The Comparison of Latent Variable Propensity Score Models to Traditional Propensity Score Models under Conditions of Covariate Unreliability

Propensity Score Weighting with Missing Data on Covariates and Clustered Data Structure

Propensity score (PS) analyses are increasingly popular in behavioral sciences. Two issues often add complexities to PS analyses, including missing data in observed covariates and clustered data structure. In previous research, methods for conducting PS analyses with considering either issue alone were examined. In practice, the two issues often co-occur; but the performance of methods for PS analyses in the presence of both issues has not been evaluated previously. In this study, we consider PS weighting analysis when data are clustered and observed covariates have missing values. A simulation study is conducted to evaluate the performance of different missing data handling methods (complete-case, single-level imputation, or multilevel imputation) combined with different multilevel PS weighting methods (fixed- or random-effects PS models, inverse-propensity-weighting or the clustered weighting, weighted single-level or multilevel outcome models). The results suggest that the bias in average treatment effect estimation can be reduced, by better accounting for clustering in both the missing data handling stage (such as with the multilevel imputation) and the PS analysis stage (such as with the fixed-effects PS model, clustered weighting, and weighted multilevel outcome model). A real-data example is provided for illustration.

Effects of herbicides and fertilization on biofilms of Pampean lotic systems: A microcosm study

We experimentally assessed the impact of the application of herbicides and fertilizers derived from agricultural activity through the individual and simultaneous addition of glyphosate, atrazine, and nutrients (nitrogen 'N' and phosphorus 'P') on the biofilm community and their resilience when the experimental factors were removed. We hypothesize that i) the presence of agrochemicals negatively affects the biofilm community leading to the simplification of the community structure; ii) the individual or simultaneous addition of herbicides and nutrients produces differential responses in the biofilm; and iii) the degree of biofilm recovery differs according to the treatment applied. Environmentally relevant concentrations of glyphosate (0.7 mgL-1), atrazine (44 μgL-1), phosphorus (1 mg P L-1 [KH2PO4]), and nitrogen (3 mg N L-1[NaNO3]) were used. Chlorophyll a, ash-free dry weight, abundance of main biofilm groups and nutrient contents in biofilm were analyzed. At initial exposure time, all treatments were dominated by Cyanobacteria; through the exposure period, it was observed a progressive replacement by Bacillariophyceae. This replacement occurred on day 3 for the control and was differentially delayed in all herbicides and/or nutrient treatments in which the abundance of cyanobacteria remains significant yet in T5. A significant correlation was observed between the abundance of cyanobacteria and the concentration of atrazine, suggesting that this group is less sensitive than diatoms. The presence of agrochemicals exerted differential effects on the different algal groups. Herbicides contributed to phosphorus and nitrogen inputs. The most frequently observed interactions between experimental factors (nutrients and herbicides) was additivity excepting for species richness (antagonistic effect). In the final recovery time, no significant differences were found between the treatments and the control in most of the evaluated parameters, evincing the resilience of the community.

Unreliable Continuous Treatment Indicators in Propensity Score Analysis

Propensity score analyses (PSA) of continuous treatments often operationalize the treatment as a multi-indicator composite, and its composite reliability is unreported. Latent variables or factor scores accounting for this unreliability are seldom used as alternatives to composites. This study examines the effects of the unreliability of indicators of a latent treatment in PSA using the generalized propensity score (GPS). A Monte Carlo simulation study was conducted varying composite reliability, continuous treatment representation, variability of factor loadings, sample size, and number of treatment indicators to assess whether Average Treatment Effect (ATE) estimates differed in their relative bias, Root Mean Squared Error, and coverage rates. Results indicate that low composite reliability leads to underestimation of the ATE of latent continuous treatments, while the number of treatment indicators and variability of factor loadings show little effect on ATE estimates, after controlling for overall composite reliability. The results also show that, in correctly specified GPS models, the effects of low composite reliability can be somewhat ameliorated by using factor scores that were estimated including covariates. An illustrative example is provided using survey data to estimate the effect of teacher adoption of a workbook related to a virtual learning environment in the classroom.