Super Crunchers – Ayres (2007) – 1/2

With the Triumph of Numbers, I read and wrote about the power of using numbers, and how the observation of empirical regularities led to the basic knowledge on how to use such numbers. Already in the triumph of numbers, it was indicated how valuable (numerical) data were regarded to be, for instance by the recollection how the first censuses were regarded as state secrets, because the information could be used to make assertions about the military strength of (rival) nations.

Unfortunately, I.B. Cohen’s Triumph of Numbers ended quite abruptly with a description of Florence Nightingale. It felt unfinished. But the use of numbers has evolved since, and quite substantially so.

How much our use of numerical data has evolved, and to what extent is has invaded our daily lives (without many of us knowing it!), is convincingly described by Ian Ayers, in his magnificent book ‘Super Crunchers’ (2007).

Companies know more and more (and more!) about you: you buy products online, you speak with the customer relations department (with a person behind a computer), you gain discounts with customer cards, and of course you are careful to make sure you receive you frequent flyer miles. Right? If not, you may have bought it all using a credit card, the transactions of which are stored anyway. … Continue Reading

Introducing Influence.ME: Tools for detecting influential data in mixed models

I’m highly excited to announce that influence.ME is now available. Influence.ME is a new software package for R, providing statistical tools for detecting influential data in mixed models. It has been developed by Rense Nieuwenhuis, Ben Pelzer, and Manfred te Grotenhuis. The basic rationale behind identifying influential data is that when iteratively single units are omitted from the data, models based on these data should not produce substantially different estimates. To standardize the assessment of how influential data is, several measures of influence are commonly used, such as DFBETAS and Cook’s Distance.

… Continue Reading

R-Sessions 25: Book – Mixed Effects Models in S and S-PLUS (Pinheiro & Bates, 2000)

Despite the reference to S and S-PLUS in the title of this book, it offers an excellent guide for the nlme-package in R-Project. Reason for this is the close resemblance between R and S. The nlme-package, available in R-Project for estimation of both linear and non-linear multilevel models, is written and maintained by the authors of this book.
… Continue Reading

R-Sessions 23: Book: Data Analysis Using Regression and Multilevel/Hierarchical Models — Gelman & Hill (2007)

Data Analysis Using Regression and Multilevel/Hierarchical Models

Andrew Gelman is known for his expertise on Bayesian statistics. Based on that knowledge he wrote a book in multilevel regression using R and WINbugs. This book aims to be a thorough description of (multilevel) regression techniques, implementation of these techniques in R and bugs, and a guide on interpreting the results of your analyses. Shortly put, the books excels on all three subjects.
… Continue Reading

R-Sessions 17: Generalized Multilevel {lme4}

Although all introductions on regression seem to be based on the assumption of data that is distributed normally, in practice this is not the case. Many other types of distributions exist. To name a few: normal distribution, binomial distribution, poisson, gaussian and so on. The lmer()-function in the lme4-package can easily estimate models based on these distributions. This is done by adding the ‘family’-argument to the command syntax, thereby specifying that not a linear multilevel model needs to be estimated, but a generalized linear model.
… Continue Reading