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

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).