Boldness: are open field and startle tests measuring the same personality trait?

Boldness, the willingness of individuals to engage in risky behaviour, is one of the most studied personality traits. It has been measured using a variety of tests; however, measuring a behaviour using different assays may lead to a jingle fallacy. The few studies that have attempted to determine whether these different assays are comparable have produced mixed results. A lack of repeatability between boldness measures under standardized and natural conditions may be the source of this variation. Here, we tested whether risk-taking behaviour of free-living African striped mice, Rhabdomys pumilio, measured in a laboratory using open field tests is comparable with measures of risk-taking behaviour from startle tests. These measures were then compared with measures of risk-taking behaviour obtained from equivalent open field and startle tests performed on the same individuals under natural conditions. During open field tests, we assessed the time an individual spent away from the wall of the arena (in the laboratory) or from its nest (in nature). During startle tests, we measured the latency to re-emerge from either a protective box (in the laboratory) or a nest (under natural conditions) after an individual was scared away. Our results showed that risk-taking behaviour measured using the open field and the startle tests were repeatable within the same context (tested twice per assay) and across contexts (laboratory, nature). However, open field measures of risk-taking behaviour were not correlated with startle measures, supporting the idea that the two tests are assaying different personality traits. The variations in the outcome observed in the two assays probably resulted from the presentation of threatening stimuli in the startle test. We propose that, at least in striped mice, the startle test is more suitable for measuring anxiety than boldness.