Inter-observer and intra-observer differences in measuring body length: a test in the common lizard, Zootoca vivipara

Abstract

Abstract. The snout-vent length (SVL), a conventional measure of overall body size in lizards and snakes, is used in a wide variety of ecological, evolutionary, and taxonomical studies. Trends in SVL variation are often analysed using data from several researchers (observers), but possible confounding effects due to inter-observer differences in measurement protocols have never been appropriately examined. This study reports inter-observer biases between eleven herpetologists who measured the same specimens of the Eurasian common lizards (21 adult specimens were examined by eight observers and additional 192 specimens by two observers). Intra-observer bias over time (1.5-15 months between measuring sessions) was also estimated. In the vast majority of comparisons, mean difference between the first author and another observer varied from −1.0 to +0.8 mm, or from −1.9 to +1.6% if expressed as a percent of the specimen's SVL value. Some non-regular effects of sex and study sample on the studied bias were revealed, and their possible reasons are discussed. We are advising the researchers who intensively collect SVL and other morphometric data to consider testing intra-observer and inter-observer biases and to establish etalon samples available for re-examinations.

Comparison of manual measurements and computer-assisted image analysis in fish morphometry

“…In contrast with other studies, we chose nonprofessional morphologists as researchers since we wanted to test both methods (manual measurement and the LUCIA system) without impact of personal experience. This probably led to bigger biases and differences obtained among several manual measurements as compared to, for example, the study of Roitberg et al (6). The mean of relative biases ranged from -4.27% to 9.48% (Table 2).…”

“…Even if experienced researchers follow standard methodology, subjective influence on the measurement due to slightly different personal habits cannot be excluded. This implies that repeated measurements by 1 researcher on identical material should grant more aligned data in comparison to measurements gained from identical material by 2 or more researchers (6,7,18). In contrast with other studies, we chose nonprofessional morphologists as researchers since we wanted to test both methods (manual measurement and the LUCIA system) without impact of personal experience.…”

“…Utilization of these types of instruments helps to make measurements more precise and less loaded with errors that commonly arise when raw data are rewritten to paper. In many scientific studies, different authors compared their morphological analysis in the same species to each other, especially in cases where the species has wide geographic range of distribution and high morphological variability can be assumed (6). Accordingly, proper methodology, repeatability, and accuracy are fundamental.…”

“…Several works that deal with measurement error of researchers or with accuracy of obtaining metric characters already exist (6,(17)(18)(19)(20)(21), but the comparison of manual measurement and computer-assisted measurement from digital images is lacking.…”

Description of a New Species of the GenusAselliscus(Chiroptera, Hipposideridae) from Vietnam

“…For this reason, the results obtained by our morphological comparison using morphometric measurements available in the literature may not be accurate due to the examined characters containing potential inter-observer variability (Lee, 1990;Yezerinac et al, 1992;Palmeirim, 1998). Indeed, the magnitude of differences between measurements taken by different and those taken by the same observers are known to differ considerably from character to character (Lee, 1990;Palmei rim, 1998;Hayek and Heyer, 2005;Roitberg et al, 2011). For small sized bats, Palmeirim (1998) considered that the both the intra-and the inter-observer variability of measurements of several craniodental characters is adequate, and morphological comparisons using these characters from different sources can be performed with reasonable confidence.…”