Misidentification Error Research Articles

Machine Learning-Based Sample Misidentification Error Detection in Clinical Laboratory Tests: A Retrospective Multicenter Study.

In clinical laboratories, the precision and sensitivity of autoverification technologies are crucial for ensuring reliable diagnostics. Conventional methods have limited sensitivity and applicability, making error detection challenging and reducing laboratory efficiency. This study introduces a machine learning (ML)-based autoverification technology to enhance tumor marker test error detection. The effectiveness of various ML models was evaluated by analyzing a large data set of 397 751 for model training and internal validation and 215 339 for external validation. Sample misidentification was simulated by random shuffling error-free test results with a 1% error rate to achieve a real-world approximation. The ML models were developed with Bayesian optimization for tuning. Model validation was performed internally at the primary institution and externally at other institutions, comparing the ML models' performance with conventional delta check methods. Deep neural networks and extreme gradient boosting achieved an area under the receiver operating characteristic curve of 0.834 to 0.903, outperforming that of conventional methods (0.705 to 0.816). External validation by 3 independent laboratories showed that the balanced accuracy of the ML model ranged from 0.760 to 0.836, outperforming the balanced accuracy of 0.670 to 0.773 of the conventional models. This study addresses limitations regarding the sensitivity of current delta check methods for detection of sample misidentification errors and provides versatile models that mitigate the operational challenges faced by smaller laboratories. Our findings offer a pathway toward more efficient and reliable clinical laboratory testing.

A field experiment demonstrating person misidentification at an appointed meeting place

AbstractMistaking a person for another sometimes occurs; however, few studies have examined it experimentally. Therefore, the reasons behind this kind of person misidentification and its occurrence rate remain unclear, and thus we aimed to demonstrate person misidentification occurs with a certain probability through a field experiment. We also sought to examine whether the similarity between two people affects the occurrence of person misidentification. When 66 undergraduate participants made a rendezvous with an acquaintance, another person who wore similar clothes to the acquaintance or had a similar face appeared. The results showed that in both the conditions, approximately half of the participants made the person misidentification error, and one‐fourth even spoke to the person mistakenly. Moreover, the results indicated that clothing contributed to person misidentification just as much as the face at first sight but became less important over time. This suggests a dynamic shift in person identification depending on time.

Imperfect detection and misidentification affect inferences from data informing water operation decisions

AbstractObjectiveManagers can modify river flow regimes using fish monitoring data to minimize impacts from water management infrastructure. For example, operation of the gate‐controlled Delta Cross Channel (DCC) in California can negatively affect the endangered Sacramento River winter‐run Chinook Salmon Oncorhynchus tshawytscha. Although guidelines have been developed for DCC operations by using real‐time juvenile fish sampling count data, there is uncertainty about how environmental conditions influence fish occupancy and the extent to which those relationships are affected by sampling and identification error.MethodsWe evaluated the effect of environmental conditions, imperfect detection, and misidentification error on salmon occupancy by analyzing data using hierarchical multistate occupancy models. A total of 14,147 trawl tows and beach seine hauls were conducted on 1058 sampling days between October and December from 1996 to 2019. During these surveys, 2803 juvenile winter‐run Chinook Salmon were identified, and approximately 29% of the sampling days had at least one winter‐run juvenile detected.ResultThe probability of misidentifying an individual juvenile winter‐run Chinook Salmon in the field was estimated to be 0.056 based on fish identification examinations and genetic sampling. Occupancy varied considerably and was related to flow characteristics, water clarity, weather, time of year, and whether occupancy was detected during the previous sampling day. However, these relationships and their significance changed considerably when accounting for imperfect detection and the probability of misidentifying individual juvenile salmon. Detection was <0.3 under average sampling conditions during a single sample and was influenced by flow, water clarity, site, and volume sampled.ConclusionOur modeling results indicate that DCC gate closure decisions could occur on fewer days when imperfect detection and misidentification error are not accounted for. These findings demonstrate the need to account for identification and detection error while using monitoring data to assess factors influencing fish occupancy and inform future management decisions.

Machine learning-based delta check method for detecting misidentification errors in tumor marker tests.

Misidentification errors in tumor marker tests can lead to serious diagnostic and treatment errors. This study aims to develop a method for detecting these errors using a machine learning (ML)-based delta check approach, overcoming limitations of conventional methods. We analyzed five tumor marker test results: alpha-fetoprotein (AFP), cancer antigen 19-9 (CA19-9), cancer antigen 125 (CA125), carcinoembryonic antigen (CEA), and prostate-specific antigen (PSA). A total of 246,261 records were used in the analysis. Of these, 179,929 records were used for model training and 66,332 records for performance evaluation. We developed a misidentification error detection model based on the random forest (RF) and deep neural network (DNN) methods. We performed an in silico simulation with 1 % random sample shuffling. The performance of the developed models was evaluated and compared to conventional delta check methods such as delta percent change (DPC), absolute DPC (absDPC), and reference change values (RCV). The DNN model outperformed the RF, DPC, absDPC, and RCV methods in detecting sample misidentification errors. It achieved balanced accuracies of 0.828, 0.842, 0.792, 0.818, and 0.833 for AFP, CA19-9, CA125, CEA, and PSA, respectively. Although the RF method performed better than DPC and absDPC, it showed similar or lower performance compared to RCV. Our research results demonstrate that an ML-based delta check method can more effectively detect sample misidentification errors compared to conventional delta check methods. In particular, the DNN model demonstrated superior and stable detection performance compared to the RF, DPC, absDPC, and RCV methods.

Interobserver error in grassland vegetation surveys: sources and implications

Abstract Aims Observer error is an unavoidable aspect of vegetation surveys involving human observers. We quantified four components of interobserver error associated with long-term monitoring of prairie vegetation: overlooking error, misidentification error, cautious error and estimation error. We also evaluated the association of plot size with pseudoturnover due to observer error, and how documented pseudochanges in species composition and abundance compared with recorded changes in the vegetation over a 4-year interval. Methods This study was conducted at Tallgrass Prairie National Preserve, Kansas. Monitoring sites contained 10 plots; each plot consisted of a series of four nested frames (0.01, 0.1, 1 and 10 m2). The herbaceous species present were recorded in each of the nested frames, and foliar cover was visually estimated within seven cover categories at the 10 m2 spatial scale only. Three hundred total plots (30 sites) were surveyed, and 28 plots selected at random were resurveyed to assess observer error. Four surveyors worked in teams of two. Important Findings At the 10 m2 spatial scale, pseudoturnover resulting from overlooking error averaged 18.6%, compared with 1.4% resulting from misidentification error and 0.6% resulting from cautious error. Pseudoturnover resulting from overlooking error increased as plot size decreased, although relocation error likely played a role. Recorded change in species composition over a 4-year interval (excluding potential misidentification error and cautious error) was 30.7%, which encompassed both pseudoturnover due to overlooking error and actual change. Given a documented overlooking error rate of 18.6%, this suggests the actual change for the 4-year period was only 12.1%. For estimation error, 26.2% of the time a different cover class was recorded. Over the 4-year interval, 46.9% of all records revealed different cover classes, suggesting that 56% of the records of change in cover between the two time periods were due to observer error.

Sexual Dimorphism in Three Species of Heterelmis Sharp (Coleoptera: Elmidae)

Sex identification of many species of adult riffle beetles (Coleoptera: Elmidae) has relied historically upon removal and examination of genitalia, which leads to death if examined individuals are initially alive. This process and its consequences present a substantial problem if investigators intend to study live beetles in captivity, or if an investigator is examining population sex ratios of threatened or at-risk beetles in the wild where destructive sampling should be avoided. We examined external morphological characteristics of three species of Heterelmis Sharp [H. comalensis (Bosse, Tuff, and Brown, 1988), Heterelmis cf. glabra (Horn), and Heterelmis vulnerata (LeConte)] and developed a non-lethal method of identifying sexes by using these characters. Across species, several external morphological characteristics differed between the sexes, but the length of the fifth abdominal ventrite was the most reliable and robust external character found to vary between the sexes. However, sex misidentification error rates using proposed species-specific length thresholds for the fifth ventrite differed among the three Heterelmis species. This easily applied technique has utility for identifying the sexes of the three Heterelmis species in studies that require live subjects and/or minimal mortality, such as in life history studies or observations of sex ratios in the field.

Misidentification error associated with classifications of white‐tailed deer images

ABSTRACTCamera traps are widely used to monitor wildlife, with important management decisions often relying on interpretation of these data. Animal misidentifications are known to be an important source of error in wildlife surveys that require the identification of unique individuals from camera‐trap data; however, the practice of broadly classifying animal images according to sex or age has received less critical attention despite the significant potential for misidentification error under certain circumstances. From 19 January to 1 April 2017, we solicited a group of 726 participants, consisting of both wildlife professionals and nonprofessionals from across the United States, to take an online survey that tested their ability to classify images of known white‐tailed deer (Odocoileus virginianus) according to sex and age. Our goal was to determine the relative influence of tested observer (i.e., experience and familiarity with classifying deer images) and image‐based factors (i.e., distance of deer from camera, day vs. night image) on accuracy of deer classifications. Our results indicated that respondents that were wildlife biologists and those with greater levels of experience viewing deer images were more accurate than others when classifying posthunting season images of deer as adult male, adult female, or fawn. However, the sex–age group of the deer was the most influential predictor of classification reliability, with branched‐antlered adult males being classified more accurately by all respondent groups than were adult females and fawns. Our findings emphasized that animal misidentifications may be an important source of survey error not only when identifying unique individuals, but also under any circumstance where comparative groups lack definitive traits. We suggest that those using camera traps to evaluate wildlife populations should select survey periods that maximize differences among classification groups, when possible, and develop species‐specific image training for observers to improve the reliability of results. Further, population demographics should be considered when evaluating the overall reliability of survey results for species where classification accuracy varies among sex–age groups. © 2019 The Wildlife Society.

Inter-Observer Error in Wetland Vegetation Surveys

Vegetation surveys are a common component of sampling wetlands. Although observer error is known to be ubiquitous in vegetation sampling in general, extremely few studies of observer error have been conducted in wetland habitats. We quantified inter-observer error in sampling emergent, scrub-shrub, and forested wetlands located in Ohio. Two expert observers independently recorded species present and estimated foliar cover within ten cover classes in 10 × 10 m modules, which were units of larger sample plots, in six wetlands. Although the numbers of species recorded by observers were similar, the identities of species recorded differed due to overlooking of species actually present, misidentification error, and not recording species to the same taxonomic level. We calculated rates of pseudoturnover due to each of these three sources of error at different spatial scales (i.e., plot vs. module level), for different vegetation strata (i.e., herbaceous vs. woody) and for the three vegetation types. We also quantified estimation error associated with cover class estimates. Several different types of VIBIs (Vegetation Index of Biotic Integrity) were calculated based on each observer’s records. Pseudoturnover rates ranged from 15 to 40%, and observers recorded different cover classes 59% of the time. VIBI categorization was affected 17% of the time.

Principles of Pediatric Patient Safety: Reducing Harm Due to Medical Care.

Pediatricians render care in an increasingly complex environment, which results in multiple opportunities to cause unintended harm. National awareness of patient safety risks has grown since the National Academy of Medicine (formerly the Institute of Medicine) published its report "To Err Is Human: Building a Safer Health System" in 1999. Patients and society as a whole continue to challenge health care providers to examine their practices and implement safety solutions. The depth and breadth of harm incurred by the practice of medicine is still being defined as reports continue to reveal a variety of avoidable errors, from those that involve specific high-risk medications to those that are more generalizable, such as patient misidentification and diagnostic error. Pediatric health care providers in all practice environments benefit from having a working knowledge of patient safety language. Pediatric providers should serve as advocates for best practices and policies with the goal of attending to risks that are unique to children, identifying and supporting a culture of safety, and leading efforts to eliminate avoidable harm in any setting in which medical care is rendered to children. In this Policy Statement, we provide an update to the 2011 Policy Statement "Principles of Pediatric Patient Safety: Reducing Harm Due to Medical Care."

A response to Gatali and Wallin (2015) Bird diversity in the savanna habitats of Akagera National Park, Rwanda, in the post-war recovery period

The paper ‘Bird diversity in the savanna habitats of Akagera National Park, Rwanda, in the post-war recovery period’ by Gatali and Wallin (Ostrich 86(3): 267–276, 2015) makes several claims for new species records for Akagera National Park and Rwanda. We found that Gatali and Wallin recorded several species considered very rare in Rwanda and new discoveries; these findings often contradicted the existing literature. We doubt the validity of 34 species records and we aim to demonstrate that these claims are mostly the result of misidentification or human error and consider that future researchers should use their findings with caution. Some findings around more common species may be validated through further studies and we encourage others to conduct fieldwork in Akagera National Park.

Effects of Photo and Genotype-Based Misidentification Error on Estimates of Survival, Detection and State Transition using Multistate Survival Models.

We simulated multistate capture histories (CHs) by varying state survival (ϕ), detection (p) and transition (ψ), number of total capture occasions and releases per capture occasion and then modified these scenarios to mimic false rejection error (FRE), a common misidentification error, resulting from the failure to match samples of the same individual. We then fit a multistate model and estimated accuracy, bias and precision of state-specific ϕ, p and ψ to better understand the effects of FRE on different simulation scenarios. As expected, ϕ, and p, decreased in accuracy with FRE, with lower accuracy when CHs were simulated under a shorter-term study and a lower number of releases per capture occasion (lower sample size). Accuracy of ψ estimates were robust to FRE except in those CH scenarios simulated using low sample size. The effect of FRE on bias was not consistent among parameters and differed by CH scenario. As expected, ϕ was negatively biased with increased FRE (except for the low ϕ low p CH scenario simulated with a low sample size), but we found that the magnitude of bias differed by scenario (high p CH scenarios were more negatively biased). State transition was relatively unbiased, except for the low p CH scenarios simulated with a low sample size, which were positively biased with FRE, and high p CH scenarios simulated with a low sample size. The effect of FRE on precision was not consistent among parameters and differed by scenario and sample size. Precision of ϕ decreased with FRE and was lowest with the low ϕ low p CH scenarios. Precision of p estimates also decreased with FRE under all scenarios, except the low ϕ high p CH scenarios. However, precision of ψ increased with FRE, except for those CH scenarios simulated with a low sample size. Our results demonstrate how FRE leads to loss of accuracy in parameter estimates in a multistate model with the exception of ψ when estimated using an adequate sample size.

Impacts of Species Misidentification on Species Distribution Modeling with Presence-Only Data

Spatial records of species are commonly misidentified, which can change the predicted distribution of a species obtained from a species distribution model (SDM). Experiments were undertaken to predict the distribution of real and simulated species using MaxEnt and presence-only data “contaminated” with varying rates of misidentification error. Additionally, the difference between the niche of the target and contaminating species was varied. The results show that species misidentification errors may act to contract or expand the predicted distribution of a species while shifting the predicted distribution towards that of the contaminating species. Furthermore the magnitude of the effects was positively related to the ecological distance between the species’ niches and the size of the error rates. Critically, the magnitude of the effects was substantial even when using small error rates, smaller than common average rates reported in the literature, which may go unnoticed while using a standard evaluation method, such as the area under the receiver operating characteristic curve. Finally, the effects outlined were shown to impact negatively on practical applications that use SDMs to identify priority areas, commonly selected for various purposes such as management. The results highlight that species misidentification should not be neglected in species distribution modeling.

DNA Barcoding through Quaternary LDPC Codes.

For many parallel applications of Next-Generation Sequencing (NGS) technologies short barcodes able to accurately multiplex a large number of samples are demanded. To address these competitive requirements, the use of error-correcting codes is advised. Current barcoding systems are mostly built from short random error-correcting codes, a feature that strongly limits their multiplexing accuracy and experimental scalability. To overcome these problems on sequencing systems impaired by mismatch errors, the alternative use of binary BCH and pseudo-quaternary Hamming codes has been proposed. However, these codes either fail to provide a fine-scale with regard to size of barcodes (BCH) or have intrinsic poor error correcting abilities (Hamming). Here, the design of barcodes from shortened binary BCH codes and quaternary Low Density Parity Check (LDPC) codes is introduced. Simulation results show that although accurate barcoding systems of high multiplexing capacity can be obtained with any of these codes, using quaternary LDPC codes may be particularly advantageous due to the lower rates of read losses and undetected sample misidentification errors. Even at mismatch error rates of 10−2 per base, 24-nt LDPC barcodes can be used to multiplex roughly 2000 samples with a sample misidentification error rate in the order of 10−9 at the expense of a rate of read losses just in the order of 10−6.

Usefulness of biological fingerprint in magnetic resonance imaging for patient verification.

The purpose of our study is to investigate the feasibility of automated patient verification using multi-planar reconstruction (MPR) images generated from three-dimensional magnetic resonance (MR) imaging of the brain. Several anatomy-related MPR images generated from three-dimensional fast scout scan of each MR examination were used as biological fingerprint images in this study. The database of this study consisted of 730 temporal pairs of MR examination of the brain. We calculated the correlation value between current and prior biological fingerprint images of the same patient and also all combinations of two images for different patients to evaluate the effectiveness of our method for patient verification. The best performance of our system were as follows: a half-total error rate of 1.59% with a false acceptance rate of 0.023% and a false rejection rate of 3.15%, an equal error rate of 1.37%, and a rank-one identification rate of 98.6%. Our method makes it possible to verify the identity of the patient using only some existing medical images without the addition of incidental equipment. Also, our method will contribute to patient misidentification error management caused by human errors.

Phenotypic polymorphism of Chrysomya albiceps (Wiedemann) (Diptera: Calliphoridae) may lead to species misidentification

Species identification is an essential step in the progress and completion of work in several areas of biological knowledge, but it is not a simple process. Due to the close phylogenetic relationship of certain species, morphological characters are not always sufficiently distinguishable. As a result, it is necessary to combine several methods of analysis that contribute to a distinct categorization of taxa. This study aimed to raise diagnostic characters, both morphological and molecular, for the correct identification of species of the genus Chrysomya (Diptera: Calliphoridae) recorded in the New World, which has continuously generated discussion about its taxonomic position over the last century. A clear example of this situation was the first record of Chrysomya rufifacies in Brazilian territory in 2012. However, the morphological polymorphism and genetic variability of Chrysomya albiceps studied here show that both species (C. rufifacies and C. albiceps) share very similar character states, leading to misidentification and subsequent registration error of species present in our territory. This conclusion is demonstrated by the authors, based on a review of the material deposited in major scientific collections in Brazil and subsequent molecular and phylogenetic analysis of these samples. Additionally, we have proposed a new taxonomic key to separate the species of Chrysomya found on the American continent, taking into account a larger number of characters beyond those available in current literature.

Maximum likelihood estimation for model Mt,α for capture-recapture data with misidentification.

We investigate model Mt,α for abundance estimation in closed-population capture-recapture studies, where animals are identified from natural marks such as DNA profiles or photographs of distinctive individual features. Model Mt,α extends the classical model Mt to accommodate errors in identification, by specifying that each sample identification is correct with probability α and false with probability 1-α. Information about misidentification is gained from a surplus of capture histories with only one entry, which arise from false identifications. We derive an exact closed-form expression for the likelihood for model Mt,α and show that it can be computed efficiently, in contrast to previous studies which have held the likelihood to be computationally intractable. Our fast computation enables us to conduct a thorough investigation of the statistical properties of the maximum likelihood estimates. We find that the indirect approach to error estimation places high demands on data richness, and good statistical properties in terms of precision and bias require high capture probabilities or many capture occasions. When these requirements are not met, abundance is estimated with very low precision and negative bias, and at the extreme better properties can be obtained by the naive approach of ignoring misidentification error. We recommend that model Mt,α be used with caution and other strategies for handling misidentification error be considered. We illustrate our study with genetic and photographic surveys of the New Zealand population of southern right whale (Eubalaena australis).

Erratum to: Opportunistic predation by Crested Owl Lophostrix cristata upon Seba’s Short-tailed Bat Carollia perspicillata

As shown in figure 1a and b, the owl identified as a Crested Owl Lophostrix cristata (Daudin, 1800) in Rocha & López-Baucells (2014) is indeed a Mottled Owl Strix virgata (Cassin, 1849). Therefore, all results and conclusions in the paper pertaining to L. cristata refer in fact to S. virgata. Both the authors and the Editor in Chief of Revista Brasileira de Ornitologia thank readers for noticing this inconsistency and apologize for this misidentification error.

Do Anglers Know What They Catch? Identification Accuracy and Its Effect on Angler Survey‐Derived Catch Estimates

AbstractAngler surveys are reliant on the ability of anglers to accurately report various aspects of their fishing trips. Misidentification of sport fishes has been postulated as a source of error among angler surveys but has received little attention. We evaluated the overall ability of anglers to identify sport fishes common to Ohio and the potential impacts on catch estimates derived from angler surveys. During angler surveys conducted on lakes and reservoirs (2007, n = 34) and Ohio River tailwaters (2010, n = 3), anglers were presented with artist‐rendered images of 18 different sport fishes common to these waters. Anglers (lake and reservoir, n = 2,442; Ohio River, n = 458) were asked to identify sport fish by their common name. On average, anglers correctly identified sport fishes 42% of the time, but accuracy varied widely among species (lake and reservoir, range = 4.4–85.1%; Ohio River, range = 9.4–71.8%), with greater accuracy exhibited for the most common species (e.g., largemouth bass Micropterus salmoides). However, by grouping angler responses into species groups (e.g., black bass Micropterus spp., sunfish Lepomis spp.) angler identification of sport fishes was more reliable (lake and reservoir, mean = 83.4%; Ohio River, mean = 83.8%). Using these estimates of angler accuracy, we simulated the potential error in sport fish catch estimates using data from an angler survey conducted at one Ohio reservoir. These results suggested that misidentification error may result in a substantial error in catch estimates. A survey of North American fisheries management agencies regarding angler surveys and angler sport fish identification revealed that the majority of agencies group similar species for analysis and cited species misidentification by anglers as the primary reason for doing so.Received April 19, 2012; accepted August 17, 2012

Predictors of hospitalized patients’ intentions to prevent healthcare harm: A cross sectional survey

BackgroundPatients can play an important role in reducing healthcare harm but little is known about the factors that may affect patients’ willingness to participate. In order to encourage the ‘active’ patient it is critical that we gain a deeper understanding of the antecedents of safety-relevant behaviours. Doing this will enable the implementation of effective interventions aimed at supporting patients to work with healthcare professionals in ensuring safe care. ObjectiveTo examine predictors of patients’ intentions to engage in two safety behaviours: (1) reminding healthcare staff to wash their hands and; (2) notifying healthcare staff if they are not wearing a hospital identification bracelet. DesignCross-sectional survey study. ParticipantsA purposive sampling method was employed to recruit 80 medical and surgical hospital inpatients aged 18–80 (mean 48) from one inner city London teaching hospital. MethodsA 42 item survey that measured the extent that patients’ control beliefs, behavioural beliefs, normative beliefs and perceived susceptibility and severity towards a hospital-acquired infection or a misidentification error could predict their intentions to ask doctors/nurses about their hand washing compliance or notify doctors/nurses if they are not wearing a hospital identification bracelet. Data was analysed using multiple regression analysis. ResultsControl beliefs, normative beliefs and perceived severity were the strongest predictors of patients’ intentions to participate in both behaviours. The regression models accounted for a smaller percentage of the variance in patients’ intentions to ask doctors/nurses if they have washed their hands (42%/37%) than notifying staff if they were not wearing an identification bracelet (54%/56%). ConclusionsIf patients understand why a behaviour is beneficial, they perceive it as acceptable to participate in and that they have control over the decision to engage in the behaviour, we hypothesise that more patients will intend to participate in that behaviour. When designing interventions aimed at encouraging the participation of patients in promoting their own safety, consideration should be given to the potential influence of patients’ control beliefs, normative beliefs and perceived severity of errors on their intentions to participate.

Estimating survival in photographic capture–recapture studies: overcoming misidentification error

Summary1. For many species, noninvasive photographic identification offers a powerful and cost–effective method for estimating demographic parameters and testing ecological hypotheses in large populations. However, this technique is prone to misidentification errors that can severely bias capture–recapture estimates.2. We present a simple ad hoc data conditioning technique that minimizes bias in survival estimates across all rates of misidentification. We use simulated data sets to characterize trade‐offs in bias, precision and accuracy of survival estimators for a range of misidentification probabilities, sampling intensities, survival rates and population sizes using this conditional approach.3. Misidentification errors resulted in mean survival estimates that were negatively biased by as much as −24·9% when errors were ignored. Applying the conditional approach resulted in very low levels of bias across parameter space. However, the main cost of conditioning is a loss of precision, which was particularly severe at low sampling intensities. Overall, the conditional approach was superior to the nonconditional approach [in terms of root mean square error (RMSE) of survival estimates] in 51% of the parameter combinations that we explored.4. We apply the data conditioning technique to a 3‐sample capture–recapture data set compiled from 2551 images of a migratory wildebeest, Connochaetes taurinus, population in northern Tanzania. We estimate the false rejection rate (i.e., the probability of failing to match two photographs of the same individual) using a test set of ‘known‐identity’ individuals. With this information, we compare survival estimates derived from conditioned data ( = 0·698 ± 0·176), unconditioned data ( = 0·706 ± 0·121) and simulated data to illustrate some of the key considerations for deciding whether to apply a conditional approach to a photographic data set.5. These analyses demonstrate that ignoring misidentification error can lead to substantial bias in survival estimates. When sampling intensity and misclassification error rates are both relatively high, use of our conditioned data approach is preferred and yields survival estimates with lower RMSE. However, when sampling intensity and misclassification error are both small, the standard approach using unconditioned data yields smaller RMSE.