> xwzzakW7{y3T(
B(
/00DArialngs`C0
0DVerdanas`C0
0" DTimes New Roman0
00DWingdingsRoman0
0
A.
@n?" dd@ @@``<!
"#%&)+,-./01245678;,r$kW7{y3Ti0AA 333@8ʚ;ʚ;g4MdMd0
0ppp@<4dddd\qt
08C<4BdBd\p
0B0___PPT10
___PPT9"(*S:3+6Teaching Analytic Methods in Maternal and Child Health77
7ATMCH C.E. Institute
Teaching Maternal and Child Health Competencies: Tools and Techniques for Graduate Programs
November 16, 2003
Deborah Rosenberg, PhD
Division of Epidemiology and Biostatistics
School of Public Health
University of Illinois at Chicago
!Teaching Analytic MethodsWhat analytic methods do MCH professionals need to understand?
What skills in using analytic methods do MCH professionals need to have?
Do all MCH professionals need to know and be skilled in analytic methods?Questions from the Field$b We are constructing an index of MCH need ... we have nearly 22 indicators and I am ranking each by neighborhood ... can I sum the ranks and create a sum of ranks variable or do i need to scale these on a similar metric (e.g. create z scores) and then sum these? ... the distributions of each vary ... some normal some not. and the indicators cover a wide range of domains ...
MPH in Epidemiology,
working for a health department2{n6Questions from the Field$v
... are you familiar or aware of any method whereby the zipcode data can be transformed or manipulated to approximate 90-95% of the population at higher levels of aggregation?are there any rules or threshholds for acceptable 'overlaps' of populations so to speak when using zipcode data?
PhD in MCH,
consultant>#n<>:A Questions from the Field$N There are several indicators that we apply a three-year moving average to in order to 'stabilize' the rate ... when computing the standard error, is the n I use the averaged population over the three-year period or the sum of the population? ... Also, is it okay to compare state data for 98-00 to US data for 2000? When is a small population too small to use one year of data?
MS in Biostatistics,
working for a state&n)Questions from the Academy $
In the low birth weight example you covered in class, you created only two dummy variables, "teen" and "plus35 (for age). Does this mean the 20-34 age group is the reference group? Why?
MPH student$ Questions from the Academy$~
I m doing a study of subtypes of preterm delivery ... do I need to run separate models for each subtype and a comparison group or can I somehow consider all subtypes jointly?
PhD student$
Questions from the Academy$
I am putting several variables in a model as sets of dummy variables. This means when I look for effect modification I have many beta coefficients to consider. How do I interpret these, especially if some are significant and some aren t?
PhD student,Questions from the Academy$
I was looking at severe preeclamptic women ... For some patients there is more than one preventability factor present ... women were given credit for each preventable factor they had. Is it possible to do comparative statistics, i.e. something, like a chi square even though some individuals were double counted??
Physician researcher $AW1 -Why Are We Teaching Analytic Methods Anyway?.."Some typical compartmentalizations:
Graduate Degree Program
Continuing education
Technical Assistance
Research
Surveillance
Assessment$B ! 33$B
g$Teaching Analytic Methods
Graduate course Graduate Seminar
Continuing
Certificate Education
Program Workshop
Technical Thesis project
Assistance 6And What Analytic Methods are We Talking About Anyway?77$Some typical compartmentalizations:
Statistical methods
Epidemiologic methods
Methods for health services research
Psychometrics
Econometrics
Other
Qualitative Methodsp$q" " $p
"Teaching Analytic MethodsReconciling Disciplinary Jargon
Synthesizing Disciplinary Approaches
a few examples
ANOVA ... regression analysis
statistical bias ... study bias ... validity
difference measures ... ratio measures
computational formulas ... notationRVoETeaching Analytic MethodslAnd then there s format and level:
Basic Advanced
Lecture hands-on
Paper/pencil Computer
Face to face Distance
Clean data Messy data
Group Work Individual work ,##4What Is Hands-On Anyway?
Specifying a conceptual framework
Specifying hypotheses
Writing computer code
Drawing table shells
Designing graphs
Writing an analysis plan
Identifying appropriate statistical tests." ("Teaching Analytic MethodsFor hands-on data analysis, what data should we use?
Historically and currently, analytic methods are often taught using datasets with few observations and few variables
pedagogical rationale
technological constraints ^8v1" 8b0)#Dataset OptionsEach option has pros and cons:
*$Teaching Analytic Methods
Small datasets keep the focus narrow, uncluttered by either data management or substantive issues
Work best for first encounters with specific methods, both basic and advanced+%Teaching Analytic Methods
Large datasets can be used with a narrow or broad focus
Works for teaching complete data analysis, integrating substantive, methodological, data management, and software issues
Teaching Analytic MethodsSIs This Graduate Education?
Data organization
Data management
Software mechanicsT" " 7" 7Teaching Analytic MethodsData Management / Software Mechanics
accessing and inputting data
linking / merging / concatenating data
subsetting data
moving files between software
organizing data for tabling
organizing data for graphing,%" " j
]Teaching Analytic MethodsOrganizing Data for Analysis
recoding variables
creating new variables
defining reference groups
defining comparison groups
handling missing values
choosing appropriate statistical procedures:" -&Teaching Analytic MethodsTo really internalize the analytic process:
Conceptual overview
Methodological overview
Demonstration of method
Bits of analysis
Z,U,U.(Teaching Analytic MethodsTo really internalize the analytic process, continued:
Methodological detail
Full analysis
Presentation of results
Interpretation of results
Discussion and reanalysisV7p, p/)Teaching Analytic MethodsTo really use the analysis:
What s the SO WHAT ?
Beyond interpretation, what are the recommendations for action?"Wt0'Teaching Analytic Methodso
Assignments can be designed in terms of grant, report, or manuscript development
methods
results
discussion
hQRSummaryCArtificial lines drawn between statistical, epidemiologic, and other methods
Artificial lines drawn between data management, software mechanics, and conducting analysis
Artificial lines drawn between theory and application
Artificial lines drawn between conducting analysis and presentation and interpretation of resultsDD%Summary
To the extent feasible, design courses that synthesize across disciplinary and pedagogical boundaries
Teach analytic methods in the context of the MCH planning cycle
1*Summary
Design MCH-specific exercises and data analyses
Textbooks aren t enough!! (and hardly exist in a format that is sufficient)~~'!Summary
Acknowledge the time and intensity required for teaching analytic methods
Acknowledge how teaching analytic methods might change other aspects of the MCH curriculum& Summary
But everyone can t do everything!
Let s just give a
consulting contract
to a statistician"]\$ ` !3̙` Q.<ffff3` 3333fff` 3K=̙fff` 3fffff` ff3ff3` aNR>ff` 3fY33` 3f3f>?" dd@&?oAd(@n<)o<6=nA+7%Z', n?" dd@ @@``PR = 7 ,`(p>>P0(
<ܕ #"`j
T Click to edit Master title style!
!$
0t "P`
RClick to edit Master text styles
Second level
Third level
Fourth level
Fifth level!
SpB
#BCE6FGIIQU*VWX@`@$ G"`
B0
`B
s*Dg"
dA"h ?Light horizontal"` Q.<ffff3___PPT10i.
"+D=' =
@B + Profile `C(
<5 #"`@_
T Click to edit Master title style!
!
07 "p`_
W#Click to edit Master subtitle style$
$
0< "``
\*
0A "`
^*
0$E "`
^*pB
#@IBCE6FGjIQU*VWX@`@$ G"p
B0
dA"h ?Light horizontal"` Q.<ffff3___PPT10i.
"+D=' =
@B +00(
x
c$,݊
x
c$ _
H
0h ? Q.<ffff380___PPT10.t>ˣ$
`$(
r
S\_j
_
r
S_P`_
H
0h ? Q.<ffff380___PPT10.@I
N(
x
c$Vj
c$]P`
*{h;@H
0h ? 33___PPT10i.I+D=' =
@B +
B(
r
SȽj
SǆP`
*ZGH
0h ? Q.<ffff3___PPT10i.ĨPk+D=' =
@B +
$B(
$r
$ S<ȧj
$ SP`
*{h;@H
$0h ? Q.<ffff3___PPT10i.ƨ+D=' =
@B +
(B(
(r
( Sj
( SeP`
*~k>CH
(0h ? Q.<ffff3___PPT10i.tQG+D=' =
@B +B
0B(
0r
0 Sdj
0 SgP`
*~k>CH
00h ? Q.<ffff380___PPT10.!B
4B(
4r
4 St_j
_
4 S_P`_
*~k>CH
40h ? Q.<ffff380___PPT10."4qB
,B(
,r
, S(j
, SdP`
*'^cH
,0h ? Q.<ffff380___PPT10.I0
P0(
x
c$dj
d
x
c$
dP`d
H
0h ? Q.<ffff380___PPT10.3Pڄ
VN(
x
c$dj
d
x
c$dP`d
2
<0d"``S
TMultiple
areas,
over
time
2
Bd"``
t"Special
topics,
compressed
time
6"
"2
B("d"`` c
\
Individually
Customized&
H
0h ? Q.<ffff380___PPT10.5P970
x0(
xx
x c$j
x
x c$؉P`
H
x0h ? Q.<ffff380___PPT10.# eB
B(
x
c$hDdj
d
c$0KdP`d
8haQH
0h ? Q.<ffff380___PPT10.D
p(
pr
p S4Tdj
d
p S^dP`d
^B
p
6DԔ^B
p
6DԔPP^B
p
6D3fԔpp^B
p
6D3Ԕ ^B
p
6D̙Ԕ00^B
p
6DԔ
H
p0h ? Q.<ffff380___PPT10.:k$
t$(
tr
t S0cdj
d
r
t S$ddP`d
H
t0h ? Q.<ffff380___PPT10.0N
N(
x
c$0ndj
d
c$tdP`d
*~k>CH
0h ? Q.<ffff380___PPT10./ W'
,$(
x
c$|dj
d
0}dP`d
x
r #
##")4#
6(d?tC#
@Small Studies /
Subset Observations and Variables
Textbook DataAA)=7,
6d?Ct#
\
Not Useful
)=7,S
<d?C#VNF___PPT9(
Few Observations6 " d# @@``
6,d?tC
b
Subset Variables
)=7,
6Pd?tC
v&e.g.
Vital Records
PRAMS
CSHCN Survey'')=7,T
<ܸd?CVNF___PPT9(
Many Observations6 " d# @@``Q
<d?tVNF___PPT9(
Few Variables8(" d
# @@``R
<d?tVNF___PPT9(
Many Variables8(" d# @@``
<8d?
T
)=7,ZB
s*Ԕ
?#ZB
s*Ԕ
?tt#ZB
s*Ԕ
?ZB
s*Ԕ
?CC`B
0Ԕ
?`B
0Ԕ
?#`B
0Ԕ
?#`B
0Ԕ
?##H
0h ? Q.<ffff380___PPT10.'@80
0(
x
c$dj
d
x
c$LdP`d
H
0h ? Q.<ffff380___PPT10./ 0
0(
x
c$|dj
d
x
c$c$
<$(
<r
< SXdj
d
r
< S,dP`d
H
<0h ? Q.<ffff380___PPT10.$ }$
$(
r
Sdj
d
r
SdP`d
H
0h ? Q.<ffff380___PPT10.04$
0$(
r
S,ej
e
r
SeP`e
H
0h ? Q.<ffff380___PPT10.PN
@N(
x
c$ej
e
c$eP`e
*~k>CH
0h ? Q.<ffff380___PPT10.\F7$
`$(
r
Sej
e
r
SeP`e
H
0h ? Q.<ffff380___PPT10.2$
p$(
r
Sd#ej
e
r
S8$eP`e
H
0h ? Q.<ffff380___PPT10.0,N
PN(
x
c$D[_j
_
c$@_P`_
*~kH
0h ? Q.<ffff380___PPT10.2xB
|B(
|r
| S*ej
e
| S(x_P`e
*~k>CH
|0h ? Q.<ffff380___PPT10.0T]$
$(
r
S3ej
e
r
S6eP`e
H
0h ? Q.<ffff380___PPT10.N$
$(
r
S;ej
e
r
S"n;Questions from the Field$N There are several indicators that we apply a three-year moving average to in order to 'stabilize' the rate ... when computing the standard error, is the n I use the averaged population over the three-year period or the sum of the population? ... Also, is it okay to compare state data for 98-00 to US data for 2000? When is a small population too small to use one year of data?
MS in Biostatistics,
working for a state&n)Questions from the Academy $
In the low birth weight example you covered in class, you created only two dummy variables, "teen" and "plus35 (for age). Does this mean the 20-34 age group is the reference group? Why?
MPH student$ Questions from the Academy$~
I m doing a study of subtypes of preterm delivery ... do I need to run separate models for each subtype and a comparison group or can I somehow consider all subtypes jointly?
PhD student$
Questions from the Academy$
I am putting several variables in a model as sets of dummy variables. This means when I look for effect modification I have many beta coefficients to consider. How do I interpret these, especially if some are significant and some aren t?
PhD student,Questions from the Academy$
I was looking at severe preeclamptic women ... For some patients there is more than one preventability factor present ... women were given credit for each preventable factor they had. Is it possible to do comparative statistics, i.e. something, like a chi square even though some individuals were double counted??
Physician researcher $AW -Why Are We Teaching Analytic Methods Anyway?.."Some typical compartmentalizations:
Graduate Degree Program
Continuing education
Technical Assistance
Research
Surveillance
Assessment$B ! 33$B $Teaching Analytic Methods
Graduate course Graduate Seminar
Continuing
Certificate Education
Program Workshop
Technical Thesis project
Assistance 6And What Analytic Methods are We Talking About Anyway?77$Some typical compartmentalizations:
Statistical methods
Epidemiologic methods
Methods for health services research
Psychometrics
Econometrics
Other
Qualitative Methodsp$q" " $p"Teaching Analytic MethodsReconciling Disciplinary Jargon
Synthesizing Disciplinary Approaches
a few examples
ANOVA ... regression analysis
statistical bias ... study bias ... validity
difference measures ... ratio measures
computational formulas ... notationRVoETeaching Analytic MethodslAnd then there s format and level:
Basic Advanced
Lecture hands-on
Paper/pencil Computer
Face to face Distance
Clean data Messy data
Group Work Individual work ,##4What Is Hands-On Anyway?
Specifying a conceptual framework
Specifying hypotheses
Writing computer code
Drawing table shells
Designing graphs
Writing an analysis plan
Identifying appropriate statistical tests." ("Teaching Analytic MethodsFor hands-on data analysis, what data should we u
!"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^{`acdefghijklmnopqrstuvy|}~Root EntrydO)N@vPicturesCurrent UserF)SummaryInformation(PowerPoint Document(ADocumentSummaryInformation80xb_00DArialngsll`Cԥ0
0DVerdanasll`Cԥ0
0" DTimes New Romanԥ0
00DWingdingsRomanԥ0
0
A.
@n?" dd@ @@``<!
"#%&)+,-./01245678;,r$kW7{y3Ti0AA 333@8ʚ;ʚ;g4MdMd
0ppp@<4dddd0qt
0l8C<4BdBd0p
0lB0___PPT10
___PPT9"(*Q:3+6Teaching Analytic Methods in Maternal and Child Health77
7ATMCH C.E. Institute
Teaching Maternal and Child Health Competencies: Tools and Techniques for Graduate Programs
November 16, 2003
Deborah Rosenberg, PhD
Division of Epidemiology and Biostatistics
School of Public Health
University of Illinois at Chicago
!Teaching Analytic MethodsWhat analytic methods do MCH professionals need to understand?
What skills in using analytic methods do MCH professionals need to have?
Do all MCH professionals need to know and be skilled in analytic methods?Questions from the Field$b We are constructing an index of MCH need ... we have nearly 22 indicators and I am ranking each by neighborhood ... can I sum the ranks and create a sum of ranks variable or do i need to scale these on a similar metric (e.g. create z scores) and then sum these? ... the distributions of each vary ... some normal some not. and the indicators cover a wide range of domains ...
MPH in Epidemiology,
working for a health department2{n6Questions from the Field$t
... are you familiar or aware of any method whereby the zipcode data can be transformed or manipulated to approximate 90-95% of the population at higher levels of aggregation?are there any rules or thresholds for acceptable 'overlaps' of populations so to speak when using zipcode data?
PhD in MCH,
consultant>"n;>:
A Questions from the Field$N There are several indicators that we apply a three-year moving average to in order to 'stabilize' the rate ... when computing the standard error, is the n I use the averaged population over the three-year period or the sum of the population? ... Also, is it okay to compare state data for 98-00 to US data for 2000? When is a small population too small to use one year of data?
MS in Biostatistics,
working for a state&n)Questions from the Academy $
In the low birth weight example you covered in class, you created only two dummy variables, "teen" and "plus35 (for age). Does this mean the 20-34 age group is the reference group? Why?
MPH student$ Questions from the Academy$~
I m doing a study of subtypes of preterm delivery ... do I need to run separate models for each subtype and a comparison group or can I somehow consider all subtypes jointly?
PhD student$
Questions from the Academy$
I am putting several variables in a model as sets of dummy variables. This means when I look for effect modification I have many beta coefficients to consider. How do I interpret these, especially if some are significant and some aren t?
PhD student,Questions from the Academy$
I was looking at severe preeclamptic women ... For some patients there is more than one preventability factor present ... women were given credit for each preventable factor they had. Is it possible to do comparative statistics, i.e. something, like a chi square even though some individuals were double counted??
Physician researcher $AW1 -Why Are We Teaching Analytic Methods Anyway?.."Some typical compartmentalizations:
Graduate Degree Program
Continuing education
Technical Assistance
Research
Surveillance
Assessment$B ! 33$B
g$Teaching Analytic Methods
Graduate course Graduate Seminar
Continuing
Certificate Education
Program Workshop
Technical Thesis project
Assistance 6And What Analytic Methods are We Talking About Anyway?77$Some typical compartmentalizations:
Statistical methods
Epidemiologic methods
Methods for health services research
Psychometrics
Econometrics
Other
Qualitative Methodsp$q" " $p
"Teaching Analytic MethodsReconciling Disciplinary Jargon
Synthesizing Disciplinary Approaches
a few examples
ANOVA ... regression analysis
statistical bias ... study bias ... validity
difference measures ... ratio measures
computational formulas ... notationRVoETeaching Analytic MethodslAnd then there s format and level:
Basic Advanced
Lecture hands-on
Paper/pencil Computer
Face to face Distance
Clean data Messy data
Group Work Individual work ,##4What Is Hands-On Anyway?
Specifying a conceptual framework
Specifying hypotheses
Writing computer code
Drawing table shells
Designing graphs
Writing an analysis plan
Identifying appropriate statistical tests." ("Teaching Analytic MethodsFor hands-on data analysis, what data should we use?
Historically and currently, analytic methods are often taught using datasets with few observations and few variables
pedagogical rationale
technological constraints ^8v1" 8b0)#Dataset OptionsEach option has pros and cons:
*$Teaching Analytic Methods
Small datasets keep the focus narrow, uncluttered by either data management or substantive issues
Work best for first encounters with specific methods, both basic and advanced+%Teaching Analytic Methods
Large datasets can be used with a narrow or broad focus
Works for teaching complete data analysis, integrating substantive, methodological, data management, and software issues
Teaching Analytic MethodsSIs This Graduate Education?
Data organization
Data management
Software mechanicsT" " 7" 7Teaching Analytic MethodsData Management / Software Mechanics
accessing and inputting data
linking / merging / concatenating data
subsetting data
moving files between software
organizing data for tabling
organizing data for graphing,%" " j
]Teaching Analytic MethodsOrganizing Data for Analysis
recoding variables
creating new variables
defining reference groups
defining comparison groups
handling missing values
choosing appropriate statistical procedures:" -&Teaching Analytic MethodsTo really internalize the analytic process:
Conceptual overview
Methodological overview
Demonstration of method
Bits of analysis
Z,U,U.(Teaching Analytic MethodsTo really internalize the analytic process, continued:
Methodological detail
Full analysis
Presentation of results
Interpretation of results
Discussion and reanalysisV7p, p/)Teaching Analytic MethodsTo really use the analysis:
What s the SO WHAT ?
Beyond interpretation, what are the recommendations for action?"Wt0'Teaching Analytic Methodso
Assignments can be designed in terms of grant, report, or manuscript development
methods
results
discussion
hQRSummaryCArtificial lines drawn between statistical, epidemiologic, and other methods
Artificial lines drawn between data management, software mechanics, and conducting analysis
Artificial lines drawn between theory and application
Artificial lines drawn between conducting analysis and presentation and interpretation of resultsDD%Summary
To the extent feasible, design courses that synthesize across disciplinary and pedagogical boundaries
Teach analytic methods in the context of the MCH planning cycle
1*Summary
Design MCH-specific exercises and data analyses
Textbooks aren t enough!! (and hardly exist in a format that is sufficient)~~'!Summary
Acknowledge the time and intensity required for teaching analytic methods
Acknowledge how teaching analytic methods might change other aspects of the MCH curriculum& Summary
But everyone can t do everything!
Let s just give a
consulting contract
to a statistician"]\$
B(
r
SȽj
SǆP`
*ZGH
0h ? Q.<ffff3___PPT10i.ĨPk+D=' =
@B +r
p3A(
/00DArialngsll`Cԥ0
0DVerdanasll`Cԥ0
0" DTimes New Romanԥ0
00DWingdi
!"#$%&'()*+,-./123456789:;<=>?@ABCDEOh+'0hp
Slide 1Deborah RosenbergeProfileUIC29Microsoft PowerPointP@N@`mS¨@>vG
o
>Zj-B( .Q<--@ !i--'ff--@ !A--ff--%A'--'@"Verdana-. 32
Teaching Analytic Methods in
!
. "System-@"Verdana-. -2
Maternal and Child Health !.-@"Verdana-. %2
@ATMCH C.E. Institute
.-@"Verdana-. Z2
7|7Teaching Maternal and Child Health Competencies: Tools i
.-@"Verdana-. =2
R$and Techniques for Graduate Programs
.-@"Verdana-. !2
FNovember 16, 2003u
.-@"Verdana-. (2
*Deborah Rosenberg, PhD
.-@"Verdana-. F2
*Division of Epidemiology and Biostatistics
.-@"Verdana-. *2
6School of Public Health
.-@"Verdana-. 92
.
!University of Illinois at Chicagoa
.-se?
Historically and currently, analytic methods are often taught using datasets with few observations and few variables
pedagogical rationale
technological constraints ^8v1" 8b0)#Dataset OptionsEach option has pros and cons:
*$Teaching Analytic Methods
Small datasets keep the focus narrow, uncluttered by either data management or substantive issues
Work best for first encounters with specific methods, both basic and advanced+%Teaching Analytic Methods
Large datasets can be used with a narrow or broad focus
Works for teaching complete data analysis, integrating substantive, methodological, data management, and software issues
Teaching Analytic MethodsSIs This Graduate Education?
Data organization
Data management
Software mechanicsT" " 7" 7Teaching Analytic MethodsData Management / Software Mechanics
accessing and inputting data
linking / merging / concatenating data
subsetting data
moving files between software
organizing data for tabling
organizing data for graphing,%" " Teaching Analytic MethodsOrganizing Data for Analysis
recoding variables
creating new variables
defining reference groups
defining comparison groups
handling missing values
choosing appropriate statistical procedures:" -&Teaching Analytic MethodsTo really internalize the analytic process:
Conceptual overview
Methodological overview
Demonstration of method
Bits of analysis
Z,U,U.(Teaching Analytic MethodsTo really internalize the analytic process, continued:
Methodological detail
Full analysis
Presentation of results
Interpretation of results
Discussion and reanalysisV7p, p/)Teaching Analytic MethodsTo really use the analysis:
What s the SO WHAT ?
Beyond interpretation, what are the recommendations for action?"Wt0'Teaching Analytic Methodso
Assignments can be designed in terms of grant, report, or manuscript development
methods
results
discussion
hQRSummaryCArtificial lines drawn between statistical, epidemiologic, and other methods
Artificial lines drawn between data management, software mechanics, and conducting analysis
Artificial lines drawn between theory and application
Artificial lines drawn between conducting analysis and presentation and interpretation of resultsDD%Summary
To the extent feasible, design courses that synthesize across disciplinary and pedagogical boundaries
Teach analytic methods in the context of the MCH planning cycle
1*Summary
Design MCH-specific exercises and data analyses
Textbooks aren t enough!! (and hardly exist in a format that is sufficient)~~'!Summary
Acknowledge the time and intensity required for teaching analytic methods
Acknowledge how teaching analytic methods might change other aspects of the MCH curriculum& Summary
But everyone can t do everything!
Let s just give a
consulting contract
to a statistician"]\$00(
x
c$,݊
x
c$ _
H
0h ? Q.<ffff380___PPT10.t>ˣ
B(
r
SȽj
SǆP`
*ZGH
0h ? Q.<ffff3___PPT10i.ĨPk+D=' =
@B +B
,B(
,r
, S(j
, SdP`
*'^cH
,0h ? Q.<ffff380___PPT10.I0
P0(
x
c$dj
d
x
c$
dP`d
H
0h ? Q.<ffff380___PPT10.3Pڄ0
x0(
xx
x c$j
x
x c$؉P`
H
x0h ? Q.<ffff380___PPT10.# e$
$(
r
Sdj
d
r
SdP`d
H
0h ? Q.<ffff380___PPT10.04r868'=_? :33+
A3