Interrupted time series analysis (ITSA) is a statistical procedure that evaluates whether an intervention causes a change in the intercept and/or slope of the time series. However, very little research has accessed causal learning in interrupted time series situations. We systematically investigated whether people are able to learn causal influences from a process akin to ITSA, and compared four different presentation formats of stimuli. We found that participants' judgments agreed with ITSA in cases in which the pre-intervention slope is zero or in the same direction as the changes in intercept or slope. However, participants had considerable difficulty controlling for pre-intervention slope when it is in the opposite direction of the changes in intercept or slope. The presentation formats didn't affect judgments in most cases, but did in one. We discuss these results in terms of two potential heuristics that people might use aside from a process akin to ITSA.

Delays between causes and effects are commonly found in cause-effect relationships in real life. However, previous studies have only investigated delays on the order of seconds. In the current study we tested whether people can learn a cause-effect relation with hour long delays. The delays between the cause and effect were either 0, 3, 9, or 21 hours, and the study lasted 16 days. Surprisingly, we found that participants were able to learn the causal relation about equally as well in all four conditions. These findings demonstrate a remarkable ability to accurately learn causal relations in a realistic timeframe that has never been tested before.

Data with multivariate count responses frequently occur in modern applications. The commonly used multinomial-logit model is limiting due to its restrictive mean-variance structure. For instance, analyzing count data from the recent RNA-seq technology by the multinomial-logit model leads to serious errors in hypothesis testing. The ubiquity of over-dispersion and complicated correlation structures among multivariate counts calls for more flexible regression models. In this article, we study some generalized linear models that incorporate various correlation structures among the counts. Current literature lacks a treatment of these models, partly due to the fact that they do not belong to the natural exponential family. We study the estimation, testing, and variable selection for these models in a unifying framework. The regression models are compared on both synthetic and real RNA-seq data.

Data with multiple responses is ubiquitous in modern applications. However, few tools are available for regression analysis of multivariate counts. The most popular multinomial-logit model has a very restrictive mean-variance structure, limiting its applicability to many data sets. This article introduces an R package MGLM, short for multivariate response generalized linear models, that expands the current tools for regression analysis of polytomous data. Distribution fitting, random number generation, regression, and sparse regression are treated in a unifying framework. The algorithm, usage, and implementation details are discussed.

Background

High attrition among research participants undermines the validity and generalizability of field research. This study contrasted two incentivizing methods--money orders (MOs) or rechargeable incentive cards (RICs)--with regard to rates of participants' study engagement and follow-up contact over a 6-month period.

Methods

Substance abusers (N = 303) in Los Angeles, California were recruited and randomized to either an MO (control) or RIC (experimental) condition. All participants were asked to call the researchers at the beginning of each calendar month for the ensuing 5 months to update their locator information, even if nothing had changed. Each call resulted in a $10 payment, issued immediately via the RIC system or by MO by mail. Research staff located and interviewed all participants at Month 6. Contact logs assessed level of effort required to locate participants and conduct follow-up interviews.

Results

Relative to controls, RIC participants, especially those with low ability to defer gratification, initiated more monthly calls. Six-month follow-up rates did not differ between RIC (75%) and controls (79%), though the RIC condition was associated with an average staff time savings of 39.8 minutes per study participant.

Discussion

For longitudinal public health research involving itinerant study participants, the RIC method produces a modest benefit in study engagement and reduced staff time devoted to participant tracking and payments. However, the overall cost-effectiveness of this approach will depend on the pricing model of the card-issuing vendor (which in turns depends on the scale of the project, with per-unit costs falling for larger projects).

Adverse neighborhood social and natural (green space) environments may contribute to the etiology of prostate cancer (CaP), but mechanisms are unclear. We examined associations between neighborhood environment and prostate intratumoral inflammation in 967 men diagnosed with CaP with available tissue samples from 1986-2009 in the Health Professionals Follow-up Study. Exposures were linked to work or residential addresses in 1988. We estimated indices of neighborhood socioeconomic status (nSES) and segregation (Index of Concentration at the Extremes (ICE)) using US Census tract-level data. Surrounding greenness was estimated using seasonal averaged Normalized Difference Vegetation Index (NDVI) data. Surgical tissue underwent pathological review for acute and chronic inflammation, corpora amylacea, and focal atrophic lesions. Adjusted odds ratios (aORs) for inflammation (ordinal) and focal atrophy (binary) were estimated using logistic regression. No associations were observed for acute or chronic inflammation. Each interquartile-range increase in NDVI within 1,230 m of the participants work or home address (aOR = 0.74, 95% confidence interval (CI): 0.59, 0.93), in ICE-income (aOR = 0.79, 95% CI: 0.61, 1.04), and in ICE-race/income (aOR = 0.79, 95% CI: 0.63, 0.99) was associated with lower odds of postatrophic hyperplasia. Interquartile-range increases in nSES (aOR = 0.76, 95% CI: 0.57, 1.02) and ICE-race/income (aOR = 0.73, 95% CI: 0.54, 0.99) were associated with lower odds of tumor corpora amylacea. Histopathological inflammatory features of prostate tumors may be influenced by neighborhood.