Teasing out biological effects and sampling artifacts when using occupancy rate in monitoring programs

ABSTRACT.   Recent declines in biodiversity stress the need for rigorous and reliable multispecies monitoring programs. A potential weakness of monitoring programs is a reliance on raw counts and the assumption either of complete detection or of constant detection probability for each species, regardless of the sampling situation. Until recently, these assumptions have largely remained untested and, therefore, to help insure accuracy, bird-monitoring programs have depended on standardization of counts and counts of longer duration. We tested the effectiveness of these strategies for providing unbiased occupancy rates using a method designed to accommodate situations where species detection probabilities are less than one and heterogeneous. We tested the effect of potential sources of heterogeneity in detection probability (vegetation structure, wind velocity, cloud cover, date, and time) on occupancy rate estimates of 13 bird species in southern France. We compared adjusted and raw occupancy rates for two sampling durations (10 and 20 min). Differences between raw and adjusted occupancy rates were low even for the shorter count duration, suggesting that standardized long counts should produce reliable estimates of occupancy rates even in the absence of correction by an appropriate method. This enhances the value of past monitoring programs where long standardized counts were used, but with designs that do not allow corrected estimates. However, we found that detection probability was heterogeneous for most species and that vegetation structure was an important source of heterogeneity. The possible effects of habitat on detection probability should be of special concern for long-term monitoring programs conducted in landscapes where habitats vary across time or space.     

Adjusting count period strategies to improve the accuracy of forest bird abundance estimates from point transect distance sampling surveys

Although point transect distance sampling methods have become widely used in surveys of forest birds, there has been no attempt to tailor field methods to maximize the accuracy of abundance estimates by minimizing the effects of violations of the method's critical assumptions, which are: (1) birds at distance 0 m are detected with certainty, (2) birds are detected at their initial location and (3) distances to objects are measured accurately. We investigate the effects on abundance estimates for Philippine forest birds of varying the count period from 2 to 10 min, and of including and excluding a pre‐count settling down period. Encounter rates were highly sensitive to count period length but density estimates from 10‐min count periods were, on average, only 13% higher than those from 2‐min periods, and in several cases were actually lower than those from periods of 6–8 min. This was because birds tended to be recorded at greater distances from the recorder as the count period went on, thus ‘stretching out’ detection functions while having little effect on detection rates close to the recorder. For some bird groups, including canopy frugivores and upperstorey gleaning insectivores, density estimates were more than twice as high without than with a settling down period. We suggest that movement away from the recorder is more common than attraction to the recorder, and that unless pilot studies show otherwise, similar studies should not use a settling down period for surveying many species. Count periods that maximized probability of bird detection close to the central point while minimizing the unwanted effects of bird movement during the count period were: 4 min for omnivores, 6 min for nectarivores and upperstorey gleaning insectivores, 8 min for understorey insectivores and canopy frugivores, and a full 10 min for sallying insectivores, ground‐dwellers, carnivores and coucals/koels. We use the results to suggest ‘group‐specific’ count period regimes that could help maximize the accuracy of density estimates from similar studies of tropical forest birds.     

Evaluating the performance of a multilocus Bayesian method for the estimation of migration rates

Bayesian methods have become extremely popular in molecular ecology studies because they allow us to estimate demographic parameters of complex demographic scenarios using genetic data. Articles presenting new methods generally include sensitivity studies that evaluate their performance, but they tend to be limited and need to be followed by a more thorough evaluation. Here we evaluate the performance of a recent method, bayesass , which allows the estimation of recent migration rates among populations, as well as the inbreeding coefficient of each local population. We expand the simulation study of the original publication by considering multi-allelic markers and scenarios with varying number of populations. We also investigate the effect of varying migration rates and F _ST more thoroughly in order to identify the region of parameter space where the method is and is not able to provide accurate estimates of migration rate. Results indicate that if the demographic history of the species being studied fits the assumptions of the inference model, and if genetic differentiation is not too low ( F _ST ≥ 0.05), then the method can give fairly accurate estimates of migration rates even when they are fairly high (about 0.1). However, when the assumptions of the inference model are violated, accurate estimates are obtained only if migration rates are very low ( m  = 0.01) and genetic differentiation is high ( F _ST ≥ 0.10). Our results also show that using posterior assignment probabilities as an indication of how much confidence we can place on the assignments is problematical since the posterior probability of assignment can be very high even when the individual assignments are very inaccurate.     

On the importance of estimating detection probabilities from at‐sea surveys of flying seabirds

The primary and accepted method used to estimate seabird densities at sea from ships is the strip transect method, designed to correct for the effect of random directional bird movement relative to that of the ship. However, this method relies on the critical assumption that all of the birds within the survey strip are detected. We used the distance sampling method from line‐transects to estimate detection probability of a number of species of flying seabirds, and to test whether distance from the ship and bird body size affected detectability. Detection probability decreased from 0.987 (SE=0.029) to 0.269 (SE=0.035) with increasing strip half‐width from 100 to 1400 m. Detection probability also varied between size‐groups of species with strip half‐width. For all size‐groups, this probability was close to 1 for strip half‐width of 100 m, but was 0.869 (SE=0.115), 0.725 (SE=0.096) and 0.693 (SE=0.091) for strip half‐width of 300 m, a typical strip width used in seabird surveys, for respectively large, medium and small size flying seabirds. For larger strip half‐width, detection probability was higher for large sized species, intermediate for medium sized species and lower for smaller sized species. For strip half‐width larger than 100 m we suggest that more attention should be paid to testing the assumption of perfect detectability, because abundance estimates may be underestimated when this assumption is violated. Finally, the effect of the speed of travel of flying seabird on the detection probability was estimated in a simulation study, which suggests that detection probability was underestimated with increasing flying speed.     

Evaluation of three methods to estimate density and detectability from roadside point counts

Roadside point counts are often used to estimate trends of bird populations. The use of aural counts of birds without adjustment for detection probability, however, can lead to incorrect population trend estimates. We compared precision of estimates of density and detectability of whistling northern bobwhites (Colinus virginianus) using distance sampling, independent double-observer, and removal methods from roadside surveys. Two observers independently recorded each whistling bird heard, distance from the observer, and time of first detection at 362 call-count stops in Ohio. We examined models that included covariates for year and observer effects for each method and distance from observer effects for the double-observer and removal methods using Akaike's Information Criterion (AIC). The best model of detectability from distance sampling included observer and year effects. The best models from the removal and double-observer techniques included observer and distance effects. All 3 methods provided precise estimates of detection probability (CV = 2.4–4.4%) with a range of detectability of 0.44–0.95 for a 6-min survey. Density estimates from double-observer surveys had the lowest coefficient of variation (2005 = 3.2%, 2006 = 1.7%), but the removal method also provided precise estimates of density (2005 CV = 3.4%, 2006 CV = 4.8%), and density estimates from distance sampling were less precise (2005 CV = 9.6%, 2006 CV = 7.9%). Assumptions of distance sampling were violated in our study because probability of detecting bobwhites near the observer was <1 or the roadside survey points were not randomly distributed with respect to the birds. Distances also were not consistently recorded by individual members of observer pairs. Although double-observer surveys provided more precise estimates, we recommend using the removal method to estimate detectability and abundance of bobwhites. The removal method provided precise estimates of density and detection probability and requires half the personnel time as double-observer surveys. Furthermore, the likelihood of meeting model assumptions is higher for the removal survey than with independent double-observers. © 2011 The Wildlife Society.     

Testing the importance of auditory detections in avian point counts

ABSTRACT.   Recent advances in the methods used to estimate detection probability during point counts suggest that the detection process is shaped by the types of cues available to observers. For example, models of the detection process based on distance-sampling or time-of-detection methods may yield different results for auditory versus visual cues because of differences in the factors that affect the transmission of these cues from a bird to an observer or differences in an observer's ability to localize cues. Previous studies suggest that auditory detections predominate in forested habitats, but it is not clear how often observers hear birds prior to detecting them visually. We hypothesized that auditory cues might be even more important than previously reported, so we conducted an experiment in a forested habitat in North Carolina that allowed us to better separate auditory and visual detections. Three teams of three observers each performed simultaneous 3-min unlimited-radius point counts at 30 points in a mixed-hardwood forest. One team member could see, but not hear birds, one could hear, but not see, and the third was nonhandicapped. Of the total number of birds detected, 2.9% were detected by deafened observers, 75.1% by blinded observers, and 78.2% by nonhandicapped observers. Detections by blinded and nonhandicapped observers were the same only 54% of the time. Our results suggest that the detection of birds in forest habitats is almost entirely by auditory cues. Because many factors affect the probability that observers will detect auditory cues, the accuracy and precision of avian point count estimates are likely lower than assumed by most field ornithologists.     

Estimating regional landbird populations from enhanced North American Breeding Bird Surveys

Estimating the size of bird populations is central to effective conservation planning and prudent management. I updated estimated regional bird populations for the East Gulf Coastal Plain of Mississippi using data from 275 North American Breeding Bird Surveys from 2009 to 2013. However, regional bird populations estimated from count surveys of breeding birds may be biased due to lack of empirical knowledge of the distance at which a species is effectively detected and the probability of detecting a species if it is present. I used data recorded within two distance classes (0–50 m and >50–400 m) and three 1‐min time intervals on 130 Breeding Bird Surveys to estimate detection probability and effective detection distance for 77 species. Incorporating these empirical estimates of detection probability and detection distance resulted in estimated regional populations for these species that were markedly greater than regional populations estimated without species‐specific estimates of detection parameters. Using the same Breeding Bird Survey data, I also estimated probability of site occupancy for 66 species and extrapolated this to the proportion of area occupied in the East Gulf Coastal Plain of Mississippi. I combined the area occupied with the reported range of breeding territory size for 54 species to obtain independent estimates of regional bird populations. Although the true population of these species is unknown, estimated populations that incorporated empirical estimates of detection probability and detection distance were more likely to be within the range of independently estimated, occupancy‐based, regional population estimates than were population estimates that lacked empirical detection and distance information.     

Estimating survival of precocial chicks during the prefledging period using a catch-curve analysis and count-based age-class data

ABSTRACT.   Estimating reproductive success for birds with precocial young can be difficult because chicks leave nests soon after hatching and individuals or broods can be difficult to track. Researchers often turn to estimating survival during the prefledging period and, though effective, mark-recapture based approaches are not always feasible due to cost, time, and animal welfare concerns. Using a threatened population of Piping Plovers ( Charadrius melodus ) that breeds along the Missouri River, we present an approach for estimating chick survival during the prefledging period using long-term (1993–2005), count-based, age-class data. We used a modified catch-curve analysis, and data collected during three 5-day sampling periods near the middle of the breeding season. The approach has several ecological and statistical assumptions and our analyses were designed to minimize the probability of violating those assumptions. For example, limiting the sampling periods to only 5 days gave reasonable assurance that population size was stable during the sampling period. Annual daily survival estimates ranged from 0.825 (SD = 0.03) to 0.931 (0.02) depending on year and sampling period, with these estimates assuming constant survival during the prefledging period and no change in the age structure of the population. The average probability of survival to fledging ranged from 0.126 to 0.188. Our results are similar to other published estimates for this species in similar habitats. This method of estimating chick survival may be useful for a variety of precocial bird species when mark-recapture methods are not feasible and only count-based age class data are available.     

Testing assumptions of mark–recapture theory in the coral reef fish Lutjanus apodus

This study tested assumptions of the Cormack–Jolly–Seber capture–mark–recapture (CMR) model in a population of the tropical snapper Lutjanus apodus in the central Bahamas using a combination of laboratory and field studies. The suitability of three different tag types [passive integrated transponder (PIT) tag, T-anchor tag and fluorescent dye jet-injected into the fins] was assessed. PIT tags were retained well, whereas T-anchor tags and jet-injected dye were not. PIT tags had no detectable effect on the rates of growth or survival of individuals. The capture method (fish trapping) was found to provide a representative sample of the population; however, a positive trap response was identified and therefore the assumption of equal capture probability was violated. This study illustrates an approach that can be used to test some of the critical assumptions of the CMR theory and it demonstrates that CMR methods can provide unbiased estimates of growth and mortality of L. apodus provided that trap response is explicitly modelled when estimating survival probability.     

Factors Influencing Survival of Radiotagged and Banded Northern Bobwhites in Georgia

Abstract: Numerous studies of behavior and ecology of northern bobwhites (Colinus virginianus) have depended on radiotagging and telemetry for data collection. Excluding the presumably short-term effects of trapping, handling, and attaching radiotransmitters, researchers often assume that little bias is associated with estimating survival and behavioral parameters associated with this technique. However, researchers have not adequately examined these effects on organisms being investigated and have thus assumed demographic information obtained from such methods are valid. In light of this conjecture, it is imperative to evaluate methodological assumptions to ensure research is statistically valid and biologically meaningful. Therefore, we used Burnham's model and program MARK to analyze survival estimates of individually banded and radiotagged bobwhites during an 8-year period (1997–2004) consisting of 6,568 individuals (2,527 radiotagged) via combined analysis of mark—recapture, dead recovery (via harvest), and radiotelemetry data to test the effects of radiotransmitters on bobwhite survival. We also compared band—recapture survival estimates to Kaplan—Meier survival estimates, and we examined the effects of various other factors (e.g., temporal, spatial) on bobwhite survival. Based on Akaike's model selection criterion, the best model including the radiotransmitter covariate (Akaike's Information Criterion adjusted for small sample size bias and overdispersion relative value = 0.72) did not explain more of the variation in survival than models without this effect. Thus, we found the effect of radiotransmitters as negligible. Bobwhite survival varied relative to spatial (e.g., site), temporal (e.g., yr and season), and gender effects. Average annual survival for the 8-year period was 22.76% (1.50 SE) for banded-only and 21.72% (1.49 SE) for radiotagged birds. Survival rate varied annually, ranging from 12.42% (7.51 SE) to 37.16% (8.27 SE), and seasonally, ranging from 23.82% (2.71 SE) to 65.06% (3.23 SE); however, between group (banded-only, radiotagged) survival differences were still inconsequential. We conclude that for our study, radiotelemetry provided reliable survival estimates of an intensively managed bobwhite population, where supplemental food was provided, and this information provided useful data to make practical habitat management decisions. We believe that future radiotelemetry studies would benefit as a whole if researchers conducted similar analyses prior to presenting their results from radiotelemetry data, especially for populations that are more food limited.     

Optimizing surveys for marsh songbirds: does timing matter?

Analysis of data from point counts, a common method for monitoring bird population trends, has evolved to produce estimates of various population parameters (e.g., density, abundance, and occupancy) while simultaneously estimating detection probability. An important consideration when designing studies using point counts is to maximize detection probability while minimizing variation in detection probability both within and between counts. Our objectives were to estimate detection probabilities for three marsh songbirds, including Marsh Wrens (Cistothorus palustris), Swamp Sparrows (Melospiza georgiana), and Yellow‐headed Blackbirds (Xanthocephalus xanthocephalus), as a function of weather covariates and to evaluate temporal variability in detection probability of these three species. We conducted paired, unlimited radius, 10‐min point counts during consecutive morning and evening survey periods for our three focal species at 56 wetlands in Iowa from 20 April to 10 July 2010. Mean detection probabilities ranged from 0.272 (SE = 0.042) for Marsh Wrens to 0.365 (SE = 0.052) for Swamp Sparrows. Time of season was positively correlated with detection probability for Swamp Sparrows, but was negatively correlated with detection probability for Yellow‐headed Blackbirds, suggesting that detection probability increased during the breeding season for Swamp Sparrows and was highest early in the breeding season for Yellow‐headed Blackbirds. Understanding how detection probabilities of marsh songbirds vary throughout the breeding season allows targeted survey efforts that maximize detection probabilities for these species. Furthermore, consistent detection probabilities of marsh songbirds during morning and evening survey periods mean that investigators have more time to conduct surveys for these birds, allowing greater flexibility to increase spatial and temporal replication of surveys that could provide more precise estimates of desired population parameters.     

Monitoring of the Flat-Tailed Horned Lizard With Methods Incorporating Detection Probability

Abstract: Difficulty in monitoring the flat-tailed horned lizard (Phrynosoma mcallii) has led to controversy over its conservation status. The difficulty in detecting this species has discouraged large-scale estimates of abundance and led to uncertainty over whether the species exists in population sizes of sufficient size for long-term persistence. We incorporated detection probability into monitoring of this species using closed mark—recapture and distance-sampling methods. Density estimation from mark—recapture abundance estimates was improved using an estimate of the proportion of time lizards were on the plot. We estimated the probability of detection on the line for distance sampling and adjusted density estimates accordingly. We estimated the populations of the Yuha Basin Management Area in 2002 and the East Mesa Management Area, Imperial County, California, USA, in 2003 to be 25,514 (95% CI 14,444-38,970) and 42,619 (95% CI 23,161-67,639), respectively. Two estimates of detection probability on the line in distance sampling by different methods were 0.45 and 0.65. Density estimates derived from distance analyses for 3 East Mesa Management Area plots and the Yuha Basin Management Area were 1.55 per ha (95% CI 0.64-3.76) and 0.41 per ha (95% CI 0.22-0.7), respectively. These are the first large-scale estimates of abundance and density for P. mcallii.     

Bird diversity: a predictable function of satellite-derived estimates of seasonal variation in canopy light absorbance across the United States

Aim  To investigate the relationships between bird species richness derived from the North American Breeding Bird Survey and estimates of the average, minimum, and the seasonal variation in canopy light absorbance (the fraction of absorbed photosynthetically active radiation, fPAR) derived from NASA's Moderate Resolution Imaging Spectroradiometer (MODIS).
Location  Continental USA.
Methods  We describe and apply a 'dynamic habitat index' (DHI), which incorporates three components based on monthly measures of canopy light absorbance through the year. The three components are the annual sum, the minimum, and the seasonal variation in monthly fPAR, acquired at a spatial resolution of 1 km, over a 6-year period (2000–05). The capacity of these three DHI components to predict bird species richness across 84 defined ecoregions was assessed using regression models.
Results  Total bird species richness showed the highest correlation with the composite DHI [ R ² = 0.88, P  < 0.001, standard error of estimate (SE) = 8 species], followed by canopy nesters ( R ² = 0.79, P  < 0.001, SE = 3 species) and grassland species ( R ² = 0.74, P  < 0.001, SE = 1 species). Overall, the seasonal variation in fPAR, compared with the annual average fPAR, and its spatial variation across the landscape, were the components that accounted for most ( R ² = 0.55–0.88) of the observed variation in bird species richness.
Main conclusions  The strong relationship between the DHI and observed avian biodiversity suggests that seasonal and interannual variation in remotely sensed fPAR can provide an effective tool for predicting patterns of avian species richness at regional and broader scales, across the conterminous USA.     

Abundance estimation of long-diving animals using line transect methods

Line transect sampling is one of the most widely used methods for estimating the size of wild animal populations. An assumption in standard line transect sampling is that all the animals on the trackline are detected without fail. This assumption tends to be violated for marine mammals with surfacing/diving behaviors. The detection probability on the trackline is estimated using duplicate sightings from double-platform line transect methods. The double-platform methods, however, are insufficient to estimate the abundance of long-diving animals because these animals can be completely missed while the observers pass. We developed a more flexible hazard probability model that incorporates information on surfacing/diving patterns obtained from telemetry data. The model is based on a stochastic point process and is statistically tractable. A simulation study showed that the new model provides near-unbiased abundance estimates, whereas the traditional hazard rate and hazard probability models produce considerably biased estimates. As an illustration, we applied the model to data on the Baird's beaked whale (Berardius bairdii) in the western North Pacific.     

Genetic relatedness in open-pollinated families of two leguminous tree species,Robinia pseudoacacia L. and Gleditsia triacanthos L.

Summary When conducting tree breeding experiments, geneticists often assume that individuals from open-pollinated families are halfsibs. The reliability of this assumption was tested using data from enzyme electrophoresis to estimate the genetic relatedness among progeny within 22 open-pollinated families of Robinia pseudoacacia L. (black locust) and 34 open-pollinated families of Gleditsia triacanthos L. (honey locust) from natural stands. An algorithm employing population estimates of fixation indices, pollen allele frequencies, and selfing rates was used to calculate the mean expected number of alleles in common across loci under assumptions of either full-sib (i.e., a single pollen parent) or half-sib (i.e., random mating) relationships. For each open-pollinated family, the average coefficient of relationship among progeny was calculated by linear interpolation from the observed number of alleles in common. For most families of both species, coefficients were significantly higher than 0.25 (half-sib relation), but were significantly lower than 0.50 (full-sib relation). These results suggest that the assumption of a half-sib relationship among progeny of open-pollinated families is violated for these tree species. More critical to the estimation of heritabilities and the prediction of genetic gains was the observation that estimates of the coefficient of relationship varied widely among open-pollinated families (for R. pseudoacacia r ₀=0.20–0.43, mean=0.34; for G. triacanthos r ₀=0.29–0.55, mean=0.36).     

Energy flux, body size and density in relation to bird species richness along an elevational gradient in Taiwan

Aim To examine the species richness of breeding birds along a local elevational gradient and to test the following assumptions of the energy limitation hypothesis: (1) the energy flux through birds is positively correlated with above‐ground net primary productivity, (2) bird density is positively correlated with total energy flux, and (3) bird species richness is positively correlated with bird density. Location An elevational gradient from 1400 to 3700 m on Mt. Yushan, the highest mountain in Taiwan (23°28′30″ N, 120°54′00″ E), with a peak of 3952 m a.s.l. Methods We established 50 sampling stations along the elevational gradient. From March to July 1992, we estimated the density of each bird species using the variable circular‐plot method. Above‐ground net primary productivity was modelled using monthly averages from weather data for the years 1961–90. Results Bird species richness had a hump‐shaped relationship with elevation and with net primary productivity. Bird energy flux was positively correlated with net primary productivity and bird species richness was positively correlated with bird density. The relationship between bird density and energy flux was hump‐shaped, which does not support one assumption of the energy limitation hypothesis. Main conclusions The results supported two essential assumptions of the energy limitation hypothesis. However, when energy availability exceeded a certain level, it could decrease species richness by increasing individual energy consumption, which reduced bird density. Thus, energy availability is a primary factor influencing bird species richness at this scale, but other factors, such as body size, could also play important roles.     

Comparison of Methods for Estimating Bird Abundance and Trends From Historical Count Data

Abstract: The use of bird counts as indices has come under increasing scrutiny because assumptions concerning detection probabilities may not be met, but there also seems to be some resistance to use of model-based approaches to estimating abundance. We used data from the United States Forest Service, Southern Region bird monitoring program to compare several common approaches for estimating annual abundance or indices and population trends from point-count data. We compared indices of abundance estimated as annual means of counts and from a mixed-Poisson model to abundance estimates from a count-removal model with 3 time intervals and a distance model with 3 distance bands. We compared trend estimates calculated from an autoregressive, exponential model fit to annual abundance estimates from the above methods and also by estimating trend directly by treating year as a continuous covariate in the mixed-Poisson model. We produced estimates for 6 forest songbirds based on an average of 621 and 459 points in 2 physiographic areas from 1997 to 2004. There was strong evidence that detection probabilities varied among species and years. Nevertheless, there was good overall agreement across trend estimates from the 5 methods for 9 of 12 comparisons. In 3 of 12 comparisons, however, patterns in detection probabilities potentially confounded interpretation of uncorrected counts. Estimates of detection probabilities differed greatly between removal and distance models, likely because the methods estimated different components of detection probability and the data collection was not optimally designed for either method. Given that detection probabilities often vary among species, years, and observers investigators should address detection probability in their surveys, whether it be by estimation of probability of detection and abundance, estimation of effects of key covariates when modeling count as an index of abundance, or through design-based methods to standardize these effects.     

Estimating treatment effects with partially observed covariates using outcome regression with missing indicators

Missing data is a common issue in research using observational studies to investigate the effect of treatments on health outcomes. When missingness occurs only in the covariates, a simple approach is to use missing indicators to handle the partially observed covariates. The missing indicator approach has been criticized for giving biased results in outcome regression. However, recent papers have suggested that the missing indicator approach can provide unbiased results in propensity score analysis under certain assumptions. We consider assumptions under which the missing indicator approach can provide valid inferences, namely, (1) no unmeasured confounding within missingness patterns; either (2a) covariate values of patients with missing data were conditionally independent of treatment or (2b) these values were conditionally independent of outcome; and (3) the outcome model is correctly specified: specifically, the true outcome model does not include interactions between missing indicators and fully observed covariates. We prove that, under the assumptions above, the missing indicator approach with outcome regression can provide unbiased estimates of the average treatment effect. We use a simulation study to investigate the extent of bias in estimates of the treatment effect when the assumptions are violated and we illustrate our findings using data from electronic health records. In conclusion, the missing indicator approach can provide valid inferences for outcome regression, but the plausibility of its assumptions must first be considered carefully.     

A comparison of methods for estimating the causal effect of a treatment in randomized clinical trials subject to noncompliance

Summary .  We consider the analysis of clinical trials that involve randomization to an active treatment ( T  = 1) or a control treatment ( T  = 0), when the active treatment is subject to all-or-nothing compliance. We compare three approaches to estimating treatment efficacy in this situation: as-treated analysis, per-protocol analysis, and instrumental variable (IV) estimation, where the treatment effect is estimated using the randomization indicator as an IV. Both model- and method-of-moment based IV estimators are considered. The assumptions underlying these estimators are assessed, standard errors and mean squared errors of the estimates are compared, and design implications of the three methods are examined. Extensions of the methods to include observed covariates are then discussed, emphasizing the role of compliance propensity methods and the contrasting role of covariates in these extensions. Methods are illustrated on data from the Women Take Pride study, an assessment of behavioral treatments for women with heart disease.     

Addressing Temporal Variability in Bird Calling with Design and Estimation: A Northern Bobwhite Example

Imprecise or biased density estimates can lead to inadequate conservation action, overexploitation of game species, or lost recreational opportunities. Common approaches to estimating density of avian populations often either ignore the probability that an individual is present within the sampling area but is not available to be sampled (e.g., not vocalizing), or do not consider covariates that could influence availability. Additionally, management decisions made at the management unit scale are often informed by inadequate monitoring practices, such as limited sampling intensity. In such cases, management agencies calculate density by applying correction factors (e.g., detection probabilities estimated using empirical data from a different study system) to count data, rather than estimating a detection function directly using statistical models. We conducted a simulation study using northern bobwhite (Colinus virginianus; bobwhite) as a model species to quantify the consequences of mis-specifying avian point count models on bias and precision of density estimates. We compared bias and precision of estimates from a fully specified distance-sampling model that estimates availability and detection to 4 different mis-specified approaches, including 2 approaches to calculating density using correction factors. Using correction factors to calculate density produced estimates with low bias but relatively lower precision compared to the fully specified model (CV of density estimates at 35 sites over 5 years: fully specified = 10%, correction factors = 25% and 30%). Although the mean precision and bias of the fully specified model improved with more data (70 sites over 5 years, CV = 9%; 35 sites over 10 years, CV = 9%), precision of correction factors did not (70 sites over 5 years, CV = 22% and 27%; 35 sites over 10 years, CV = 24% and 29%). The fully specified model captured the underlying temporal variation in detection and availability. Increasing sampling duration from 5 to 10 years improved modeled estimates of growth rate, even for mis-specified models, but not derived growth rates using pre-determined detection functions. We demonstrated that conducting point counts 3 times/year at a feasible number of sites can produce relatively unbiased estimates of bobwhite density. Pre-determined detection functions can be fortuitously unbiased for certain years, but they are not a reliable method for determining density or identifying trends in density over time. © 2020 The Wildlife Society.