Forced-choice and free-choice trials in response priming with moving primes

Published: 11 March 2021| Version 2 | DOI: 10.17632/3btrkch437.2
Christina Bermeitinger


Basically, we conducted a response priming experiment with moving row-of-dots-primes (motion direction: leftwards, rightwards, neutral [to the center or the boarders of the screen; inwards or outwards, respectively]; see Bermeitinger, 2013). In forced-choice trials, the target was a directional arrow (<< or >>) which has to be classified (leftwards or rightwards). In free-choice trials, the target was an ambigous arrow composition (<> or ><) and subjects should freely choose the left or right response. Hypothesis: positive compatibility/congruency effects (PCE) with the shorter stimulus onset asynchrony (SOA) and reduced, perhaps even negative compatibility/congruency effects (NCE) with the longer SOA. (SOA was varied between subjects, short SOA = 147 ms, longer SOA = 360 ms.) In general, we expected the same pattern of results in forced-choice and free-choice trials. Further, we set out to determine whether there are differences in the responses to trials with compatible neutral primes (i.e., inwards motions and >< targets; outwards motions and <> targets) and to trials with incompatible neutral primes (i.e., inwards motions and <> targets; outwards motions and >< targets). Last but not least, we analyzed differences between >< and <> targets, focusing on leftwards and rightwards prime motions in free-choice trials. According to Cole and Kuhn (2010), responses should be faster or slower depending on the number of attentional turns needed between prime and response. For <> targets, we thus expected faster and more likely prime-congruent (than prime-incongruent) responses. In contrast, a >< target should result in faster and more probable prime-incongruent responses. Results: In forced-choice trials: Compatible trials = primes and targets move/point in the same direction; incompatible trials: primes and targets move/point in opposite directions; In free-choice trials, congruency was determined by the responses given, i.e. congruent response = left/right button press after leftwards/rightwards motion; incongruent response = left/right button press after rightwards/leftwards motion. Compatibility/congruency effect = response time in incompatible/incongruent trials - response time in compatible/congruent trials. Stimulus Onset Asynchrony (SOA) = time from the beginning of the prime display to the beginning of the target display We found (non-significant) PCEs in the short SOA (for forced-choice and free-choice trials) and (significant) NCEs in the longer SOA (for forced-choice and free-choice trials). There were no congruency effects regarding neutral conditions (i.e. <> and outwards motion; >< and inwards motion). For directional motions in free-choice trials, we found (non-significantly) faster prime-congruent than prime-incongruent responses after >< targets and the reverse (significant) result after <> targets (which is contrary to Cole & Kuhn, 2010).


Steps to reproduce

For presentation of the material and data acquisition, we used E-Prime software (version 2). We used CRT screens with a refresh rate of 75 Hz. Exclude participants who were outliers regarding their mean error rate in forced-choice trials (i.e. more than 3 interquartile ranges above the third quartile; Tukey, 1977). Notes to the files: Dependent variable Reaction Time can be found in the column "Turget.RT". Dependent variable Accuracy can be found in the column "Turget.ACC". Given response can be found in the column "Turget.RESP" (1 = left response, 3 = right response). The target presented in the trial can be found in column "Target". List1 (column "Running") = practice block. "Procedure" indicates the prime's motion direction: links = leftwards; rechts = rightwards; neutralInnen = neutral inwards; neutralAussen = neutral outwards SOA can be seen from the last number in the column "ExperimentName": _150 = SOA 147 ms; _360 = SOA 360 ms For RT analysis, errouneous responses have to be excluded, as well as outlying responses (i.e. responses that were 1.5 interquartile ranges above the third quartile with respect to the individual distribution (Tukey, 1977), were above 1’500 ms, or were below 200 ms were discarded for the forced-choice as well as the free-choice trials).


Cognitive Psychology