Michael Bishop
  • Member for 10 years, 10 months
  • Last seen more than a month ago
References containing arguments against null hypothesis significance testing?
44 votes

Chris Fraley has taught a whole course on the history of the debate (the link seems to be broken, even though it's still on his official site; here is a copy in Internet Archive). His summary/...

View answer
How to annoy a statistical referee?
16 votes

Using causal language to describe associations in observational data when omitted variables are almost certainly a serious concern.

View answer
Endogeneity versus unobserved heterogeneity
Accepted answer
13 votes

The terms endogeneity and unobserved heterogeneity often refer to the same thing but usage varies somewhat, even within economics, the discipline I most associate with the terms. In a regression ...

View answer
Are there any good movies involving mathematics or probability?
10 votes

Proof was pretty good.

View answer
Including the interaction but not the main effects in a model
9 votes

Arguably, it depends on what you're using your model for. But I've never seen a reason not to run and describe models with main effects, even in cases where the hypothesis is only about the ...

View answer
What is the single most influential book every statistician should read?
8 votes

Andrew Gelman's interesting book recommendations are here: http://thebrowser.com/interviews/andrew-gelman-on-statistics

View answer
Causality test for logistic regression
7 votes

Estimating causal relationships is hard. You can't expect to be given an easy test that can be applied the same way regardless of your data and works reliably. The Granger causality test can't offer ...

View answer
Statistical intuition/data sense
6 votes

In the example you mention, the core issue is causal inference. A good place to start for causal inference is this triple-book-review by Andrew Gelman, and the books reviewed therein. In addition to ...

View answer
Can cross validation be used for causal inference?
6 votes

To respond to the follow-up @Andy posted as an answer here... Although I could not say which estimate is correct and which is false, doesn't the inconsistency in the Assault Conviction and the Gun ...

View answer
Ziliak (2011) opposes the use of p-values and mentions some alternatives; what are they?
6 votes

I recommend focusing on things like confidence intervals and model-checking. Andrew Gelman has done great work on this. I recommend his textbooks but also check out the stuff he's put online, e.g. ...

View answer
Understanding multiple regression output
Accepted answer
6 votes

It seems like you need an introduction to regression. People made book recommendations here. Free book recommendations here. It's hard to make sure you're doing the analysis right when we don't ...

View answer
Graphical data overview (summary) function in R
5 votes

I have found this function helpful... the original author's handle is respiratoryclub. f_summary <- function(data_to_plot) { ## univariate data summary require(nortest) #data <- as.numeric(scan ...

View answer
When to drop a term from a regression model?
5 votes

What are you using this model for? Is parsimony an important goal? More parsimonious models are preferred in some situations, but I wouldn't say parsimony is a good thing in itself. Parsimonious ...

View answer
How to reduce question set of a poorly designed questionnaire?
Accepted answer
5 votes

@Peter and @ Jonathan have some good ideas. Without repeating, I'll add two more. If you need to shorten your survey and you won't suffer by reducing the sample size you could give different people ...

View answer
What is your favorite data visualization blog?
4 votes

Andrew Gelman doesn't limit himself to visualization, but he comments on it frequently. Statistical Modeling, Causal Inference, and Social Science

View answer
Does it make sense to study plots of residuals with respect to the dependent variable?
4 votes

Is it possible you are confusing fitted/predicted values with the actual values? As @gung and @biostat have said, you hope there is no relationship between fitted values and residuals. On the other ...

View answer
What is the expected gain from taking a winner's bet?
Accepted answer
4 votes

@Cardinal hasn't elaborated on his/her correct answer so I'll do so... To learn the expected value of the whole we calculate the expected value of the parts. Let's start with the optional winner's ...

View answer
Recommend references on survey sample weighting
4 votes

Lucky for me, Andrew Gelman decided to discuss this topic on his blog last week! There I found the following books recommended in the comments: Applied Survey Data Analysis by Heeringa, West & ...

View answer
Structural equation modeling for experimental design data
Accepted answer
4 votes

There is no simple yes or no answer. People constantly attempt to make inferences about causal relationships. The question is what assumptions you have to make, and how sensitive your inferences are ...

View answer
How to group-center / standardize variables in R?
3 votes

Here is a data.table solution. It is definitely faster than plyr (relevant only for large data sets). Maybe later I'll do up a dplyr example. # generate example data raw.data <- data.frame( ...

View answer
Importance of multiple linear regression assumptions when building predictive regression models
3 votes

@James makes a great point on estimating prediction intervals, but I don't think you want to say that the only thing that matters for point prediction is validating to avoid overfitting. Some of ...

View answer
Experiment design: Can unbalanced dataset be better than balanced?
3 votes

As you surmised, your level of interest in accurately describing the treatment vs. control groups is relevant, as is the prior information you have on each. Perhaps you have a theoretical reason to ...

View answer
Are there any good movies involving mathematics or probability?
3 votes

The documentary about Andrew Wiles proof of Fermat's Last Theorem is fantastic: http://www.pbs.org/wgbh/nova/proof/ Available on youtube: http://www.youtube.com/watch?v=7FnXgprKgSE

View answer
Appropriate Statistical Analyses - HLM?
3 votes

HLM, aka multi-level models, or mixed models, are a good idea here. The devil is in the details of how you setup such a model though. Consider the following questions: Does each individual student'...

View answer
Statistics for online dating sites
3 votes

You asked for a simple model. Here's how I would start with R code: glm(match ~ outdoorDif*outdoorImport + optimistDif*optimistImport, family=binomial(link="logit")) outdoorDif = the ...

View answer
Statistical significance in Excel
3 votes

Someone else will probably answer the precise question that you ask, but you might learn more if you explain what deeper questions you are interested in and get feedback on the best way to answer ...

View answer
Can binning a continuous predictor or DV variable improve large data sets fit?
2 votes

How would binning a variable "reduce noise"? It seems to me, whatever sort of measurement error your variables have, binning is always adding additional measurement error in your variables. I'm ...

View answer
What method can be used to correlate variables
Accepted answer
2 votes

You don't describe the data in great detail but regression sounds like a reasonable approach here. There are different flavors of regression depending on how the outcome variable is distributed. The ...

View answer
Visualise and interpret large set of univariate skewed data
2 votes

A histogram is a fine choice when you have lots of data. I'm guessing your problems are that with a histogram: your high outliers won't be noticed. the visual is too small because the x-axis is ...

View answer
Resources for learning Stata
2 votes

Alex Tabarrok posted a great list of resources here on marginalrevolution.com Gabriel Rossman shares a good introductory guide here.

View answer