SUMMARY OF THE RESULTS Sample Clauses

SUMMARY OF THE RESULTS. We conclude the results section with a summary of our findings as they pertain to our three re- search hypotheses. We find strong support for Hypothesis 1. The higher bonus amounts do in- duce higher orders. We do not find general support of Hypothesis 2. The relationship between the order amount and the length of the review period does not generally confirm theoretical pre- dictions. In fact, what we find is evidence that for medium and high bonuses, longer review pe- riods induce higher orders, independent of the bonus. For low bonuses, orders under both review periods tend to be above their theoretical benchmarks and not different from one another. Of course since our laboratory study included review periods of only two different lengths, the gen- erality of this conclusion is limited. For medium bonuses (B=25) the optimal order for T=2 is 70 and for T=8 it is 86, so the fact that the longer review period induces higher ordering levels is consistent with the theoretical prediction. But for high bonuses (B = 50) higher orders in the T=8 conditions are the opposite of the theoretical prediction. In those treatments, participants start by ordering too low and adjust in the direction of the optimal order, but because of the high probability of observing the fill rate of 100% in the T=2 treatment, they do not adjust their orders sufficiently. In the T=8 treatment, the feedback is more reliable, and consequently participants are able to come closer to the opti- mal order. One way to bring a common framework to our results is by noting a combination of be- havioral effects. As we can see from Table 4, a bonus impacts players’ orders in two ways: (1) in the initial order, and (2) in the direction of the order-up-to level adjustment. The anchoring effect on the initial order may be influenced by the availability heuristic (▇▇▇▇▇▇▇ and ▇▇▇▇▇▇▇▇ 1973) because a lump-sum bonus is more “available” than the holding or backorder costs, caus- ing players to adjust orders accordingly. Because higher bonuses also increase optimal order-up- to levels, this has an impact on the location of players’ orders relative to theoretical predictions. Players seem to over-adjust the order-up-to level amount when given small bonuses and under- adjust when given higher bonuses. When no bonus is present no adjustment takes place. This brings us to Hypothesis 3 (Anchoring and insufficient adjustment) for which we do find some support. Interestingly, our data suggests that the anchor is ...

SUMMARY OF THE RESULTS. For all measured dependent variables (except deviation) the main effects of tool, velocity, and movement were significant at the 5 % significance level (Table 6). The main effect of location was significant for pleasure and arousal. At the starting point of the approach toward the robot participants experienced higher levels of pleasure and lower levels of arousal. Participants moved closer to the robot, experienced higher levels of pleasure and lower levels of arousal, perceived the robot as safer, and had higher intentions to collaborate with the robot when the robot moved slowly, used a safe tool, and moved up/down, left/right, or performed a circular left/right movement. Participants stopped earlier in front of the robot, experienced lower levels of pleasure and higher levels of arousal, perceived the robot as less safe, and had lower intentions to collaborate with the robot when the robot used a dangerous tool, moved quickly, and performed forward/backwards, random, and circular forward/backwards movements. The effect of the tool was by far the strongest, which could be due to the fact that the robot was using a really dangerous tool – a large kitchen knife – which can normally elicit higher arousal. The effect of velocity was, on average, the second largest effect. The tool of the robot (safe versus dangerous) and velocity of the robot (slow versus fast) were found to be the most important factors of human-robot interaction. The third largest effect was the main effect of movement type, which also should not be neglected when designing a human-robot interaction. Participants felt less safe, experienced higher levels of arousal and unsafety and lower level of pleasure and intention to collaborate when the movements were random and forward/backward. Random movement is probably perceived as more unpredictable and unsmooth. Forward/backward movement can represent a movement that mimics the movement of an attacking or hitting something and therefore provokes unpleasant feelings. The most frequent significant two-way interactions were between velocity and movement and between location and tool. All two-way interactions were weak, especially compared to main effects. There was just one significant three-way interaction (i.e., between location, tool and velocity on pleasure) and no significant four-way interaction. Location / / / ▯ ▯ Tool type ▯ ▯ ▯ ▯ ▯ ▯ Velocity ▯ ▯ ▯ ▯ ▯ ▯ Movement type ▯ ▯ ▯ ▯ ▯ ▯ LxT / / / ▯ ▯ ▯ ▯ LxV / / / ▯ LxTxV / / / ▯ TxM ▯ ▯ VxM ▯...

SUMMARY OF THE RESULTS. In Chapter 4, the development of the Dutch posture-verb progressive construction was analyzed in detail based on the corpus data. The results were summarized in Table 22 in Chapter 4, repeated here as Table 1. Table 1. Development of the Dutch posture-verb progressive construction Stage 1 [pre-1200] Biclausal/bipredicative or monopredicative S PVfin Advloc en(de) (S) V2fin Stage 2 [1200–1600] Biclausal/monopredicative S PVfin Advloc en(de) V2fin Stage 3 [1600–1700] Biclausal/monopredicative S PVfin Advloc en(de) V2fin Monoclausal/monopredicative S PVfin Advloc te V2inf Stage 4 [1800–now] Monoclausal/monopredicative S PVfin Advloc te V2inf As discussed in 4.5.2., the data are interpreted as indicating that one construction replaced another. These two constructions do not seem to be historically related. The older construction with en(de) as a connector was mostly attested in Middle Dutch (13th–16th century, cf. Stage 2 in Table 1). This construction was gradually superseded by the newer construction with te as a connector, which increased in frequency mostly in the 17th century (Stage 3 in Table 1). From around the 18th century, the te construction became the only posture-verb construction with a progressive meaning in the language (Stage 4 in Table 1). This situation seems to have remained stable to the present day: the proportions of locative and temporal modification and the most frequent types of second verb are comparable between the 18th century and Modern Dutch. en(de) construction and the unambiguously monoclausal te construction, the two constructions share some commonalities. For example, both are strongly associated with a locative meaning, as reflected by the regular occurrence of locative modifiers. The stability of the two constructions in terms of their locative meaning can also be seen in the semantics of the second verb, such as the fact that the second verb must describe an activity compatible with the particular posture; that is, it appears that the stable and salient locative semantics of the posture verbs has prevented the second verb from becoming semantically more diverse. This semantic profile seems to be found with both the en(de) and the te construction, and it may have been precisely this overlap in semantic profile that made the two constructions competitors, with one of them eventually ‘taking over’. For German, Chapter 5 reports the current status of the posture-verb construction in the present-day language. According to the an...