Inducing a Blind Spot: Blinding Data Collectors in an Investigation of Experimental Pain

Abstract

Objective: Methods of blinding have been infrequently investigated. This study assessed data collector blinding within an investigation of endogenous pain modulation. Design: Participants (N=33; 52% women) were randomly assigned to an order of an exercise bout with intramuscular control injections, quiet rest with intramuscular algesic injections, and a control condition. Data collectors (N=4; 50% women) recorded participants' pain responses and then left the room before the injection and exercise procedures were administered by other personnel. Immediately after the procedures, the data collectors returned to collect additional pain responses and record their suspicions regarding the assigned condition, their confidence level in their suspicions (0-100mm scale), and the reason/s for their suspicions. Results: Data collectors correctly identified control and algesic injections in 90.4 and 73.1% of the sessions, respectively, and quiet rest and exercise in 94.2 and 69.2% of the sessions, respectively. The confidence of the data collectors in their suspicions was 60.17mm (SD=34.54) for injections and 62.19mm (SD=34.95) for exercise. However, data collectors only had correct suspicions and confidence of at least 90mm for 26.2% of the injections and 29.4% of the exercise bouts. Participants' pain responses were the primary reason for data collectors' suspicions. Neither the amount of experience nor the sexual composition of the participant-data collector dyad influenced the blinding. Conclusions: Collection of participants' pain responses led to frequent episodes of unblinding. However, it may be misleading to only consider the frequency of correct suspicions as successful or unsuccessful blinding.