Fear of humans can have a major impact on farm animals, not only in terms of their welfare, but also their productivity. A measure of ‘fear of humans’ is included in most welfare assessment protocols for many farmed species. Welfare indicators need to be valid, reliable and repeatable, but also need to be practical. While fear of humans in sheep has been studied using a range of behaviour tests, a validated practical test that can be used on-farm in extensive farm settings has not yet been developed. It has been suggested that the exit speed after close confinement is a measure of fear, and may be related to fear of humans if sheep are released in the presence of a human. We compared the behaviour of 85 slaughter lambs in a novel arena test, an isolation test, a human avoidance test and a human approach test (with other sheep located behind the human)  with the exit speed out of an enclosed stall. Spearman’s rank correlations of several measurements from each test found exit speed was negatively related with vocalisations in the isolation box (r=-0.223; p=0.04) and positively correlated with movement in a novel arena (r=0.336; p=0.002) and distance from the human at the end of the human approach test (r=0.222; p=0.04), but not with other measures. While the positive correlation between exit speed and the human approach test is promising, the correlation is weak. It appears that each test measures a different aspect of fear and more research is required to develop and identify a valid, practical test that specifically measures fear of humans in extensively-managed sheep.