andrew_gelman_stats andrew_gelman_stats-2013 andrew_gelman_stats-2013-1908 knowledge-graph by maker-knowledge-mining
Source: html
Introduction: Mike Johns writes: Are you familiar with the work of Ai and Norton on interactions in logit/probit models? I’d be curious to hear your thoughts. Ai, C.R. and Norton E.C. 2003. Interaction terms in logit and probit models. Economics Letters 80(1): 123-129. A peer ref just cited this paper in reaction to a logistic model we tested and claimed that the “only” way to test an interaction in logit/probit regression is to use the cross derivative method of Ai & Norton. I’ve never heard of this issue or method. It leaves me wondering what the interaction term actually tests (something Ai & Norton don’t discuss) and why such an important discovery is not more widely known. Is this an issue that is of particular relevance to econometric analysis because they approach interactions from the difference-in-difference perspective? Full disclosure, I’m coming from a social science/epi background. Thus, i’m not interested in the d-in-d estimator; I want to know if any variables modify the rela
sentIndex sentText sentNum sentScore
1 Mike Johns writes: Are you familiar with the work of Ai and Norton on interactions in logit/probit models? [sent-1, score-0.227]
2 A peer ref just cited this paper in reaction to a logistic model we tested and claimed that the “only” way to test an interaction in logit/probit regression is to use the cross derivative method of Ai & Norton. [sent-10, score-1.295]
3 It leaves me wondering what the interaction term actually tests (something Ai & Norton don’t discuss) and why such an important discovery is not more widely known. [sent-12, score-0.862]
4 Is this an issue that is of particular relevance to econometric analysis because they approach interactions from the difference-in-difference perspective? [sent-13, score-0.449]
5 Thus, i’m not interested in the d-in-d estimator; I want to know if any variables modify the relationship between the focal IV and DV. [sent-15, score-0.191]
6 The standard method of calculating and testing the interaction term seems perfectly reasonable for answering this question. [sent-16, score-0.935]
7 My reply: My quick answer is that with nonlinear functions, there is no single summary that tells the whole story. [sent-17, score-0.238]
8 Different coefficients and different average predictive quantities are appropriate in different contexts. [sent-18, score-0.419]
wordName wordTfidf (topN-words)
[('ai', 0.47), ('norton', 0.404), ('interaction', 0.333), ('interactions', 0.154), ('ref', 0.141), ('term', 0.128), ('johns', 0.12), ('probit', 0.118), ('derivative', 0.118), ('disclosure', 0.115), ('econometric', 0.112), ('iv', 0.11), ('modify', 0.11), ('method', 0.109), ('answering', 0.107), ('estimator', 0.105), ('calculating', 0.105), ('cross', 0.102), ('issue', 0.099), ('leaves', 0.098), ('nonlinear', 0.097), ('quantities', 0.097), ('logit', 0.094), ('tested', 0.093), ('different', 0.089), ('letters', 0.088), ('mike', 0.088), ('discovery', 0.088), ('peer', 0.085), ('perfectly', 0.085), ('relevance', 0.084), ('cited', 0.082), ('functions', 0.082), ('relationship', 0.081), ('claimed', 0.081), ('widely', 0.08), ('tells', 0.077), ('logistic', 0.077), ('coefficients', 0.076), ('curious', 0.075), ('reaction', 0.074), ('familiar', 0.073), ('wondering', 0.07), ('predictive', 0.068), ('material', 0.068), ('testing', 0.068), ('hear', 0.068), ('http', 0.067), ('tests', 0.065), ('summary', 0.064)]
simIndex simValue blogId blogTitle
same-blog 1 1.0 1908 andrew gelman stats-2013-06-21-Interpreting interactions in discrete-data regression
Introduction: Mike Johns writes: Are you familiar with the work of Ai and Norton on interactions in logit/probit models? I’d be curious to hear your thoughts. Ai, C.R. and Norton E.C. 2003. Interaction terms in logit and probit models. Economics Letters 80(1): 123-129. A peer ref just cited this paper in reaction to a logistic model we tested and claimed that the “only” way to test an interaction in logit/probit regression is to use the cross derivative method of Ai & Norton. I’ve never heard of this issue or method. It leaves me wondering what the interaction term actually tests (something Ai & Norton don’t discuss) and why such an important discovery is not more widely known. Is this an issue that is of particular relevance to econometric analysis because they approach interactions from the difference-in-difference perspective? Full disclosure, I’m coming from a social science/epi background. Thus, i’m not interested in the d-in-d estimator; I want to know if any variables modify the rela
2 0.22511294 888 andrew gelman stats-2011-09-03-A psychology researcher asks: Is Anova dead?
Introduction: A research psychologist writes in with a question that’s so long that I’ll put my answer first, then put the question itself below the fold. Here’s my reply: As I wrote in my Anova paper and in my book with Jennifer Hill, I do think that multilevel models can completely replace Anova. At the same time, I think the central idea of Anova should persist in our understanding of these models. To me the central idea of Anova is not F-tests or p-values or sums of squares, but rather the idea of predicting an outcome based on factors with discrete levels, and understanding these factors using variance components. The continuous or categorical response thing doesn’t really matter so much to me. I have no problem using a normal linear model for continuous outcomes (perhaps suitably transformed) and a logistic model for binary outcomes. I don’t want to throw away interactions just because they’re not statistically significant. I’d rather partially pool them toward zero using an inform
3 0.19044496 781 andrew gelman stats-2011-06-28-The holes in my philosophy of Bayesian data analysis
Introduction: I’ve been writing a lot about my philosophy of Bayesian statistics and how it fits into Popper’s ideas about falsification and Kuhn’s ideas about scientific revolutions. Here’s my long, somewhat technical paper with Cosma Shalizi. Here’s our shorter overview for the volume on the philosophy of social science. Here’s my latest try (for an online symposium), focusing on the key issues. I’m pretty happy with my approach–the familiar idea that Bayesian data analysis iterates the three steps of model building, inference, and model checking–but it does have some unresolved (maybe unresolvable) problems. Here are a couple mentioned in the third of the above links. Consider a simple model with independent data y_1, y_2, .., y_10 ~ N(θ,σ^2), with a prior distribution θ ~ N(0,10^2) and σ known and taking on some value of approximately 10. Inference about μ is straightforward, as is model checking, whether based on graphs or numerical summaries such as the sample variance and skewn
4 0.18334363 1257 andrew gelman stats-2012-04-10-Statisticians’ abbreviations are even less interesting than these!
Introduction: From AC, AI, and AIH to WAHM, WOHM, and WM. P.S. That was all pretty pointless, so I’ll throw in this viral Jim Henson link (from the same source) for free.
5 0.16877016 823 andrew gelman stats-2011-07-26-Including interactions or not
Introduction: Liz Sanders writes: I viewed your 2005 presentation “Interactions in multilevel models” and was hoping you or one of your students/colleagues could point me to some readings about the issue of using all possible vs. only particular interaction terms in regression models with continuous covariates (I think “functional form validity” is the term I have encountered in the past). In particular, I am trying to understand whether I would be mis-specifying a model if I deleted two of its interaction terms (in favor of using only 2-way treatment interaction terms). The general full model, for example, is: Y = intercept + txt + pre1 + pre2 + txt*pre1 + txt*pre2 + pre1*pre2 + txt*pre1*pre2, where txt is effect coded (1=treatment, -1=control) and pre1 and pre2 are two different pretests that are assumed normally distributed. (The model is actually a multilevel model; the error terms are not listed for brevity.) The truncated model, on the other hand, would only test 2-way treatment inte
7 0.13421635 1686 andrew gelman stats-2013-01-21-Finite-population Anova calculations for models with interactions
8 0.13166574 1887 andrew gelman stats-2013-06-07-“Happy Money: The Science of Smarter Spending”
9 0.11915234 2274 andrew gelman stats-2014-03-30-Adjudicating between alternative interpretations of a statistical interaction?
11 0.11145376 753 andrew gelman stats-2011-06-09-Allowing interaction terms to vary
12 0.10992305 1719 andrew gelman stats-2013-02-11-Why waste time philosophizing?
13 0.10676562 1786 andrew gelman stats-2013-04-03-Hierarchical array priors for ANOVA decompositions
14 0.097126588 39 andrew gelman stats-2010-05-18-The 1.6 rule
15 0.093712144 251 andrew gelman stats-2010-09-02-Interactions of predictors in a causal model
16 0.088891439 368 andrew gelman stats-2010-10-25-Is instrumental variables analysis particularly susceptible to Type M errors?
17 0.082082674 288 andrew gelman stats-2010-09-21-Discussion of the paper by Girolami and Calderhead on Bayesian computation
18 0.081810631 782 andrew gelman stats-2011-06-29-Putting together multinomial discrete regressions by combining simple logits
19 0.08024019 1431 andrew gelman stats-2012-07-27-Overfitting
20 0.078274719 2294 andrew gelman stats-2014-04-17-If you get to the point of asking, just do it. But some difficulties do arise . . .
topicId topicWeight
[(0, 0.128), (1, 0.079), (2, 0.014), (3, -0.029), (4, 0.035), (5, 0.013), (6, -0.021), (7, -0.031), (8, 0.051), (9, 0.059), (10, 0.006), (11, 0.015), (12, -0.008), (13, -0.036), (14, 0.01), (15, 0.007), (16, -0.014), (17, -0.016), (18, -0.019), (19, -0.02), (20, 0.035), (21, 0.012), (22, 0.02), (23, -0.025), (24, 0.006), (25, -0.019), (26, 0.026), (27, -0.026), (28, -0.01), (29, -0.003), (30, 0.023), (31, 0.062), (32, 0.035), (33, 0.007), (34, 0.007), (35, -0.048), (36, -0.007), (37, 0.025), (38, -0.004), (39, 0.031), (40, -0.006), (41, -0.046), (42, 0.079), (43, 0.004), (44, 0.006), (45, 0.01), (46, 0.004), (47, -0.037), (48, 0.016), (49, 0.013)]
simIndex simValue blogId blogTitle
same-blog 1 0.97650468 1908 andrew gelman stats-2013-06-21-Interpreting interactions in discrete-data regression
Introduction: Mike Johns writes: Are you familiar with the work of Ai and Norton on interactions in logit/probit models? I’d be curious to hear your thoughts. Ai, C.R. and Norton E.C. 2003. Interaction terms in logit and probit models. Economics Letters 80(1): 123-129. A peer ref just cited this paper in reaction to a logistic model we tested and claimed that the “only” way to test an interaction in logit/probit regression is to use the cross derivative method of Ai & Norton. I’ve never heard of this issue or method. It leaves me wondering what the interaction term actually tests (something Ai & Norton don’t discuss) and why such an important discovery is not more widely known. Is this an issue that is of particular relevance to econometric analysis because they approach interactions from the difference-in-difference perspective? Full disclosure, I’m coming from a social science/epi background. Thus, i’m not interested in the d-in-d estimator; I want to know if any variables modify the rela
2 0.8106637 257 andrew gelman stats-2010-09-04-Question about standard range for social science correlations
Introduction: Andrew Eppig writes: I’m a physicist by training who is transitioning to the social sciences. I recently came across a reference in the Economist to a paper on IQ and parasites which I read as I have more than a passing interest in IQ research (having read much that you and others (e.g., Shalizi, Wicherts) have written). In this paper I note that the authors find a very high correlation between national IQ and parasite prevalence. The strength of the correlation (-0.76 to -0.82) surprised me, as I’m used to much weaker correlations in the social sciences. To me, it’s a bit too high, suggesting that there are other factors at play or that one of the variables is merely a proxy for a large number of other variables. But I have no basis for this other than a gut feeling and a memory of a plot on Language Log about the distribution of correlation coefficients in social psychology. So my question is this: Is a correlation in the range of (-0.82,-0.76) more likely to be a correlatio
3 0.79207879 1967 andrew gelman stats-2013-08-04-What are the key assumptions of linear regression?
Introduction: Andy Cooper writes: A link to an article , “Four Assumptions Of Multiple Regression That Researchers Should Always Test”, has been making the rounds on Twitter. Their first rule is “Variables are Normally distributed.” And they seem to be talking about the independent variables – but then later bring in tests on the residuals (while admitting that the normally-distributed error assumption is a weak assumption). I thought we had long-since moved away from transforming our independent variables to make them normally distributed for statistical reasons (as opposed to standardizing them for interpretability, etc.) Am I missing something? I agree that leverage in a influence is important, but normality of the variables? The article is from 2002, so it might be dated, but given the popularity of the tweet, I thought I’d ask your opinion. My response: There’s some useful advice on that page but overall I think the advice was dated even in 2002. In section 3.6 of my book wit
4 0.78493792 888 andrew gelman stats-2011-09-03-A psychology researcher asks: Is Anova dead?
Introduction: A research psychologist writes in with a question that’s so long that I’ll put my answer first, then put the question itself below the fold. Here’s my reply: As I wrote in my Anova paper and in my book with Jennifer Hill, I do think that multilevel models can completely replace Anova. At the same time, I think the central idea of Anova should persist in our understanding of these models. To me the central idea of Anova is not F-tests or p-values or sums of squares, but rather the idea of predicting an outcome based on factors with discrete levels, and understanding these factors using variance components. The continuous or categorical response thing doesn’t really matter so much to me. I have no problem using a normal linear model for continuous outcomes (perhaps suitably transformed) and a logistic model for binary outcomes. I don’t want to throw away interactions just because they’re not statistically significant. I’d rather partially pool them toward zero using an inform
5 0.77031153 1462 andrew gelman stats-2012-08-18-Standardizing regression inputs
Introduction: Andy Flies, Ph.D. candidate in zoology, writes: After reading your paper about scaling regression inputs by two standard deviations I found your blog post stating that you wished you had scaled by 1 sd and coded the binary inputs as -1 and 1. Here is my question: If you code the binary input as -1 and 1, do you then standardize it? This makes sense to me because the mean of the standardized input is then zero and the sd is 1, which is what the mean and sd are for all of the other standardized inputs. I know that if you code the binary input as 0 and 1 it should not be standardized. Also, I am not interested in the actual units (i.e. mg/ml) of my response variable and I would like to compare a couple of different response variables that are on different scales. Would it make sense to standardize the response variable also? My reply: No, I don’t standardize the binary input. The point of standardizing inputs is to make the coefs directly interpretable, but with binary i
6 0.7383464 1121 andrew gelman stats-2012-01-15-R-squared for multilevel models
7 0.7355361 14 andrew gelman stats-2010-05-01-Imputing count data
8 0.72858655 1686 andrew gelman stats-2013-01-21-Finite-population Anova calculations for models with interactions
10 0.72292572 753 andrew gelman stats-2011-06-09-Allowing interaction terms to vary
12 0.70979971 1294 andrew gelman stats-2012-05-01-Modeling y = a + b + c
13 0.70704818 375 andrew gelman stats-2010-10-28-Matching for preprocessing data for causal inference
14 0.69645888 684 andrew gelman stats-2011-04-28-Hierarchical ordered logit or probit
15 0.69190621 1070 andrew gelman stats-2011-12-19-The scope for snooping
16 0.68223941 1663 andrew gelman stats-2013-01-09-The effects of fiscal consolidation
17 0.68056041 251 andrew gelman stats-2010-09-02-Interactions of predictors in a causal model
18 0.67778587 451 andrew gelman stats-2010-12-05-What do practitioners need to know about regression?
19 0.67594618 1218 andrew gelman stats-2012-03-18-Check your missing-data imputations using cross-validation
20 0.67591745 1981 andrew gelman stats-2013-08-14-The robust beauty of improper linear models in decision making
topicId topicWeight
[(15, 0.287), (16, 0.054), (17, 0.013), (21, 0.032), (22, 0.018), (24, 0.107), (30, 0.014), (63, 0.043), (76, 0.014), (89, 0.014), (99, 0.288)]
simIndex simValue blogId blogTitle
1 0.9832201 908 andrew gelman stats-2011-09-14-Type M errors in the lab
Introduction: Jeff points us to this news article by Asher Mullard: Bayer halts nearly two-thirds of its target-validation projects because in-house experimental findings fail to match up with published literature claims, finds a first-of-a-kind analysis on data irreproducibility. An unspoken industry rule alleges that at least 50% of published studies from academic laboratories cannot be repeated in an industrial setting, wrote venture capitalist Bruce Booth in a recent blog post. A first-of-a-kind analysis of Bayer’s internal efforts to validate ‘new drug target’ claims now not only supports this view but suggests that 50% may be an underestimate; the company’s in-house experimental data do not match literature claims in 65% of target-validation projects, leading to project discontinuation. . . . Khusru Asadullah, Head of Target Discovery at Bayer, and his colleagues looked back at 67 target-validation projects, covering the majority of Bayer’s work in oncology, women’s health and cardiov
2 0.97503018 834 andrew gelman stats-2011-08-01-I owe it all to the haters
Introduction: Sometimes when I submit an article to a journal it is accepted right away or with minor alterations. But many of my favorite articles were rejected or had to go through an exhausting series of revisions. For example, this influential article had a very hostile referee and we had to seriously push the journal editor to accept it. This one was rejected by one or two journals before finally appearing with discussion. This paper was rejected by the American Political Science Review with no chance of revision and we had to publish it in the British Journal of Political Science, which was a bit odd given that the article was 100% about American politics. And when I submitted this instant classic (actually at the invitation of the editor), the referees found it to be trivial, and the editor did me the favor of publishing it but only by officially labeling it as a discussion of another article that appeared in the same issue. Some of my most influential papers were accepted right
3 0.96307349 439 andrew gelman stats-2010-11-30-Of psychology research and investment tips
Introduction: A few days after “ Dramatic study shows participants are affected by psychological phenomena from the future ,” (see here ) the British Psychological Society follows up with “ Can psychology help combat pseudoscience? .” Somehow I’m reminded of that bit of financial advice which says, if you want to save some money, your best investment is to pay off your credit card bills.
4 0.96160245 1081 andrew gelman stats-2011-12-24-Statistical ethics violation
Introduction: A colleague writes: When I was in NYC I went to this party by group of Japanese bio-scientists. There, one guy told me about how the biggest pharmaceutical company in Japan did their statistics. They ran 100 different tests and reported the most significant one. (This was in 2006 and he said they stopped doing this few years back so they were doing this until pretty recently…) I’m not sure if this was 100 multiple comparison or 100 different kinds of test but I’m sure they wouldn’t want to disclose their data… Ouch!
5 0.96130407 1541 andrew gelman stats-2012-10-19-Statistical discrimination again
Introduction: Mark Johnstone writes: I’ve recently been investigating a new European Court of Justice ruling on insurance calculations (on behalf of MoneySuperMarket) and I found something related to statistics that caught my attention. . . . The ruling (which comes into effect in December 2012) states that insurers in Europe can no longer provide different premiums based on gender. Despite the fact that women are statistically safer drivers, unless it’s biologically proven there is a causal relationship between being female and being a safer driver, this is now seen as an act of discrimination (more on this from the Wall Street Journal). However, where do you stop with this? What about age? What about other factors? And what does this mean for the application of statistics in general? Is it inherently unjust in this context? One proposal has been to fit ‘black boxes’ into cars so more individual data can be collected, as opposed to relying heavily on aggregates. For fans of data and s
7 0.95322859 1624 andrew gelman stats-2012-12-15-New prize on causality in statstistics education
8 0.94905484 1394 andrew gelman stats-2012-06-27-99!
9 0.94258767 945 andrew gelman stats-2011-10-06-W’man < W’pedia, again
10 0.94032478 133 andrew gelman stats-2010-07-08-Gratuitous use of “Bayesian Statistics,” a branding issue?
same-blog 11 0.94023073 1908 andrew gelman stats-2013-06-21-Interpreting interactions in discrete-data regression
12 0.93732572 2278 andrew gelman stats-2014-04-01-Association for Psychological Science announces a new journal
13 0.93398905 1794 andrew gelman stats-2013-04-09-My talks in DC and Baltimore this week
14 0.90630603 762 andrew gelman stats-2011-06-13-How should journals handle replication studies?
15 0.90322709 1833 andrew gelman stats-2013-04-30-“Tragedy of the science-communication commons”
16 0.90113926 1998 andrew gelman stats-2013-08-25-A new Bem theory
17 0.88721609 1800 andrew gelman stats-2013-04-12-Too tired to mock
18 0.8841812 1393 andrew gelman stats-2012-06-26-The reverse-journal-submission system
20 0.88198924 1779 andrew gelman stats-2013-03-27-“Two Dogmas of Strong Objective Bayesianism”