Resources for 634 lecture
on Rates (Incidence)
(EPIB-634 Survival Analysis
and Related Topics)
Updated: January 27, 2020
Graphs Figures Tables
Articles, Chapters, ...
Software, Data, Simulations, ...
- Poisson Process Excel
Macro
- { 'unprotect' Excel sheets if you
wish to alter the layout .. under Tools menu -> Protection}
- Exact confidence limits on a Poisson parameter: Excel worksheet
- { 'unprotect' Excel sheets if you
wish to alter layout .. under Tools menu -> Protection}
- Perceived age as clinically useful
biomarker of ageing: cohort study
Article
Aggregated Data
R code
Stata .do file
Regression Models for Rates
- Sex- Age- and CalendarTime-specific Mortality Rates
Denmark
R code for plots & informal fit of a regression model
- Analysis of Count data using SAS (PROC GENMOD) and
Stata
-
- Rates_regressionR.txt [needlestick
injuries, Que. death rates, Uganda RCT, CH table 22.6 via R, SAS and Stata]
- quedata.txt
- CircumcisionHIV_Uganda.pdf
- ugandaTrial
Data.txt
- CircumcisionHIV_Kenya.pdf
- JUPITER dataset (in 10 PT segments)
- R code for fitting rate (ID)
functions to these data via Poisson regression.
.
- Condo Prices -- Montreal, 2005 and 2020
- 2005 prices of these condos
- [data+SAScode+Rcode]
- condoPricesDataR.txt [data, with header]
- 2020 prices of these condos
data
R code
- Analysis of IHD data in Table 22.6 of Clayton & Hills
* R code for additive and multiplicative models
- Québec mortality data , 1971, 1976, 1980-1992 2002
* Females(.txt) |
Males(.txt) |
Rcode & explanations(.txt) |
SAS(.txt)
-
LONGEVITY OF ACTORS/ACTRESSES WHO HAVE BEEN NOMINATED FOR AN OSCAR
original article (2001), re-analysis (2006), editorial, Response
CBC Radio Interview(2010)
Discussion of longevity begins 9 min. into Hour Three
(Debunking of Medical Myths).
New York Times Interview(2010)
Think the Answer's Clear? Look Again
Forbes.com(2011)
another 2011 blog
(a) dataset (1 record per performer, .csv format with headers)
nom1yr = year 1st nominated; win1yr = year 1st won (-1 if never won)
(b) individual time-slices dataset [with headers, and spaces as separators]
Lexis rectangles for time lived by performer as nominee, plus
Lexis rectangles for time lived by performer as winner.
20782 records
lex.id .... performer (1 - # performers, created by Lexis fn)
per ....... calendar year ('period')
age ....... age
lex.dur ... length of Lexis time-slice
lex.Xst ... status at exit from slice (1=dead, 0=alive)
id ........ performer id (assigned by Redelmeier)
born ...... year born
imale ..... 1 if male, 0 if female
nom1yr .... year first nominated
win1yr .... year first won (-1 if never won)
lastyr .... last year in study
dead ...... 1 if dead, 0 if alive, at LAST year
as.winner . 1 if these years were lived as a winner, 0 if not
start ..... year began the present state
winner.r .. 1 if EVER won, 0 if not (Redelmeier)
(c) aggregated performer-time (p-t) data
Lexis rectangles for each sex-age-period-Oscar pattern, where
Oscar = w.cat = 0 if time lived as nom'ee, 1 if as winner.
467 records
a.cat ....age-category(1=5-9, 2=10-14, 19=95-99)[floor(age/5)]
p.cat ....period (category) (2=1920- 9=1990- 10=2000-)
m.cat ....male? 1=yes 0=no
w.cat ....winner? time lived as winner? 1=yes 0=no
lex.Xst ..no. persons who exited Lexis rectangle by death
lex.dur ..no. performer-years lived in this "cell"
(d) aggregated performer-time (p-t) data, SOME-INCORRECTLY-ALLOCATED
Lexis rectangles for each sex-age-period-Oscar pattern, where
Oscar = 0 if performer never won, 1 if performer ever won Oscar.
Variables as defined in aggregated performer-time (p-t) data, EXCEPT...
467 records
w.cat .... winner? Did performer EVER win? 1=yes 0=no
(***) Data Analysis - some R code
* Software and R code for creating Person-Years datasets
* Links to material re "Immortal-time" bias
|