2024
Two approaches to approximating sums of chisquareds
2024/09/13
National Land Transport Plan graphs
2024/09/03
Ihaka Lectures
2024/08/29
The missing test in survey regression models
2024/08/27
Another way to not sample without replacement
2024/08/26
A Bayesian t-test, again
2024/07/16
Stage vs phase
2024/06/28
Estimator vs estimate
2024/06/27
Automatic transformation of standard errors?
2024/06/15
S3 method dispatch on other arguments
2024/06/04
Crossvalidation in complex survey data
2024/05/21
Choosing frame weights in dual-frame surveys
2024/05/10
Another update on non-transitive dice
2024/04/29
Multiple frame sampling
2024/04/26
Importance weights
2024/04/19
Assumptions
2024/04/14
Quantitative graphics?
2024/04/05
Symbolically nested
2024/04/01
Factors as factors
2024/03/22
Small-area estimates by smoothing direct estimates
2024/03/09
New in the survey package
2024/03/08
Ordinal outcomes: the LOCT DOOR
2024/01/25
Recurrent events: increased susceptibility or latent risk?
2024/01/12
Asymptotics for linear mixed models
2024/01/09
2023
Why do the Rao-Scott tests have good size?
2023/12/18
How good is the leading eigenvalue approximation to quadratic forms?
2023/12/14
Why not REML?
2023/12/12
Sparse correlation and the Central Limit Theorem
2023/12/04
svy2lme: the preprint
2023/11/24
Linear mixed models with pairwise likelihood
2023/11/21
Benchmark Archaeology
2023/08/24
Quoting and requoting
2023/08/04
Blank-cheque inheritance and statistical methods objects
2023/06/07
Pairwise likelihood and cluster sizes
2023/05/05
New in the survey package
2023/05/04
Ranks in survey data
2023/04/18
Class imbalance: bug or feature?
2023/03/27
Which infinite sequence?
2023/03/20
The fourth-root thing
2023/03/07
Determinant of correlation matrix
2023/03/06
Sandwiches and aggregation
2023/02/21
When is population mean rank a thing?
2023/02/10
Checking proportionality of odds
2023/02/09
Linkage and multiple imputation
2023/01/06
2022
Pairwise and joint independence
2022/12/09
A short note on effect sizes
2022/12/06
The sandwich and the t-test
2022/12/01
Bus pruning
2022/11/05
Improving a graph
2022/11/03
Code archaeology: polynomial distributed lags
2022/10/14
A plug-in uniform law of large numbers
2022/09/28
Looking back
2022/09/15
Tracking down a Real Data Set(tm)
2022/08/14
ASCII and beyond — will it play in Peoria?
2022/07/26
Tidying rimu
2022/07/23
Combining a survey and other data
2022/07/15
Self-promotion: an actual multiwave two-phase design
2022/07/06
Getting strings into code in base R
2022/06/23
Design |> Data: Ihaka Lectures 2022
2022/06/21
Tables with zeroes
2022/04/17
stringsAsFactors=do_you_feel_lucky
2022/03/31
Nine and sixty ways
2022/02/20
Comparing tests for generalised linear models in survey data
2022/01/28
Optimal design for raking/AIPW estimation
2022/01/06
Per capita, in mice
2022/01/02
2021
Top posts from 2021
2021/12/31
Is it binary?
2021/11/10
Crossed clustering and parallel invention
2021/09/18
Score tests: surprisingly annoying
2021/09/10
Ordinal data, metadata, and models
2021/09/03
Pictures of code are not code
2021/08/21
Wellington buses
2021/08/13
The New Oil
2021/08/07
Maintenance of Headway
2021/08/04
Subsets and subpopulations in survey inference
2021/07/22
What's new in the survey package
2021/07/20
Not all strictly monotone functions are additive
2021/07/16
Housing unaffordability hexmaps
2021/06/11
Generalisability, prediction, and causation
2021/05/02
A modest proposal for matrix multiplication
2021/04/01
Phobos and Deimos and public speaking
2021/03/19
Two-phase sampling notation
2021/02/15
Co-linearity
2021/02/11
They're back!
2021/02/02
2020
Top posts in 2020
2020/12/31
Emma Lathen e-books
2020/12/31
Planning a new data management course
2020/12/18
Neyman Allocation, only exact
2020/11/05
You will probably not be eaten by a grue
2020/10/30
When the sky didn't fall
2020/09/26
MOAR survey regression models
2020/09/24
A Bayesian t-test?
2020/09/17
Weights in statistics
2020/08/04
Sourdough happens
2020/06/23
New in the survey package
2020/04/03
Changing strata mid-stream
2020/03/27
Mapping NZ cases of COVID-19
2020/03/26
Not cross buns
2020/03/23
Quadratic trend tests in survey package
2020/02/28
The Ihaka Lectures, Episode 4
2020/02/19
Survey package news
2020/01/22
Multifactor interventions and interactions
2020/01/09
2019
Computer says no
2019/12/30
What is 'Data Science Practice'?
2019/12/25
How many giraffes?
2019/12/01
Hexmaps for NZ District Health Boards
2019/11/07
Some things I don’t like about the Oxford-Munich Code of Conduct
2019/10/01
How to review a book
2019/09/13
Why isn't rimu tidy?
2019/09/10
(What’s up with the brackets?)
2019/09/10
A package for multiple-response data
2019/09/05
Adding new functions to the survey package
2019/07/16
Denominator degrees of freedom in svyglm
2019/06/26
Wald, score, LRT: the picture
2019/06/20
Analysing the mouse microbiome autism data
2019/06/16
Confidence intervals: not a very strong property
2019/06/11
Design degrees of freedom: brief note
2019/06/08
Mean People Tweet
2019/05/24
The Reeferendum
2019/05/07
Local asymptotic minimax, and nearly-true models
2019/04/30
Survey package update
2019/04/28
That’s for remembrance
2019/04/24
Handling ‘plausible values’ in surveys
2019/04/21
Progress on linear mixed models for surveys
2019/04/19
Hypergraph network meta-analysis
2019/03/26
The school climate strike
2019/03/12
Normal horizontiles
2019/03/04
Displaying bus punctuality
2019/03/01
Absolutely no warranty?
2019/02/18
What have I got against the Shapiro-Wilk test?
2019/02/09
How do you tell what packages to trust?
2019/02/04
Recognising when you don’t know
2019/02/01
Two quick survey items
2019/01/26
Another way to see why mixed models in survey data are hard:
2019/01/18
The Ihaka Lectures 3: Rise of the Machine Learners
2019/01/11
Bayesian Surprise — the Shiny app
2019/01/04
2018
What are packages for?
2018/12/17
svycontrast
2018/12/10
Finding principal components without even looking?
2018/11/26
Come work with us
2018/11/04
Progress on svy2lme
2018/10/19
Survey package update
2018/10/12
The Kiwi PRNG
2018/10/04
How to write a racist AI in R without really trying
2018/09/27
Journalism and cyber-bullying
2018/09/11
What can data science add to statistics education?
2018/08/28
ISCB/ASC talk
2018/08/26
Leaflet and buses
2018/08/14
Testing probability distribution generators
2018/08/01
Quoting and macros in R
2018/07/30
e-bike: the reboot
2018/07/17
Interlingual
2018/07/11
Spell my name with a ‘v'
2018/06/24
Survey analysis in SQL
2018/06/09
Statistical software matters
2018/06/09
New blog home
2018/06/05
Graduation
2018/05/14
svylme
2018/04/01
Small p hacking
2018/03/23
Chebyshev’s inequality and `UCL’
2018/03/15
Why pairwise likelihood?
2018/03/13
Faster generalised linear models in largeish data
2018/03/05
Useful debugging trick
2018/01/31
The Ihaka Lectures
2018/01/22
More tests for survey data
2018/01/22
As far as it goes
2018/01/20
breakInNamespace
2018/01/15
2017
e-bike-onomics
2017/12/30
Statistics on pairs
2017/12/26
How to add chi-squareds
2017/12/06
Secret Santa collisions
2017/11/25
When all U-shaped curves look the same to you
2017/11/23
Means of maximums
2017/11/08
Haere mai, statistical computing folks
2017/09/26
A genome analogy
2017/09/25
Bayesian surprise
2017/09/22
Visual design of diagnostics
2017/09/06
Causes and counterfactuals
2017/08/23
Wilcoxon and polymath: another update
2017/08/19
The bus bot
2017/08/10
Tail bounds under sparse correlation
2017/07/26
Psychoactive substances and Peter Dunne
2017/07/26
Information and control
2017/07/25
Probabilities not bounded away from zero
2017/07/09
Two-day course: survival analysis
2017/07/05
A possibly unsurprising bootstrap observation
2017/06/11
Stupid word games
2017/06/05
Pipeable survey analysis in R
2017/05/29
Peer review and community endorsement
2017/05/22
A ‘polymath’ project on the Wilcoxon test?
2017/05/12
Value of a degree
2017/05/01
Prerequisites
2017/03/29
Come work with us
2017/03/28
Why I like the Convolution Theorem
2017/03/27
Flat Earthers
2017/03/27
Case-control efficiency
2017/03/18
Order and quotient topologies
2017/03/14
“Meritocracy” and “public good”
2017/03/11
Hearing things
2017/03/05
The Ihaka Lectures
2017/02/02
When the bootstrap doesn’t work
2017/02/01
Te Reo Māori in schools
2017/01/31
Case-control sampling and pseudo-Rsquareds
2017/01/27
A bus-watching bot
2017/01/17
Mature and premature optimisation
2017/01/12
Fixing an infelicity in ‘leaps’
2017/01/09
Learning the Monty Hall problem
2017/01/03
2016
The ‘iris’ data
2016/12/30
Making survey statistics boring and inefficient
2016/11/23
Brief quake summary for overseas people
2016/11/14
Changes in turnout and preference
2016/11/10
Cuts to ‘Growing Up in New Zealand’
2016/10/18
Terms to eschew
2016/10/12
Large quadratic forms
2016/09/27
The hard problem of AI and other stories
2016/09/22
Come work with us
2016/09/07
On permuting all the things
2016/09/06
The lithium-powered space bike
2016/09/04
“The” multiple comparisons problem
2016/08/27
Like a crossword
2016/08/20
Simulations and modes of convergence
2016/08/14
Etymology
2016/08/02
A modest proposal: Lazy Ambiguous Single Transferable Vote
2016/07/29
One scoRe years
2016/07/28
How do we prove the Central Limit Theorem?
2016/07/04
Computing the (simplest) sandwich estimator incrementally
2016/06/04
Are there any news?
2016/06/03
Size matters
2016/04/14
Sufficiently advanced technology
2016/04/10
The Great Kiwi Cherry Ripe Scandal
2016/03/29
Mostly dead
2016/03/28
Artistic verisimilitude
2016/03/24
The conservative Bonferroni correction
2016/03/20
Trace estimators and impact factors
2016/03/15
A gene for celibacy?
2016/03/13
Truthy and Sciency
2016/03/02
Coding linear splines
2016/02/29
Cheap tricks
2016/02/28
Two cheers for crowdfunding
2016/02/26
No-one’s forcing you to read the Herald
2016/02/07
Stochastic SVD
2016/02/05
Is it that time of day?
2016/01/20
What does ‘design-consistent’ even mean?
2016/01/13
Another view of the ‘nearly true’ model
2016/01/13
2015
Circumspice
2015/12/31
Superfood sourcing
2015/12/30
The Muntab Question Strikes Back
2015/12/24
Potential energy and kinetic energy
2015/12/22
Case-control estimation is more complicated than you think
2015/12/20
The Muntab Question
2015/12/14
A simple probability problem
2015/12/14
Serious tongue-twister
2015/11/27
Poetry visualisation
2015/11/14
Should SPRINT have stopped?
2015/11/10
Prefiltering very large numbers of tests
2015/10/19
Double robustness
2015/10/18
Convergent evolution and NZ Bird of the Year
2015/10/05
NZ Flag Referendum pseudorandom numbers
2015/09/22
Oranges and lemons
2015/09/21
Good reasons for assuming a spherical cow
2015/09/14
(high-dimensional) Space is Big.
2015/09/14
Net Reclassification Index: surprisingly weird.
2015/08/29
Colour names from XKCD in R
2015/08/20
A conservation tragedy
2015/08/20
Fox fails statistics; does NYT?
2015/08/06
JSM2015: notes on Seattle from an ex-resident
2015/08/05
Pianos, heaps, and ethics of randomisation
2015/08/01
Te Wiki o Te Reo Māori
2015/07/27
stringsAsFactors = <sigh>
2015/07/25
Pi day
2015/07/02
A much-needed gap
2015/06/20
Countermatching
2015/06/03
Zero-inflated Poisson from complex samples
2015/05/26
Call me, Ishmael
2015/05/20
Superefficiency
2015/05/12
Precise answers, but not necessarily to the right question
2015/05/04
What’s the right proof of the Continuous Mapping Theorem?
2015/05/03
Eppur si muove
2015/04/02
Pharmacy ethics
2015/03/29
Paper helicopters at a science fair
2015/03/28
What does measurability mean?
2015/03/07
How hard did you look: equivalence and non-inferiority
2015/02/27
Clinically proven ingredients
2015/02/26
Science and statistical inference
2015/02/17
Assumptions and testing
2015/01/15
A transitive test is a test for a univariate parameter
2015/01/14
Tomato, tomato
2015/01/12
New header picture
2015/01/12
Different questions can have different answers
2015/01/11
Variation explained and log transformation
2015/01/03
2014
How not to treat Ebola
2014/12/23
Citations: credit or blame
2014/12/14
What science should everyone know?
2014/12/08
It depends on what you mean by 'cost'
2014/11/30
This is just to say
2014/11/06
A people set apart
2014/11/05
Semiparametric efficiency and nearly-true models
2014/10/25
Miasma and Contagion
2014/10/25
Broman's Socks and the Nature of Scientific Reporting
2014/10/20
Is it good or bad when confounding adjustment makes no difference?
2014/09/24
Rhetorical sensitivity analysis
2014/08/29
On dialect
2014/08/29
O necessary sinpi
2014/08/27
Taking meta-analysis heterogeneity seriously
2014/08/24
Survey package update
2014/08/15
Feynman and the Suck Fairy
2014/07/12
Herd Immunity simulations
2014/06/01
Monotonicity and smoothness
2014/05/22
Anchoring bias
2014/05/18
Randomisation without consent
2014/05/14
Einstein, Wikiquote, and fact checking
2014/03/14
My likelihood depends on your frequency properties
2014/03/04
Chemical nerdview
2014/02/25
This is a wug. Now you have two of them.
2014/02/09
2013
At risk of vanishing
2013/12/14
Moving the goalposts?
2013/11/15
From labhacks: the $25 scrunchable scientific poster
2013/11/04
A diversity of gifts, but the same spirit
2013/10/30
Interaction: 'real' and statistical
2013/10/27
Barren proxies
2013/10/20
Google completions and sexism
2013/10/18
Do you know where it's been?
2013/10/10
Today we have shaming of prats
2013/10/06
Rock, paper, scissors, Wilcoxon test
2013/10/06
Auckland's top news story
2013/10/03
Statins and the causal Markov property
2013/09/23
PBRF consultation response consultation
2013/09/22
An absolutely minimal way to increase invited speaker diversity
2013/09/13
What I said on StatsChat only shorter and with more swearing
2013/09/04
On the persistence of variation in horn size among Soay sheep
2013/08/23
A layperson's view of a science communication problem
2013/08/13
SPEED sessions at JSM 2013
2013/08/09
In defense of theory
2013/08/08
Some failure modes of statistics research talks
2013/08/04
Welfare as an addictive drug
2013/07/15
Graphs and counterfactuals
2013/07/15
Sparse linear systems and calibration of weights
2013/07/08
Big data linear models
2013/07/08
Problems with faithfulness and the causal Markov property (II)
2013/07/06
Problems with faithfulness and the causal Markov property (I)
2013/07/02
Upcoming talks and stuff
2013/06/30
Two simple notes on error in regression models
2013/06/28
When is Bayesian introductory statistics better?
2013/06/27
My Setup
2013/06/08
Talks in the near future
2013/06/07
Hello World
2013/06/07
1520
Lorem Ipsum
1520/01/01