Honors STAT 1000
Bone marrow transplantation is one of the
most common treatments for acute
The purpose of this study was to consider the
effects of bone marrow transplantation with a
radiation-free conditioning regimen.
Patients were given different drugs in two groups at
four different hospitals (two in US and two in
137 patients were studied (97 in the US and 40 in
Australia) from 1984 to 1989.
This study was….
a) Retrospective observational study
b) Prospective observational study
c – This was an experiment because the experimenters controlled
different variables (specifically, the types of drugs given) in order to
see their effect on recovery from or relapse into acute leukemia. The
study was two sample since the study looked at the effects of two
independent drug treatments.
The experimenters grouped individuals into
many different risk categories.
Donor age and recipient age was one of the
Since we start with a recipient who needs a
The explanatory variable is recipient age
The response variable is donor age.
S = 6.55008
The regression equation is
R-Sq = 54.0%
Recipient = 8.89 + 0.684 Recipients
r = .734846 (fairly strong + relationship)
Positive slope of .790 means that for every
additional year in the age of the donor, we
expect the age of the recipient to go up
around 9.6 months.
Fairly strong positive correlation supports
None of the recipients were less than 1 year
Intercept is of no interpretive value since it
deals with a recipient age of 0
The s value of 6.55008 meant that ages
predicted by the regression line tended to be
off on average by around 6.5 years from the
This difference makes sense when looking at
the r value of the correlation r = .734846
(fairly strong + relationship)
Why are ages positively correlated?
Bone marrow transplantation need exact
matches of Human leukocyte antigen
Highest percentage of matches found
Parent donation vs. Sibling donation
Was this study reported from a sample or population?
The experimenters tried to categorize risk
factors prior to transplantation in order to
reduce confounding variables.
Objective determinations such as “death”
and “remission” do not involve bias.