When the sense of fluency triggers an attentional bias

Abstract

Spatial attention can be captured automatically by an exogenous stimulus (e.g., digital interruption) or by an endogenous stimulus (e.g., valence of the stimulus). In this study, we investigated whether a non-perceptual characteristic (e.g., sense of fluency) has an impact on attention. To this end, we used the conceptual fluency paradigm developed by Whittlesea combined with the dot-probe task developed by MacLeod et al. In three experiments, we measured the response times for each experimental situation (i.e., Valid and Non-valid situations). At each trial, participants were presented in three consecutive displays on a screen: (1) an incomplete and predictive sentence stem; (2) a pair of words, one of which was semantically compatible with the previous sentence stem; and (3) a circle appeared at the spatial location of one of the words. Then, participants had to perform a Go (i.e., a filled circle) and No-go (i.e., an empty circle) task. The analysis found that response times were significantly faster when the Go stimulus appeared at the same location as the semantically compatible word (i.e., Valid situations). Overall, our results show that the sense of fluency triggers attentional capture. Thus, they replicate those of Gardner et al. using another experimental paradigm. Our finding might be helpful to better understand the consequences of digital interruptions on behavioural performance.

Keywords

Attention conceptual fluency discrepancy dot-probe task interruption

Public significance statement

In this article, we investigate the relationship between sense of fluency and attention. Three experiments demonstrate that the sense of fluency captured attention manipulating the compatibility between predictive sentence stem and words. The sense of fluency created by the context seems to have an effect on the attention of individuals. These observations can help to better understand the relationship between fluency and attention, and the consequences of digital interruptions on human performance.

Introduction

Interruptions and their consequences are classically examined in terms of attentional processes (Altmann & Trafton, 2007). In this sense, interruptions are exogenous stimuli that capture people’s attention. Several researchers showed that when an interruption occurs, our attention automatically shifts to the stimulus that causes it (Addas & Pinsonneault, 2015). Other studies showed that individuals cannot ignore the sudden appearance of a stimulus in their visual field (Yantis & Jonides, 1984). The attentional capture can be associated with a non-perceptual stimulus characteristic (e.g., valence of the stimuli; Lien et al., 2020). When we present two stimuli simultaneously, attention will be allocated more quickly to the spatial location containing emotionally non-neutral stimuli (Carretié, 2014; MacLeod et al., 1986; van Rooijen et al., 2017). Furthermore, Gardner et al. (2020) showed that word processing fluency induces involuntary attentional shift. In that context, the purpose of the current study was to investigate whether the sense of fluency could capture attention, using a combination of the conceptual fluency paradigm developed by Whittlesea (1993) and the dot-probe task (MacLeod et al., 1986).

The study of attentional capture is part of a long-standing debate: is attentional capture the response to stimulus properties (Belopolsky et al., 2007) or to subject properties (Folk et al., 1992)? For the former, it would be the salience of the stimulus that would determine attentional capture (i.e., bottom-up). For the second, it would be the individual’s attentional state that would determine attentional capture (i.e., top-down). Authors have shown that the degree of activation of the attentional state could determine the capacity of individuals to detect a target (Wolfe et al., 2005). Specifically, using the dot-probe task (MacLeod et al., 1986), it has been shown that individuals’ emotional state impacts attention (Arndt & Fujiwara, 2012). In this task, stimuli are presented in pairs; usually one is neutral and the other emotional, with each stimulus on a different side of the screen. Immediately after they disappear, a target is presented at the location previously occupied by one of the stimuli. Target detection or discrimination is an emotionally neutral task. If attention is more likely to be directed towards the emotional stimulus, the target at that spatial location will be detected more quickly than it appears at the location of the neutral stimulus. Usually, situations in which the target appears at the spatial location of the emotional stimulus are defined as valid, while situations in which it appears at the position of the neutral stimulus are defined as non-valid. The fact that there is a neutral response (i.e., button pressing) to a neutral stimulus is a significant advantage because it eliminates the possibility of response bias (MacLeod et al., 1986).

More recently, Gardner et al. (2020) have shown that involuntary attentional shifting can be associated with a non-perceptual stimulus characteristics, namely, word processing fluency (i.e., the ease with which information is processed). Subjects were asked to focus their gaze on a fixation cross and maintain their gaze, even if words appeared around the cross. The results revealed that involuntary attention was increasingly directed to high-fluency words than low-fluency words. It should be noted that high-fluency words were real words, and low-fluency words were pseudowords. This study suggests that a non-conscious phenomenological dimension is able to capture attention (Gardner et al., 2020).

According to Neisser (1976), human perception is characterised by an interconnection between the environment, incoming sensory information, and prior knowledge. In a specific context, these co-activations would allow the emergence of predictions and the orientation of attention towards relevant stimuli (Bar, 2007). Thus, stimuli can be either expected or unexpected. Regarding expected stimuli, their presence and characteristics are predicted based on prior knowledge (Horstmann, 2015). If expectations are validated by the occurrence of an expected stimulus, then a sense of fluency will develop. If expectations are not validated, then a disruption of the sense of fluency will occur. As individuals are not aware of the result of this perceived discrepancy, they are highly likely to attribute this subjective feeling to the stimulus (Whittlesea, 2002; Whittlesea & Leboe, 2000; Whittlesea & Williams, 1998, 2001). From an evolutionary perspective, individuals’ attention should be focused on surprising (i.e., unexpected) stimuli (Wang et al., 1994). However, studies have shown that familiar (i.e., expected) stimuli capture more attention than unfamiliar stimuli (Christie & Klein, 1995; Gibson & Jiang, 1998).

To study the relationship between fluency and attention, we integrated these concepts into a single unifying framework. We combined the conceptual fluency paradigm developed by Whittlesea (1993) with the dot-probe task (MacLeod et al., 1986). Specifically, we manipulated target words that appeared after a predictive sentence stem. Two target words were presented simultaneously in two different locations on the screen for 1,000 ms. One word was semantically incompatible (i.e., unexpected stimulus) and the other one was semantically compatible (i.e., expected stimulus). Immediately after their disappearance, a target was displayed at each spatial location previously occupied by the two stimuli. Participants had to perform a Go/No-go task. A filled circle and an empty circle were, respectively, used as Go and No-go stimuli. In Experiment 1, we tested our protocol with an inter-stimulus interval (ISI) of 250 ms between the offset of the predictive stem and the presentation of the words. This design allowed us to test whether attentional capture was influenced by the sense of fluency. We expected that Valid situations would result in faster detections of Go stimulus than Non-valid situations. In Experiment 2, we removed the ISI used in Experiment 1 to test whether this delay had an impact on attentional capture in our protocol. We expected to observe a smaller difference in response time between Valid and Non-valid situations. In Experiment 3, we removed the presentation time of the predictive sentence stem which was previously fixed at 3,000 ms in Experiments 1 and 2. We aimed at evaluating whether this modification of the protocol will have an impact on the relationship between conceptual fluency and attention. Consequently, participants had all the time they needed to read the predictive sentence stem. Once the reading was over, participants had to trigger by themselves the presentation of the words. We expected to observe the same pattern of results as in Experiments 1 and 2.

Experiment 1

The goal of Experiment 1 was to verify that attention was influenced by the sense of fluency. To this end, we reproduced the paradigm of conceptual fluency (Whittlesea, 1993) using pre-tested predictive sentence stems. We associated it with a dot-probe task (MacLeod et al., 1986). We thought that the combination of the paradigm of conceptual fluency with a dot-probe task would provide a robust starting point for evaluating the impact of conceptual fluency on attentional capture. Since we are using Whittlesea’s (2002) protocol, we replicated it by taking into account that the ISI between predictive sentence stem and the two target words was experimentally manipulated. An ISI of 250 ms is necessary for the process of conceptual fluency to occur, according to Whittlesea (2002).

Method

Participants

A power analysis was conducted in XLSTAT (Addinsoft, 2019) for a repeated-measures analysis of variance (ANOVA) (between factors) with an expected medium effect size ( $η_{p}^{2} = . 05$ ) and a targeted power of 0.8. These analyses suggested a sample size of at least 15 participants for each group, but in an effort to maximise the robustness of our experiment, we increased our sample size to 17. The latter was calculated for a theoretical minimum effect size of interest for a first study on a topic (Anderson et al., 2017).

Accordingly, 17 volunteers were recruited (M = 24.00 years, SD = 4.70). All participants received some general information about the respective studies and signed a statement of informed consent. They were all native French speakers and were tested individually in an isolated room. All had normal or corrected-to-normal vision. French ethical approval was given for this study.

Materials

The stimulus consisted of 26 predictive sentence stems (e.g., “She cleaned the kitchen floor with a. . .”). Each of them had both an associated semantically compatible word (SCW) (e.g., “BROOM”) and a semantically incompatible word (SIW) (e.g., “FRAME”). All words used were 4–11 letters long.

To select the predictive sentence stems, we first conducted a pre-test involving 15 subjects. We presented the latter with incomplete sentence stems, and for each one of them, they had to produce a word that would fit the best with it. Consecutively, we only selected the incomplete sentence stems that presented the highest predictive association with an SCW. In a second pre-test, for each predictive sentence stem and their corresponding SCW, participants (N = 15) were asked to rate the association between them. Responses were made using a Likert-type scale from 1 (strongly disagree) to 6 (strongly agree). Subsequently, we randomly selected 26 sentences among those whose association was rated in average greater than 4. Furthermore, participants from the pre-tests were not allowed to take part in Experiment 1.

Thus, contrary to Whittlesea’s (1993) study, words that are semantically incompatible with the predictive sentence stem are not semantically plausible. We wanted to avoid any inhibition/suppression effect (Gernsbacher et al., 2001) that could affect the performance (Carretti et al., 2014; Syssau et al., 2000). The word pair presented on each trial was balanced in terms of the number of letters and frequency of use.

Design

Regarding our experimental protocol, we have 2 × 2 design: target position (right or left) and SCW position (right or left). However, the variables were counterbalanced and were analysed in a combined way for the response times in the four modalities.

We adopted a repeated-measures ANOVA: Valid situation right, Valid situation left, Non-valid situation target right, or Non-valid situation target left. All trials were counterbalanced.

Procedure

Participants were seated in front of a 13-inch laptop placed at the centre of a table and aligned with the median plane of their trunk. Each trial started with the presentation of a predictive sentence stem at the centre of the screen for 3,000 ms (e.g., “She cleaned the kitchen floor with a. . . .”) This duration was adjusted to give participants time to read the sentence, following the rules provided by Gernsbacher and Faust (1991). When the stem disappeared, the screen went blank for 250 ms before the target words (e.g., “BROOM” and “FRAME”) were presented. The latter were displayed in uppercase, with one word appearing on each side of the screen. The two target words were displayed for 1,000 ms before disappearing. Finally, either the Go (i.e., filled circle) or No-go (i.e., empty circle) stimulus appeared for 100 ms at the location of one of the target words. Participants were asked to ignore the empty circle (i.e., the No-go stimulus) but respond to the filled circle (i.e., the Go stimulus) as fast as possible by pressing the space bar. Each participant performed a total of 208 trials (Figure 1). The duration of the experiment was approximately 35 min.

Figure 1.

The figure illustrates a schematic overview of the procedure used in each trial for Experiment 1. The sequence of several screen images depicts the time of each event constituting a trial.

Results

To determine whether there were differences in the experimental situation across ISI conditions, we performed a repeated-measures ANOVA. We additionally reported the estimated Bayes factors (BF₁₀) as revealed by comparable Bayesian statistics using JASP software (JASP Team, 2019). We relied on BF₁₀ to interpret our results. Null-hypothesis (H0) significance tests and their accompanied p-values have several limitations. More reliable alternative approaches such as BF₁₀ have been proposed (Schönbrodt & Wagenmakers, 2018). Following the recommendations of Schönbrodt and Wagenmakers (2018), it seems expensive and time-consuming to gather strong evidence for H0 in the fixed-n design with a predetermined sample size. The authors suggest relaxing expectations for H0, aiming at strong support for H1. This reduces the likelihood of providing misleading and inconclusive evidence (Schönbrodt & Wagenmakers, 2018). BF₁₀ is the ratio of p(D∣H1), the probability of observing the data under the alternative hypothesis, and p(D∣H0), the probability of observing the data under the null hypothesis. Thus, the BF₁₀ is a measure of the probability that the data match the alternative hypothesis rather than the null hypothesis (Rouder et al., 2009). Values greater than 1 indicate an increased evidence for H1 relative to H0, and in the opposite way, values lesser than 1 indicate a decreased evidence for H1 relative to H0. We used the guidelines proposed by Lee and Wagenmakers (2013) and adjusted from Jeffreys (1961) to interpret BF₁₀ (Table 1).

Table 1.

Interpretation of Bayes factors (BF₁₀) (Lee & Wagenmakers, 2013; adjusted from Jeffreys, 1961).

BF₁₀	Support for hypothesis
>100	Extreme evidence for H1
30–100	Very strong evidence for H1
10–30	Strong evidence for H1
3–10	Moderate evidence for H1
1–3	Anecdotal evidence for H1
1	No evidence
1/3–1	Anecdotal evidence for H0
1/10–1/3	Moderate evidence for H0
1/30–1/10	Strong evidence for H0
1/100–1/30	Very strong evidence for H0
<1/100	Extreme evidence for H0

Regarding No-go trials, we found no significant difference in commission errors (going on No-go trials), F(1, 15) = 0.35, p = .57, BF₁₀ = 0.55. We therefore only analysed response times for Go trials (Table 2).

Table 2.

Mean response times (in milliseconds) and standard deviation (SD) as a function of the experimental situation in Experiment 1.

Experimental situations	Valid situation right	Valid situation left	Non-valid situation target right	Non-valid situation target left
M	307.61	308.18	340.57	344.06
SD	56.31	56.03	54.70	56.39

The results of Experiment 1 are depicted in Figure 2 and listed in Table 3. Diagrams show average response times for each experimental situation (i.e., Valid situation right and left and Non-valid situation with a target on the right and with the target on the left). For more details, valid situations refer to the situation where the Go stimulus was shown at the location of the SCW, which can be either on the right side of the screen or on the left side. Non-valid situations refer to the situation when the Go stimulus was shown at the location of the SIW. The Go stimulus can be located on the right side of the screen and the SCW on the left (i.e., Non-valid situation target right) and vice versa (i.e., Non-valid situation target left).

Figure 2.

Mean difference in response times according to experimental situations in Experiment 1.

Table 3.

Inferential statistics for the repeated-measures and Bayesian analyses of variance (ANOVA) and post hoc comparisons in Experiment 1.

			Repeated measures			Bayesian
			F	p	$η_{p}^{2}$	BF ₁₀
Effect	Experimental situations		6.45	<.001	.30	45.12
			Post hoc comparisons			Bayesian
			t	p _bonf		BF ₁₀
Experimental situations	Valid situation right	Non-valid situation target right	−3.04	<.05		2.25
		Non-valid situation target left	−3.36	<.01		5.87
		Valid situation left	−0.05	1		0.25
	Valid situation left	Non-valid situation target right	−2.99	<.05		2.65
		Non-valid situation target left	−3.31	<.01		9.73
	Non-valid situation target right	Non-valid situation target left	−0.32	1		0.30

This analysis yielded a significant main effect, F(3, 48) = 6.75, p < .001, $η_{p}^{2} = . 30$ , BF₁₀ = 45.12. For the post hoc comparisons, given that such repeated testing increases the chance of obtaining a significant effect, a Bonferroni correction was applied (Neter & Wasserman, 1974). The mean response time was lower in the Valid situations compared with Non-valid situations: Valid situation right (M = 307.61, SD = 56.31) versus Non-valid situation target right (M = 340.57, SD = 54.70), t(48) = −3.04, p_bonf ⩽ .05, BF₁₀ = 2.25; Valid situation right versus Non-valid situation target left (M = 344.06, SD = 56.39), t(48) = −3.36, p_bonf < .01, BF₁₀ = 5.87; Valid situation left (M = 308.18, SD = 56.03) versus Non-valid situation target right, t(48) = −2.99, p_bonf < .05, BF₁₀ = 2.65; Valid situation left versus Non-valid situation target left, t(48) = −3.31, p_bonf < .01, BF₁₀ = 9.73. Finally, a comparison of the two Valid situations (right and left) revealed no significant effect, t(48) = −0.05, p_bonf = 1, BF₁₀ = 0.25. The same result was observed when comparing the two Non-valid situations (target right and target left), t(48) = −0.32, p_bonf = 1, BF₁₀ = 0.30.

Discussion

In Experiment 1, we sought to test the influence of the sense of fluency on attention. The results provide statistical evidence of the impact of conceptual fluency on attentional bias, and support the earlier findings presented above. Response times were quicker in Valid situations than Non-valid situations. When the Go stimulus appeared at the same spatial location as the SCW (i.e., the Valid situations), the sense of fluency reduced response times. Conversely, in Non-valid situations, where the Go stimulus appeared at the same spatial location as the SIW, response times were longer. The latter observation indicates a discrepancy effect and attentional bias towards the SCW.

We propose that the attentional bias observed during this experiment is the consequence of the manipulation of conceptual fluency. When stems were presented, participants focused on the spatial location of the SCW. This made it easier for them to discriminate the Go stimulus from the No-go Stimulus, resulting in faster response times. In the influent situation, attention was focused on the location of the SCW. More generally, Valid situations would support a better sense of fluency, while Von-valid situations would create an attentional bias.

The presence of an ISI between the predictive sentence stem and the target words was the condition for observing the conceptual fluency effect. In Experiment 1, we took this ISI into account, according to the experimental framework of Whittlesea (2002). In his work, Whittlesea observed that the absence of an ISI of 250 ms led to the suppression of this effect. Following these observations, we tested our experimental protocol by removing the ISI in Experiment 2.

Experiment 2

We observed in Experiment 1 that attention was influenced by the sense of fluency. We thought that the innovative paradigm combination would provide a robust starting point for evaluating the relationship between conceptual fluency and attentional capture. In Experiment 2, we replicated this protocol by suppressing the ISI.

Method

Participants

As in Experiment 1, 17 volunteers were recruited (M = 24.29 years, SD = 4.90). All participants received some general information about the respective studies and signed a statement of informed consent. They were all native French speakers and were tested individually in an isolated room. All had normal or corrected-to-normal vision. French ethical approval was given for this study.

Materials

Materials were identical to those used in Experiment 1.

Design

Design was identical to those used in Experiment 1.

Procedure

The procedure and data analysis were the same as for Experiment 1, except for the ISI that was suppressed.

Results

The data analysis was identical to that in Experiment 1. To determine whether there were any differences in the experimental situation, we performed a repeated-measures ANOVA. Regarding No-go trials, we found no significant difference in commission errors between groups, F(1, 15) = 1.28, p = .28, BF₁₀ = 0.89. We therefore only analysed response times for Go trials. The results of Experiment 2 are depicted in Figure 3 and listed in Table 4. The pattern replicates that of Experiment 1, with faster response times in Valid situations (right and left) than in Non-valid situations (target right and target left). This analysis yielded a significant main effect of the experimental situations, F(3, 48) = 4.59, p < .01, $η_{p}^{2} = . 23$ , BF₁₀ = 44.77.

Figure 3.

Mean difference response times according to experimental situations in Experiment 2.

Table 4.

Inferential statistics for the repeated-measures and Bayesian analyses of variance (ANOVA) and post hoc comparisons in Experiment 2.

			Repeated measures			Bayesian
			F	p	$η_{p}^{2}$	BF ₁₀
Effect	Experimental situations		4.59	<.01	.23	44.77
			Post hoc comparisons			Bayesian
			t	p _bonf		BF ₁₀
Experimental situations	Valid situation right	Non-valid situation target right	−3.21	<.01		15.64
		Non-valid situation target left	−3.15	<.05		4.68
		Valid situation left	−0.07	1		0.25
	Valid situation left	Non-valid situation target right	−3.15	<.05		50.94
		Non-valid situation target left	−3.08	<.05		2.66
	Non-valid situation target right	Non-valid situation target left	−0.07	1		0.25

For the post hoc comparisons, given that such repeated testing increases the chance of obtaining a significant effect, a Bonferroni correction was applied (Neter & Wasserman, 1974). Mean response time was lower in the Valid situations compared with Non-valid situations: Valid situation right (M = 347.44, SD = 106.16) versus Non-valid situation target right (M = 373.01, SD = 126.12), t(48) = −3.21, p_bonf < .01, BF₁₀ = 15.64; Valid situation right versus Non-valid situation target left (M = 372.45, SD = 108.51), t(48) = −3.15, p_bonf < .05, BF₁₀ = 4.68; Valid situation left (M = 347.96, SD = 116.28) versus Non-valid situation target right, t(48) = −3.15, p_bonf < .05, BF₁₀ = 50.94; Valid situation left versus Non-valid situation target left, t(48) = −3.08, p_bonf < .05, BF₁₀ = 2.66. Finally, a comparison of the two Valid situations (right and left) revealed no significant effect, t(48) = −0.07, p_bonf = 1, BF₁₀ = 0.25. The same result was observed when comparing the two Non-valid situations (i.e., target right and target left), t(48) = −0.07, p_bonf = 1, BF₁₀ = 0.25.

Discussion

In Experiment 2, we wanted to test whether the absence of the ISI would have an impact on performance. Our results confirmed what we expected to observe: a smaller difference in response time difference between Valid and Non-valid situations compared with Experiment 1. However, the response pattern is the same as in Experiment 1. We found an impact of conceptual fluency on attention. The ISI does not seem to play a significant role in individuals’ performance. Response times were faster in Valid situations than in Non-valid situations, with or without ISI.

Contrary to previous studies, the effect of conceptual fluency on attentional capture does not appear only when there is an ISI between the predictive sentence stem and the target words (Whittlesea, 2002; Whittlesea & Williams, 2001). The presence of the ISI (250 ms) made the observation of attentional bias significantly more obvious, but it does not seem to be a necessary condition for the observation of this effect. When reading predictive sentence stems, participants appear to perceive a sense of fluency that leads to attentional capture towards the SCW. The specific context created by the experimental protocol in this study allows individuals to integrate information in a coherent manner.

Critically, both Experiments 1 and 2 imposed a sentence reading time. The results may suggest that individuals automatically processed the meaning of the target sentences and words. To test this, we wanted to see whether individuals’ engagement could affect response patterns.

Experiment 3

We observed in Experiments 1 and 2 that the ISI had no significant effect on performance. Only Valid and Non-valid situations highlighted the impact of conceptual fluency on attentional bias.

Experiment 3 aimed at testing whether our protocol would allow us to evaluate the effect of conceptual fluency on the attentional process in the absence of an imposed reading time for predictive sentences. To this end, we replicated our experimental protocol by removing the fixed reading time (3,000 ms) for the predictive sentence stem. If conceptual fluency is due to the effects of Valid and Non-valid situations on attentional bias, then participant engagement should enhance this effect.

Method

Participants

As in Experiments 1 and 2, 17 volunteers (M = 24.65 years, SD = 7.10) were involved in the experiment. All participants received some general information about the respective studies, and signed a statement of informed consent. They were all native French speakers and were tested individually in an isolated room. All had normal or corrected-to-normal vision. French ethical approval was given for this study.

Materials

Materials were identical to those used in Experiment 1.

Design

Design was identical to those used in Experiment 1.

Procedure

The procedure and data analysis were the same as for Experiments 1 and 2, except for the duration of appearance of the predictive sentence stem. In Experiment 3, the predictive sentence stem remained on the screen until the participant had read it and pressed the space bar. Each trial, therefore, began with the presentation of a predictive sentence stem, in the centre of the screen, for an unlimited period of time. The participant was asked to press the space bar, after reading the sentence aloud. As in Experiments 1 and 2, each participant performed a total of 208 trials.

Results

The data analysis was identical to that in Experiments 1 and 2. To determine whether there were any differences in the experimental situation, we performed a repeated-measures ANOVA. The results of Experiment 3 are depicted in Figure 4 and listed in Table 5. The pattern replicates that of the previous experiments. Removing the time for reading the predictive sentence, stem did not appear to change response times. This analysis yielded a significant main effect of the experimental situation, F(3, 48) = 11.35, p < .001, $η_{p}^{2} = . 42$ , BF₁₀ = 1751.91.

Table 5.

Inferential statistics for the repeated-measures and Bayesian analyses of variance (ANOVA) and post hoc comparisons in Experiment 2.

			Repeated measures			Bayesian
			F	p	$η_{p}^{2}$	BF ₁₀
Effect	Experimental situations		11.35	<.001	.42	1.75 × 10³
			Post-hoc comparisons			Bayesian
			t	p _bonf		BF ₁₀
Experimental situations	Valid situation right	Non-valid situation target right	−3.63	<.05		18.75
		Non-valid situation target left	−4.19	<.05		52.07
		Valid situation left	−0.41	1		0.27
	Valid situation left	Non-valid situation target right	−3.03	<.05		6.44
		Non-valid situation target left	−4.78	<.001		150.15
	Non-valid situation target right	Non-valid situation target left	1.62	0.75		0.74

Figure 4.

Mean difference response times for each experimental situation in Experiment 3.

For the post hoc comparisons, given that such repeated testing increases the chance of obtaining a significant effect, a Bonferroni correction was applied (Neter & Wasserman, 1974). Mean response time was lower in the Valid situations compared with Non-valid situations: Valid situation right (M = 321.82, SD = 89.05) versus Non-valid situation target right (M = 345.26, SD = 85.11), t(48) = −3.63, p_bonf < .05, BF₁₀ = 18.75; Valid situation right versus Non-valid situation target left (M = 352.11, SD = 82.61), t(48) = −4.19, p_bonf < .05, BF₁₀ = 52.07; Valid situation left (M = 324.68, SD = 82.23) versus Non-valid situation target right, t(48) = −3.03, p_bonf < .05, BF₁₀ = 6.44; Valid situation left versus Non-valid situation target left, t(48) = −4.78, p_bonf < .001, BF₁₀ = 150.15. Finally, a comparison of the two Valid situations (right and left) revealed no significant effect, t(48) = −0.41, p_bonf = 1, BF₁₀ = 0.27. The same result was observed when comparing the two Non-valid situations (i.e., target right and target left), t(48) = 1.62, p_bonf = 0.75, BF₁₀ = 0.74.

Discussion

In Experiment 3, we observed that the absence of a fixed reading time had no effect on response times. These findings converge with those of Experiments 1 and 2 and confirm that conceptual fluency affects attentional bias. In Valid situations, fluency translates into faster response times, and its manipulation has consequences for performance, notably attention capacities. When there is a discrepancy (i.e., a Non-valid situation), response times increase. As attention is focused on the SCW, when the Go stimulus (i.e., the target) appears at the spatial location of the SIW, information processing is slower. This observation suggests that there was no impact on the pattern of results. Individuals’ level of engagement in the task did not impact the results. Whether the participant is in control of the experiment (Experiment 2) or not (Experiment 1), conceptual fluency continues to affect attentional bias.

General discussion

The aims of this study were twofold: (1) to examine the impact of sense of fluency (i.e., conceptual fluency) on attention during a visual discrimination task and (2) to study the effect of the ISI on individual performance. The results show that the sense of fluency generated by conceptual fluency attracts the attention of individuals to words that are semantically compatible with the predictive sentence stem. When the Go stimulus was located at the same place as the SCW (i.e., Valid situations), responses were faster than when it was located at the same place as the SIW (i.e., Non-valid situations). Thus, our results replicate those of Gardner et al. (2020) using another experimental paradigm. Furthermore, our results indicated that the ISI and the level of engagement do not impact performance.

These results suggest that attention is oriented towards the SCW. The manipulation of conceptual fluency (i.e., the sense of perceived fluency) seems to have an important role in the attentional process. Responses were slower in situations where the Go stimulus did not appear at the location of the SIW. Attention was—incorrectly—oriented towards the SCW, when the Go stimulus appeared under the SIW. This suggests that the environmental context influences the subjective experience of the perceiver (James, 1890). Attention can be manipulated with Valid or Non-valid spatial cues, and as Posner (1980) demonstrated, we can observe the “bad” placement of attention.

Our experiments combine the paradigm of conceptual fluency (Whittlesea, 1993) and the dot-probe task (MacLeod et al., 1986) to demonstrate the influence of conceptual fluency on attention. Although the dot-probe task is classically used to test for emotional bias, it is, nevertheless, useful in detecting attentional bias (van Rooijen et al., 2017). Our work supports the idea that this task is a useful way to understand the mechanisms underlying attentional bias.

Our observations occur under well-defined circumstances. Prior exposure to predictive sentence stems associated with target words leads to attentional capture by the SCW, resulting in an attentional bias. Contrary to what was suggested by Whittlesea (2002), the absence of a pause between the predictive stem and the target words (i.e., ISI) does not affect performance (Experiments 1 and 2). Furthermore, participants’ engagement in reading the predictive sentence stem does not affect performance (Experiment 3). In Valid situations, expectations are satisfied, resulting in a sense of fluency. When expectations are not satisfied (i.e., Non-valid situations), it causes a sense of surprise and consequently, a fluency break (Brouillet et al., 2016; Whittlesea & Williams, 2001) that is observed by an increase in response time.

Moreover, while a large majority of research hypothesises that novel (i.e., unfamiliar) stimuli will capture individuals’ attention, to the neglect of familiar stimuli, our study shows that individuals’ attention is captured by familiar and contextually fluent stimuli. In other words, in a specific context, the use of incomplete predictive sentences put participants in an expectancy situation. With the presentation of the words, their expectations for the last word of the sentence were met by the SCW. It is this expectation that captures attention: the SCW is fluent with the presented sentence, so it captures the attention of individuals. This observation is contrary to the assumption that novel or unexpected stimuli will capture individuals’ attention. If this had been the case, we should have observed faster response times when the target was at the spatial location of the SIW with the sentence.

Thus, we believe that this work can provide new insight into the deleterious effects of digital interruptions. It would not be simply the sudden appearance of a notification that would explain these deleterious effects, but the fact that it intervenes while our attention is captured by what we expect. For example, when we read a text, we constantly anticipate the word that will follow. If the word is the one that is expected, our attention is entirely captured by the word. If at that moment we receive a notification, we can understand that it will have a deleterious effect. However, if the word read is not the expected one, then the notification will not have such a negative effect.

Future directions

It would be interesting to repeat this study using an eye-tracking device, as the observation of eye movements could provide additional support for our findings. Eye-tracking devices have been used in various applications in the field of learning in a digital context (Noh et al., 2014) and semantic processing analysis (Buchmann & Mihali, 2009). Eye-tracking allows continuous and uninterrupted monitoring of automatic eye movements and provides a better understanding of information processing, particularly in the context of the reception of visual stimuli (Wickens, 2008).

Our results show that the absence or presence of an ISI does not influence the impact of conceptual fluency on attentional bias. From there on, it would be relevant to know whether the fair distribution between valid and non-valid situations plays a role on individual performance. By modifying the contextual distribution of stimuli (e.g., 70% of valid situations and 30% of invalid situations), this would allow us to address the question of the impact of fluency on attentional bias in a new way. We plan to address these questions in our future work.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Sibylle Turo

Denis Brouillet

References

Addas

Pinsonneault

(2015). The many faces of information technology interruptions: A taxonomy and preliminary investigation of their performance effects: Information technology interruptions taxonomy and performance effects. Information Systems Journal, 25(3), 231–273. https://doi.org/10.1111/isj.12064

Altmann

E. M.

Trafton

J. G.

(2007). Timecourse of recovery from task interruption: Data and a model. Psychonomic Bulletin & Review, 14(6), 1079–1084. https://doi.org/10.3758/BF03193094

Anderson

S. F.

Kelley

Maxwell

S. E.

(2017). Sample-size planning for more accurate statistical power: A method adjusting sample effect sizes for publication bias and uncertainty. Psychological Science, 28(11), 1547–1562. https://doi.org/10.1177/0956797617723724

Arndt

J. E.

Fujiwara

(2012). Attentional bias towards angry faces in trait-reappraisal. Personality and Individual Differences, 52(1), 61–66. https://doi.org/10.1016/j.paid.2011.08.030

Bar

(2007). The proactive brain: Using analogies and associations to generate predictions. Trends in Cognitive Sciences, 11(7), 280–289. https://doi.org/10.1016/j.tics.2007.05.005

Belopolsky

A. V.

Zwaan

Theeuwes

Kramer

A. F.

(2007). The size of an attentional window modulates attentional capture by color singletons. Psychonomic Bulletin & Review, 14(5), 934–938. https://doi.org/10.3758/BF03194124

Brouillet

Milhau

Heurley

Vagnot

Brunel

(2016). Word-to-picture recognition is a function of motor components mappings at the stage of retrieval: Motor action mappings in recognition. International Journal of Psychology, 51(5), 397–402. https://doi.org/10.1002/ijop.12210

Buchmann

R. A.

Mihali

S. T.

(2009). Semantic processing based on eye-tracking metrics. WSEAS Transactions on Computers, 8(10), 1701–1710.

Carretié

(2014). Exogenous (automatic) attention to emotional stimuli: A review. Cognitive, Affective, & Behavioral Neuroscience, 14(4), 1228–1258. https://doi.org/10.3758/s13415-014-0270-2

10.

Carretti

Cornoldi

Beni

R. D.

Palladino

(2014). What happens to information to be suppressed in working-memory tasks? Short and long term effects. Quarterly Journal of Experimental Psychology Section A, 57(6), 1059–1084.

11.

Christie

Klein

(1995). Familiarity and attention: Does what we know affect what we notice? Memory & Cognition, 23(5), 547–550. https://doi.org/10.3758/BF03197256

12.

Folk

C. L.

Remington

R. W.

Johnston

J. C.

(1992). Involuntary covert orienting is contingent on attentional control settings. Journal of Experimental Psychology: Human Perception and Performance, 18(4), 1030–1044.

13.

Gardner

Walker

E. B.

Gazzaley

Morsella

(2020). Involuntary attentional shifts as a function of set and processing fluency. Acta Psychologica, 203, Article 103009. https://doi.org/10.1016/j.actpsy.2020.103009

14.

Gernsbacher

M. A.

Faust

M. E.

(1991). The mechanism of suppression: A component of general comprehension skill. Journal of Experimental Psychology: Learning, Memory, and Cognition, 17(2), 245–262.

15.

Gernsbacher

M. A.

Keysar

Robertson

R. R. W.

Werner

N. K.

(2001). The role of suppression and enhancement in understanding metaphors. Journal of Memory and Language, 45(3), 433–450. https://doi.org/10.1006/jmla.2000.2782

16.

Gibson

B. S.

Jiang

(1998). Surprise! An unexpected color singleton does not capture attention in visual search. Psychological Science, 9(3), 176–182. https://doi.org/10.1111/1467-9280.00034

17.

Horstmann

(2015). The surprise-attention link: A review. Annals of the New York Academy of Sciences, 1339(1), 106–115. https://doi.org/10.1111/nyas.12679

18.

JASP Team. (2019). Jasp. Computer software.

19.

Jeffreys

(1961). Small corrections in the theory of surface waves. Geophysical Journal International, 6(1), 115–117. https://doi.org/10.1111/j.1365-246X.1961.tb02965.x

20.

James

(1890). The perception of reality. Principles of psychology, 2, 283–324.

21.

Lee

M. D.

Wagenmakers

E.-J.

(2013). Bayesian cognitive modeling: A practical course. Cambridge University Press. https://doi.org/10.1017/CBO9781139087759

22.

Lien

M. C.

Ruthruff

Goodin

Remington

R. W.

(2008). Contingent attentional capture by top-down control settings: Converging evidence from event-related potentials. Journal of Experimental Psychology: Human Perception and Performance, 34(3), 509. http://doi.org/10.1037/0096-1523.34.3.509

23.

MacLeod

Mathews

Tata

(1986). Attentional bias in emotional disorders. Journal of Abnormal Psychology, 95, 15–20.

24.

Neisser

(1976). Cognition and reality. W. H. Freeman and Company.

25.

Neter

Wasserman

(1974). Bonferroni multiple comparison. Applied Linear Statistical Models, 480.

26.

Noh

K.-B.

Song

K.-S.

Nam

S. C.

Park

S. Y.

(2014). The effects of multimedia elements on learning achievements in digital content. Journal of Physics: Conference Series, 1114, Article 012019.

27.

Rouder

J. N.

Speckman

P. L.

Sun

Morey

R. D.

Iverson

(2009). Bayesian t tests for accepting and rejecting the null hypothesis. Psychonomic Bulletin & Review, 16(2), 225–237. https://doi.org/10.3758/PBR.16.2.225

28.

Posner

M. I.

(1980). Orienting of attention. Quarterly Journal of Experimental Psychology, 32(1), 3–25. https://doi.org/10.1080/00335558008248231

29.

Schönbrodt

F. D.

Wagenmakers

E.-J.

(2018). Bayes factor design analysis: Planning for compelling evidence. Psychonomic Bulletin & Review, 25(1), 128–142. https://doi.org/10.3758/s13423-017-1230-y

30.

Syssau

Brouillet

Groen

. (2000). The study of inhibitory processes in aging with the Faust et al. Paradigm (1997): Methodological issues. Brain and Language, 72(2), 150–157. https://doi.org/10.1006/brln.1999.2282

31.

van Rooijen

Ploeger

Kret

M. E

. (2017). The dot-probe task to measure emotional attention: A suitable measure in comparative studies? Psychonomic Bulletin & Review, 24(6), 1686–1717. https://doi.org/10.3758/s13423-016-1224-1

32.

Yantis

Jonides

(1984). Abrupt visual onsets and selective attention: Evidence from visual search. Journal of Experimental Psychology: Human Perception and Performance, 10(5), 601.

33.

Wang

Cavanagh

Green

(1994). Familiarity and pop-out in visual search. Perception & Psychophysics, 56(5), 495–500. https://doi.org/10.3758/BF03206946

34.

Whittlesea

B. W. A.

(1993). Illusions of Familiarity. Journal of Experimental Psychology: Learning, Memory, and Cognition, 19(6), 1235–1253.

35.

Whittlesea

B. W. A.

(2002). Two routes to remembering (and another to remembering not). Journal of Experimental Psychology: General, 131(3), 325–348. https://doi.org/10.1037//0096-3445.131.3.325

36.

Whittlesea

B. W. A.

Leboe

J. P.

(2000). The Heuristic basis of remembering and classification: Fluency, generation, and resemblance. Journal of Experimental Psychology: General, 129, 84–106.

37.

Whittlesea

B. W. A.

Williams

L. D.

(1998). Why do strangers feel familiar, but friends don’t? A discrepancy-attribution account of feelings of familiarity. Acta Psychologica, 98(2–3), 141–165. https://doi.org/10.1016/S0001-6918(97)00040-1

38.

Whittlesea

B. W. A.

Williams

L. D.

(2001). The discrepancy-attribution hypothesis: I. The heuristic basis of feelings and familiarity. Journal of Experimental Psychology: Learning, Memory, and Cognition, 27(1), 3–13. https://doi.org/10.1037/0278-7393.27.1.3

39.

Wickens

C. D.

(2008). Multiple Resources and Mental Workload. Human Factors: The Journal of the Human Factors and Ergonomics Society, 50(3), 449–455. https://doi.org/10.1518/001872008X288394

40.

Wolfe

J. M.

Horowitz

T. S.

Kenner

N. M.

(2005). Rare items often missed in visual searches. Nature, 435(7041), 439–440. https://doi.org/10.1038/435439a