The Necessity Of Collecting Equal Baseline Reaction Times In Priming Experiments: A Step That Cannot Be Omitted

This paper aims to study the necessity of collecting baseline reaction times (RTs) in priming experiments by investigating whether equal word frequency leads to equal baseline RTs.In a priming experiment there are at least two groups of materials, an experimental group and its corresponding control group. There is a priming effect if reaction times to targets in the experimental group are shorter than those to targets in the control group. The priming effect is taken as evidence for the special prime-target relationship in the mental lexicon. That is, prime-targets pairs in the experimental group are more closely located or have stronger connections than prime-targets pairs in the control. However, a possible confounding factor for this conclusion is that the RT difference between the two groups may be attributed to the targets themselves, not to the prime-target relations. The best and most direct way to exclude the influence of this factor is to select targets in such a way that no RTs difference exists in the targets between the two groups when they are not preceded by primes, e.g., to make sure that there is no significant difference between the baseline RTs to the targets in the two groups. But there are studies that replaced this step to collect baseline RTs by selecting two groups of targets that are not different in frequency, on the assumption that equal word frequency leads to equal baseline RTs.We tried to test this assumption in the task of lexical decision in both the L1 and the L2 of our subjects (Chinese learners of English). And we found that,1) For a second / foreign language, matching of frequencies for two lists of words do not lead to equal baseline RTs for the two lists in most cases, no matter how closely matched they are. But occasionally if the frequencies are high, it may lead to insignificant RT differences.2) For the first language, matching of frequencies for two lists of words do not necessarily lead to equal baseline RTs. If the two lists of words are closely matched (one-to-one match) AND if they are words of high frequencies, they lead to equal baseline RTs in most cases. But otherwise it may not be the case.The results show that the step to collect baseline RTs cannot be omitted in priming experiments, esp. for the subjects' second language.