Fear conditioning is a laboratory paradigm commonly used to investigate aversive learning and memory. In context fear conditioning, a configuration of elemental cues (conditioned stimulus, CS) predicts an aversive event (unconditioned stimulus, US). To quantify context fear acquisition in humans, previous work has used startle eye-blink responses (SEBR), skin conductance responses (SCR) and verbal reports, but different quantification methods have rarely been compared. Moreover, it is unclear how to induce, and measure context fear memory retention over several days. First, we used a semi-immersive virtual reality paradigm. In two experiments, we found successful declarative learning and memory retention over seven days, but no evidence of conditioned responses. Next, we used a configural fear conditioning paradigm with five static room images as CSs in two experiments. Besides successful declarative learning and memory retention after seven days, SCR and pupil dilation to CS onset differentiated CS+/CS- during acquisition training, and SEBR and pupil dilation differentiated CS+/CS- during the recall test, with medium to large effect sizes for the most sensitive indices (SEBR: Hedge’s g = 0.56 and 0.69; pupil dilation: Hedge’s g = 0.99 and g = 0.88). Our results demonstrate that with a suitable experimental paradigm, context fear memory retention can be demonstrated over seven days, and we provide robust and replicable measurement methods.