on
probability of harm ratings F (1, 430) = .006, MSE= 1,694.271, p=.937, r=.004.
This is misleading for those who give these types of assessments weekly and
work in a forensic facility, with those who had low usefulness ratings
reporting higher probability harm than those with high usefulness ratings. It
was misleading for those who give assessments weekly and do not work in a
forensic facility with the opposite occurring, those with low usefulness
ratings reporting lower probability of harm than those with high usefulness
ratings.
Probability of Rehospitalization
As hypothesized,
there was a significant three way interaction of whether or not someone worked
in a forensic facility, how useful they found these types of judgements for
making decisions (usefulness rating), and the frequency of providing formal
assessments of “dangerousness” on probability of rehospitalization
ratings, F (2, 436) = 3.625, MSE= 1.326, p=.0270, r=.129, refer to Table 3 and
Figure 2. As expected, there was not a lot of variability among ratings of rehospitalization. Pairwise comparisons revealed that for
those who work in a forensic facility, probability of rehospitalization
ratings were equivalent across the three levels of frequency whether the
respondent high or low usefulness ratings. Among those who do not work in a
forensic facility, probability of rehospitalization
ratings were also equivalent across the three levels of frequency but only if
the respondent had low usefulness ratings. For those who did not work in a
forensic facility and had high usefulness ratings, probability
of rehospitalization ratings were equivalent
for those who gave dangerousness assessments never/rarely to monthly. However
if they gave assessments weekly, respondents had higher probability of rehospitalization ratings than those who gave them never/rarely
(r=.304) or monthly (r=.229).
Among those who had
high usefulness ratings, probability of rehospitalization
ratings were equivalent between those who work in forensics and those who do
not across all three levels of frequency. Among those who had low usefulness
ratings, probability of rehospitalization ratings
were equivalent between those who worked in forensics and those who don’t only
if they gave assessment never/rarely and monthly. If respondents gave
assessments weekly, those who work in a forensic facility gave higher rehospitalization ratings than those who do not (r=.305).
Those who provide assessments of dangerousness
never/rarely and monthly gave equivalent ratings of rehospitalization
whether they worked in a forensic facility or not and had high and low useful
ratings. Among those who provide assessments weekly and worked in a forensic
facility, ratings of rehospitalization were
equivalent between those who had high and low useful ratings. However for those
who gave assessments weekly and did not work in a forensic lab, those who had
high usefulness ratings gave higher rehospitalization
than those with low usefulness ratings (r=.331).
There was not a
significant two way interaction of whether or not someone worked in a forensic
facility and how frequently they gave assessments of dangerousness as they
relate to ratings of rehospitalization, (F (2, 436) =
.294, MSE= 1.326, p=.745, r=.037). This was misleading for those who don’t work
in a forensic facility and had high usefulness ratings as those who gave
assessments weekly reported greater likelihood of rehospitalization
than those who gave them never/rarely and monthly. This interaction was also
misleading for those who had low usefulness ratings and gave these types of
assessments weekly with those who worked in a forensic facility reporting a
greater likelihood of rehospitalization.
There was not a
significant two way interaction of usefulness ratings and frequency as they
relate to ratings of rehospitalization, F (2, 436) =
.840, MSE= 1.326, p=.432, r=.062. This was misleading for those who don’t work
in a forensic facility and had high usefulness ratings as those who gave
assessments weekly reported greater likelihood of rehospitalization
than those who gave them never/rarely and monthly. This was misleading for
those who gave assessments weekly and do not work in a forensic facility, with
those who had high usefulness ratings reporting greater likelihood of rehospitalization than those with low usefulness ratings.
There was not a
significant two way interaction of usefulness ratings and whether or not
someone worked in a forensic facility, F (1, 436) = 2.603, MSE= 1.326, p=.107,
r=.077. This interaction was misleading for those who had low usefulness
ratings and gave these types of assessments weekly with those who worked in a
forensic facility reporting a greater likelihood of rehospitalization.
This interaction was also misleading for those who gave assessments weekly and
do not work in a forensic facility, with those who had high usefulness ratings
reporting greater likelihood of rehospitalization
than those with low usefulness ratings.
There was not a
significant main effect of frequency of providing formal assessments of
“dangerousness” on likelihood of rehospitalization
ratings F (2, 436) = 1.885, MSE= 1.326, p=.242, r=.093. This was misleading for
those who don’t work in a forensic facility and had high usefulness ratings
with those who gave assessments weekly reporting greater likelihood of rehospitalization than those who give them never/rarely and
monthly.
There was not a
significant main effect of whether a participant worked in a forensic facility
or not, F (1, 436) = 1.996, MSE= 1.326, p=.158, r=.068, on likelihood of rehospitalization ratings. This was misleading for those
who had low usefulness ratings and gave assessments weekly, with those who work
in a forensic facility giving higher ratings of rehospitalization
than those who do not.
There was not a
significant main effect of usefulness ratings of these types of assessments on
likelihood of rehospitalization ratings F (1, 436)
=.027, MSE= 1.326, p=.870, r=.008. This is misleading for those who give these
types of assessments weekly and do not work in a forensic lab, with those who
had high usefulness ratings reporting greater likelihood of rehospitalization
than those with low usefulness ratings.
Index Introduction Methods Discussion Table 1 Table 2 Table 3 Figure 1 Figure 2 References