Effects of Individual Differences and Situational Features on Age Differences in Mindless Reading

Transcription

1 Shake, M.C., Shulley, L.J., Soto-Freita, A. (2015) Effects of individual differences and situational features on age differences in mindless reading. Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, doi: /geronb/gbv012. Effects of Individual Differences and Situational Features on Differences in Mindless Reading Matthew C. Shake, Leah J. Shulley, and Angelica M. Soto-Freita Department of Psychological Sciences, Western Kentucky University, Bowling Green. Objective. Mindless reading occurs when an individual shifts their attention away from the text and toward other offtask thoughts. This study examined whether previously reported age-related declines in mindless reading episodes are due primarily to (a) situational features related to the text itself (e.g., text genre or interest in the text) and/or (b) individual differences in cognitive ability. Method. Participants read 2 texts written in different genres but about the same topic. During reading, they were randomly probed to indicate whether they were on-task or mind-wandering. They also indicated their perceptions regarding the interest and difficulty of the text, and completed a battery of cognitive ability measures. Results. The results showed that (a) text genre may engender some age differences in mindless reading and (b) greater age and perceived interest in the text were each uniquely predictive of reduced mindless reading for both text genres. Individual differences in cognitive abilities (e.g., working memory, vocabulary) did not account for additional significant variance in mindless reading after interest and age were taken into account. Discussion. Our findings are discussed in terms of implications for age differences in lapses of attention during reading and predictors of mind-wandering generally. Key Words: Attention Cognition Executive function Mind-wandering Reading Self-regulation. Mind-wandering (MW) refers to the pervasive phenomenon wherein an individual decouples their attention from the goal or task at hand and shifts to task-unrelated thoughts (TUTs; Barron, Riby, Greer, & Smallwood, 2011; Smallwood, 2013; Smallwood & Schooler, 2006). This shift of attention unto TUTs is quite common and may occupy as much as 30% 50% of daily awake thoughts (Giambra, 1989; Kane et al., 2007; Killingsworth & Gilbert, 2010). Research suggests that MW has detrimental effects in domains such as reading comprehension and some aspects of cognitive performance (though it can be beneficial in other areas; see Mooneyham & Schooler, 2013). Recently, there has been significant interest and debate among researchers as to how and why the mind wanders (e.g., Franklin, Mrazek, Broadway, & Schooler, 2013; McVay & Kane, 2012a, 2012b; McVay, Meier, Touron, & Kane, 2013; Smallwood, 2013). In the context of MW during reading, researchers such as Smallwood (2011) have highlighted the relative paucity of research elucidating the roles of situational features (e.g., text characteristics, interest, difficulty, etc.) and individual differences (e.g., aging, cognitive ability, working memory [WM] capacity, etc.) in giving rise to mindless reading. On the most basic level, understanding whether aging affects the nature and frequency of mindless reading is obviously important because age-related declines in text memory are well documented (Johnson, 2003). One might, therefore, hypothesize that some of those memory deficits could be due to age-related increases in MW. However, there have now been a few studies examining age differences in MW, all of which have found that older adults (OAs) mind-wander less than younger adults (YAs), during both reading and nonreading tasks, and across varying levels of task complexity (Carriere, Seli, & Smilek, 2013; Jackson & Balota, 2012; Krawietz, Tamplin, & Radvansky, 2012; McVay et al., 2013; Zavagnin, Borella, & De Beni, 2014). However, the aforementioned situational features and/or individual differences driving this age-related decline in mindless reading are still unclear. In this study, we partly address this gap by examining the effects of (a) situational features of text genre, interest, and difficulty, and (b) individual differences in aspects of cognitive aging, upon mindless reading. Our motivation for studying these specific factors with regard to aging was derived from the substantial literature implicating these same factors in older and YAs differential allocation of on-task attention during reading (e.g., McVay et al., 2013; Payne & Stine-Morrow, 2014; Shake & Stine-Morrow, 2011; Stine-Morrow, Gagne, Morrow, & Dewall, 2004; Stine-Morrow, Miller, Gagne, & Hertzog, 2008; Stine-Morrow, Miller, & Hertzog, 2006; Stine-Morrow, Shake, et al., 2010; Zabrucky & Moore, 1999). For example, several of the aforementioned studies have found that characteristics of the text genre interact with age to affect reading patterns and subsequent text memory; however, whether such a Genre interaction affects lapses of attention (i.e., MW) during reading has not been The Author Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please journals.permissions@oup.com. Received August 20, 2014; Accepted January 30, 2015 Decision Editor: Myra Fernandes, PhD Page 1 of 13

2 Page 2 of 13 SHAKE ET AL. studied. We therefore hypothesized that such a factor could explain some of the age differences in the frequency and nature of the TUTs that older and younger readers experience. Before describing our experimental approach, we briefly describe the nature of mindless reading, then turn to how it may manifest differently in older and younger readers as a function of situational features of the text and/or cognitive ability. Mindless Reading Although MW has now been studied during a variety of tasks, reading provides an extremely practical window into investigating MW because people frequently report a positive correlation between the two. Such mindless reading is often described by individuals as a feeling of the eyes scanning the words on a page, but no attention being paid to their meaning. As one would expect, mindless reading has been shown to have negative effects on reading comprehension (e.g., McVay & Kane, 2010; Mooneyham & Schooler, 2013; Reichle, Reineberg, & Schooler, 2010; Smallwood, McSpadden, & Schooler, 2008). For example, Smallwood and colleagues (2008) found that readers who mind-wandered during reading were less likely to generate critical inferences necessary for good comprehension, and Schooler, Reichle, and Halpern (2004) found that readers who zoned-out during reading (i.e., mind-wandered unaware until they were interrupted and asked) had poorer recognition memory on a later test. The reasons why mindless reading leads to negative comprehension and text memory likely have to do with the nature of attention allocation during reading. That is, inattentive and attentive reading appear to have distinct and dissociable patterns of time allocation to linguistic features (e.g., Foulsham, Farley, & Kingstone, 2013; Reichle et al., 2010; Schad, Nuthmann, & Engbert, 2012). Attentive reading involves several parallel processes in which the reader must identify the individual words (i.e., surface form), encode and parse these words into meaningful propositions (Kintsch, 1988), and then ultimately build a situation model that incorporates what they read with general world knowledge and inferences (Graesser, Millis, & Zwaan, 1997). Recently, Smallwood (2011; also see Schad et al., 2012) compellingly argued that MW disrupts each of these aspects of linguistic processing, leading to impoverished text representations and poorer comprehension and memory. However, there are significant unresolved issues in mindless reading research, particularly the need to better explicate the role of situational features and individual differences (also see Dixon & Bortolussi, 2013). These include the effects of (a) the structure of the text itself, (b) perceived interest in the text, and (c) individual differences in cognitive ability inherent in normal aging. As we mentioned earlier, although numerous studies have documented how such factors affect age differences in self-regulation of attention during on-task reading (e.g., Noh et al., 2007; Shake, Noh, & Stine-Morrow, 2009; Shake & Stine-Morrow, 2011; Stine-Morrow, Noh, & Shake, 2010; Stine-Morrow, Shake, et al., 2010; Stine-Morrow, Shake, Miles, & Noh, 2006), little research has examined age differences in mindless reading. Differences in Effects of Situational Features on Mindless Reading Effects of text genre. One significant limitation of MW studies conducted so far is that the range of texts has been extremely limited. By far, the modal text in mindless reading studies thus far seems to be Tolstoy s War and Peace (e.g., (Jackson & Balota, 2012; Krawietz et al., 2012; McVay & Kane, 2012b; Sayette, Reichle, & Schooler, 2009; Sayette, Schooler, & Reichle, 2010; Schooler et al., 2004). This novel is a historical fiction replete with narrative genre structures. It is well known that text genre influences readers self-regulated attention during reading. For example, narrative genres often require tracking aspects of the characters in a story, resulting in a complex situation model in which the reader tracks oftenimplicit dimensions of space, temporality, causality, and so forth (Zwaan, Magliano, & Graesser, 1995). Expository genres, in contrast, are generally used to convey information or ideas in a descriptive, essay-based form. Thus, although both narrative and expository genres require the reader to build a discourse-level representation of the text, they engender different patterns of attention allocation (Britton, 1994). More importantly, text genre engenders age differences in the allocation of attention during on-task reading (e.g., Stine-Morrow et al., 2004, 2006, 2008). Studies in this area often show age equivalence (or even an aging-related advantage) for the processing and comprehension of texts with strong narrative genre structures (Morrow, Leirer, & Altieri, 1992; Radvansky & Dijkstra, 2007; Radvansky, Copeland, Berish, & Dijkstra, 2003; Radvansky, Curiel, Zwaan, & Copeland, 2001), which suggests that OAs find these types of texts more engaging than YAs. We therefore hypothesized that some of the age-related reduction in mindless reading reported in previous studies could be partly explained simply by the text genre itself. We tested this possibility in the current study. Effects of perceived interest in the text. Some of the previously conducted studies have suggested that differences in task interest are the culprit behind age-related declines in MW. For example, Jackson and Balota (2012) compared older and YAs in both a reading task (War and Peace; Tolstoy, 1942) and a sustained attention to response task and found that OAs mind-wandered less in both tasks. They argued this age difference in MW was due to (a) individual differences in personality (i.e., higher scores on conscientiousness among OAs) as well as (b) greater self-reported interest and difficulty on some tasks for the OAs. As such, they suggested that the degree to which laboratory tasks may be more interesting and engaging to the typical older

3 AGE DIFFERENCES IN MIND-WANDERING Page 3 of 13 adult participant could be a primary factor driving agerelated reductions in MW. However, in the reading task in their study, there was age equivalence in the interest and difficulty s, so in the context of mindless reading, it is not clear whether those were the most causal factors in mindless reading. In another recent report by Krawietz and colleagues (2012), an age-related reduction in MW was found in two reading studies (again using War and Peace); however, when self-reported interest was included as a covariate in the analysis, there was age equivalence in MW. Thus, they concluded that the higher level of interest among the OAs, rather than age per se, was the predictive factor for age differences in MW. The argument for interest as a significant factor in MW is consistent with earlier research on YAs, which found that perceived interest in the text was strongly related to frequency of TUTs (Grodsky & Giambra, 1990; also see McVay & Kane, 2009; Unsworth & McMillan, 2013). In this study, we sought to replicate these effects of interest on age differences in MW and to test this with different text genres. Differences in Effects of Cognitive Aging on Mindless Reading There continues to be a significant body of diverse research focused on understanding how aging affects the regulation of attention during cognitive tasks (e.g., Hayes, Kelly, & Smith, 2013; Sylvain-Roy, Lungu, & Belleville, 2014; Vasquez, Binns, & Anderson, 2014). Given that OAs have widely documented deficits in WM and other executive functions, there has been interest in studying age differences in MW, as a way to test hypotheses regarding the role of individual differences in cognitive ability. For example, McVay and Kane have found in a series of studies (e.g., McVay & Kane, 2009, 2010, 2012a) that individuals with poorer WM tend to MW more and have used this finding to support their contention that cognitive control is one factor which predicts MW episodes (i.e., individuals with poorer WM are more likely to experience lapses in attentional control and thus, TUTs). One might therefore guess that OAs would be expected to MW more; however, multiple aforementioned studies have shown that OAs MW less. The reasons for this are still unclear; however, it has been suggested that OAs may have fewer current concerns in the context of university lab-based studies, which may lead to reduced MW (but increased task-related interference; TRI) relative to younger college-age participants (McVay & Kane, 2010; McVay et al., 2013). Perhaps most problematically, the relationship between WM and MW has not been consistent across the previous aging studies. In their reading experiment, Jackson and Balota (2012) found a robust correlation between WM and MW (r =.52) regardless of age. In contrast, Krawietz and colleagues (2012) found no correlation between WM and MW. Because both studies used the same text (War and Peace), it is difficult to blame the inconsistency on text characteristics, although narrative genre processing does not always correlate well with measures of WM (e.g., Radvansky & Dijkstra, 2007; Radvansky et al., 2001). Furthermore, the role of other aspects of cognitive aging changes (e.g., vocabulary, processing speed, etc.) on MW has not been widely studied. It therefore remains unclear the extent to which aspects of cognitive ability are reliable predictors of mindless reading. Overall then, the aim of this study was to test whether age differences in mindless reading are influenced by situational features of the text (i.e., text genre, interest, and difficulty) and to examine whether individual differences in fluid and/or crystallized cognitive ability are actually related to age differences in the type and frequency of mindless reading episodes. Method Participants We report how we determined our sample size and all data exclusions. To detect at least small-sized effects (f = 0.25) with a power of 0.95 and α = 0.05 in our design, we collected and analyzed data from 34 YAs (18 women) aged years (mean [M] = 19.53, standard error [SE] = 0.26) and 34 OAs (17 women) aged years (M = 67.12, SE = 0.10). An additional two younger and two OAs participated but were removed prior to data analysis, either for near-chance performance on comprehension questions or mean reading times (RTs) exceeding 2.5 standard deviations [SDs] of the sample. All individuals were screened beforehand to ensure they were native speakers of English and did not have any severe neurological or medical impairments. The OA sample was 91.2% European American and 8.8% unknown; regarding ethnicity, 91.2% were non-hispanic, 5.9% were Hispanic, and 2.9% were unknown. The YA sample was 67.6% European American, 26.5% African American, 2.9% multiple race, and 2.9% unknown; 94.1% were non-hispanic and 5.9% were Hispanic. YAs were recruited from courses at Western Kentucky University and received partial course credit for participation, whereas OAs were recruited from the community and received $20. Additional participant demographic data are shown in Table 1, including age comparisons. OAs had significantly more years of education. Self-reported measures of health, vision, and hearing were completed on a 5-point Likert scale (1 = poor, 5 = excellent). There was age equivalence in self-reported health and vision, but OAs reported worse hearing than YAs. On an objective vision test, YAs (M = 20/21, SE = 0.004) and OAs (M = 20/22, SE = 0.005) were equivalent in near-sightedness, but YAs (M = 20/21, SE = 0.004) had better far-sighted vision than the OAs (M = 20/27, SE = 0.02; p =.003). All participants who needed corrective lenses wore them during the entirety of the study. We next report all manipulations and measures in the study.

4 Page 4 of 13 SHAKE ET AL. Measure Younger adult Table 1. Participant Characteristics Older adult M SE M SE Years of education (66) <.001* Vision (66).11 Hearing (66) <.001* Overall health (66).87 Working memory (65).52 Vocabulary (66) <.001* Processing speed (66) <.001* Inductive reasoning (66).08 Notes. M = mean; SE = standard error. *Significant age difference. t (df) p Value Materials Cognitive measures. Participants completed a battery of cognitive tests (see Table 1 for M and SE), including a measure of Vocabulary (ETS-KFT Advanced Vocabulary; Ekstrom, 1976), Processing Speed (WAIS-III Digit Symbol; Wechsler, 1997), Working Memory Span (RSPAN; Conway & Kane, 2005), and Inductive Reasoning (ETS-KFT Letter Sets; Ekstrom, 1976). The older participants scored better in vocabulary (p <.001), poorer in processing speed (p <.001) but similar to younger participants both inductive reasoning (p =.08) and working memory span (p =.52). Texts. For this experiment, we selected two texts with clearly different genres (narrative and expository), but similar levels of readability and the same overarching topic. For the narrative text, we chose Survival in Auschwitz (Levi, Woolf, & Roth, 1996), which is an autobiographical narrative of the author s experiences in the Nazi death camp. The text provides a structure typical of situation model discourse in narratives (e.g., a protagonist, goals, etc.). The expository text was Anatomy of the Auschwitz Death Camp (Gutman & Berenbaum, 1994), which provides a historical, factual description of the death camp s structure and operations. The texts were equivalent in objective readability (Gunning-Fog Index of 14.2 for both texts, and a Flesch- Kincaid Grade Level of 11.6 and 11.1, for the narrative and expository texts, respectively). None of the participants reported having read either of the two texts previously. Thought probes. Following procedure from other MW studies, the participants were periodically presented with thought probes every 2 4 min (randomly sampled from a uniform distribution) during reading to assess whether they were MW or not. The probe that appeared on the screen followed a bell sound and asked, Can you describe what was going through your head just before you heard the sound? When probed with this question, participants pressed a button to indicate whether just before the sound (a) their mind was blank, (b) they were aware that thoughts about other things (not the actual text) popped into their heads, (c) they were thinking of other things, but unaware of it until the sound, or (d) their minds were on the task and thus not MW. Following terminology from Schooler and colleagues (2004), we will refer to these four options as spacing out, tuning out, zoning out, and being on-task, respectively. Procedure After giving consent, completing demographics, and taking a vision test, participants were seated in front of a computer to complete the readings. Participants were told to read at a normal pace. A brief description of MW was provided, and participants were informed that occasionally there would be a bell sound and a question about their MW experience would follow, which they were to answer honestly and accurately. Following the initial instructions, the experimenter left the room and the participant read the first of the two texts for 30 min. The order of the two texts was randomized. Text was presented on the computer screen one page at a time (one page consisted of ~20 lines of text); participants pressed the space bar to advance to each subsequent page. Rereading previous pages was not permitted. Whenever participants were randomly probed for their current thoughts and pressed a button to indicate their response, the program returned them to the page they had been previously reading. After reading the text for 30 min, participants were asked to rate how interesting the text was and how difficult it was to understand. For each, they selected s on 5-point Likert scales (higher numbers indicated more interest and more difficulty). Next, they were presented with a series of true false comprehension questions about the text. The number of questions participants received were based on how many pages they had read in 30 min (ranging from 5 to 20 questions). After repeating the entire process again for the second text, participants completed the cognitive ability measures in the following order: RSPAN, Advanced Vocabulary Test, Digit Symbols Task, and the Letter Sets Test. Results Ms and SEs are reported for the thought probe measures in Table 2. Correlations between the various measures are presented separately for each age group in Table 3 (YAs) and Table 4 (OAs).

5 AGE DIFFERENCES IN MIND-WANDERING Page 5 of 13 Table 2. Ms and SEs for Mind-Wandering Probes, Reading Times, Comprehension Scores, and Ratings of and Difficulty Measure M SE M SE M SE M SE Total probes encountered Overall MW Tune-outs Zone-outs Space-outs RT/syllable prior page off-task RT/syllable same page off-task RT/syllable prior page on-task RT/syllable same page on-task Comprehension Difficulty Notes. M = mean; RT = reading time; SE = standard error. Younger adults Thought Probes The thought probe means are reported as probe-caught ratios, which were calculated by dividing the frequency of self-reported MW episodes by the total overall number of MW probes. We first submitted overall rates of MW (i.e., tune-outs [TOs], zone-outs [ZO], and space-outs [SO] combined) to a 2 () 2 (Genre: Narrative vs. Expository) analysis of variance (ANOVA), before examining each subtype separately. Effect sizes are reported as partial η 2 values. Overall MW. Participants reported being off-task fairly frequently during reading; on average, they reported being offtask on about 30% of the probes during the narrative text and 42% of the probes during the expository text. The main effect of Genre was significant, F(1, 66) = 16.34, p <.001, η 2 = 0.20, indicating participants mind-wandered more during the expository text. Additionally, we replicated the main effect of found by previous studies, such that OAs reported significantly less overall MW, F(1, 66) = 43.73, p <.001, η 2 = However, these main effects were subsumed by a significant Genre interaction, F(1, 66) = 3.93, p =.05, η 2 = 0.06, which showed that while both age groups tended to MW more for the expository text, the difference was driven primarily by the YAs, t(33) = 3.58, p <.01 (the p value for the OAs was marginally significant at.07). Given this interaction on overall MW, we next analyzed each sub-type of MW individually. Tune-outs. Frequencies of aware MW were fairly high, with probe response rates across all participants ranging from 14% for the narrative text to 23% for the expository text. The results for TOs were similar to the overall MW rates; a main effect of Genre indicated that participants were less likely to TO during the narrative text, F(1, 66) = 17.05, p <.001, η 2 = OAs were also less likely to TO than YAs, F(1, 66) = 16.88, p <.001, η 2 = A significant Genre interaction, F(1, 66) = 6.61, p =.01, η 2 = 0.09, showed that YAs engaged in TOs during the expository text more often than the narrative text, t(33) = 4.12, p <.001; in Older adults Narrative Expository Narrative Expository contrast, OAs did not show any difference in TOs as a function of Genre, t(33) = 1.34, ns. Zone-outs. Frequencies of unaware MW were fairly modest, with probe response rates across all participants ranging from 10% for the narrative text to 12% for the expository text; both of these rates of ZOs were significantly above zero, t(67) = 5.91 and t(67) = 7.06, ps <.001, and OAs zoned-out less than YAs, F(1, 66) = 10.54, p <.01, η 2 = However, there was no difference as a function of Genre, F(1, 66) = 1.47, ns, and the interaction was not significant, F < 1. Space-outs. Frequencies of SOs were fairly low, with probe response rates across all participants ranging from 5% for the narrative text to 6% for the expository text. Nevertheless, rates of SOs were significantly above zero, even for OAs (all one-sample t-tests, p <.05). As with the other sub-types, OAs had fewer space-outs than YAs, F(1, 66) = 11.97, p <.01, η 2 = Consistent with the ZO data, there was no effect of Genre, and no interaction, both F < 1. RTs and Comprehension Performance Some studies have found that during MW, reading patterns change (e.g., Reichle et al., 2010; Schad et al., 2012). Krawietz and colleagues (2012) examined age differences in RTs as a function of MW, but did not find any effect on RTs in either age group. To examine this question in this study, we analyzed participants RTs separately for each text because the thought probe data suggested that Genre and were significant. To do this, we took the page RT in milliseconds (ms) and divided it by the total number of syllables on that page; we completed the RT/syll calculations separately for (a) the current page that participants were reading when randomly probed, and (b) the immediately preceding page (Ms and SEs are shown in Table 2). For the immediately preceding page RT/syll, we did not find any difference in reading rates as a function of on-task versus off-task thought, F(1, 41) = 2.53, ns. However, for the

6 Page 6 of 13 SHAKE ET AL. Table 3. Younger Adult Correlations Between Mind-Wandering,, Difficulty, Comprehension, Cognitive Ability, and Reading Speed Variable Narrative space-outs 2. Narrative tune-outs Narrative zone-outs Narrative overall **.610** mind-wandering 5. Expository.377* space-outs 6. Expository tune-outs.350*.483** Expository zone-outs * * Expository overall * * **.527** mind-wandering 9. Narrative interest.372* *.602** Narrative difficulty * Expository interest.349* * *.446** Expository difficulty Narrative comprehension score 14. Expository comprehension * score 15. Working memory (RSPAN) 16. Vocabulary (KFT * advanced) 17. Processing speed.443** * (digit symbol) 18. Inductive reasoning (KFT letter sets) 19. Narrative RT/syll * for off-task page 20. Narrative RT/syll * * ** for on-task page 21. Expository RT/syll * * **.599** for off-task page 22. Expository RT/syll * *.372*.603**.724**.724** for on-task page Notes. RT = reading time. Higher numbers for RT/syll indicate slower reading speed. Higher numbers for processing speed indicate faster processing. *p <.05. **p <.01.

7 AGE DIFFERENCES IN MIND-WANDERING Page 7 of 13 Table 4. Older Adult Correlations Between Mind-Wandering,, Difficulty, Comprehension, Cognitive Ability, and Reading Speed Variable Narrative space-outs 2. Narrative tune-outs Narrative zone-outs Narrative overall **.647** mind-wandering 5. Expository space-outs * * 6. Expository tune-outs **.368*.562** Expository zone-outs ** Expository overall *.574**.552** **.554** mind-wandering 9. Narrative interest 10. Narrative difficulty 11. Expository interest 12. Expository difficulty 13. Narrative comprehension score 14. Expository comprehension score 15. Working memory (RSPAN) 16. Vocabulary (KFT advanced) 17. Processing speed (digit symbol) 18. Inductive reasoning (KFT letter sets) 19. Narrative RT/syll for off-task page 20. Narrative RT/syll for on-task page 21. Expository RT/syll for off-task page 22. Expository RT/syll for on-task page * * * * *.496**.374*.436** * *.362* **.524** ** * * *.357*.395* * ** * * ** * * *.525* * ** ** **.434* * **.499**.865** Notes. RT = reading time. Higher numbers for RT/syll indicate slower reading speed. Higher numbers for processing speed indicate faster processing. *p <.05. **p <.01.

8 Page 8 of 13 SHAKE ET AL. current page RT/syll, there was a reliable difference in reading rates in the expected direction, such that participants tended to read more slowly when reporting being off-task, F(1, 41) = 8.28, p <.01, η 2 = However, this effect did not interact with Genre (F < 1) or (F = 1.02, ns). For comprehension questions, a 2 2 ANOVA showed a main effect of Genre such that participants answered slightly more questions about the expository text (M = 9.32, SE = 0.34) than the narrative text (M = 8.71, SE = 0.32), F(1, 66) = 7.61, p <.01, η 2 = However, OAs and YAs did not differ in the number of questions answered, F(1, 66) = 1.07, ns, and there was no interaction of Genre and, F(1, 66) < 1, ns. Performance was generally good across all participants (M prop. correct =0.82; SD = 0.12). A 2 2 ANOVA showed no difference in performance based on Genre, F(1, 66) < 1, ns. OAs tended to perform better than YAs, F(1, 66) = 8.08, p <.01, η 2 = 0.11, but the interaction with Genre was only marginal, F(1, 66) = 3.07, p =.08, η 2 = Further examination indicated that the age groups performed similarly on the narrative questions, t(66) < 1, whereas OAs performed somewhat better than YAs on the expository questions, t(66) = 3.61, p <.01. Synopsis Our data clearly show that OAs mind-wander less during reading than YAs, and that text genre moderates this age difference. Furthermore, the Genre interaction was largely localized to a specific type of MW TOs which occur when an individual is fully aware they are off-task. This type of brazen MW represents a sort of willful disengagement with the reading task, suggesting that older readers were comparatively either more willing or more able to maintain attention to the demands of the task at hand. The Roles of Situational Features and Cognitive Aging on MW To potentially explain some of the aforementioned differences, we next examined the predictive power of (a) situational features and (b) individual differences in cognitive abilities. As mentioned earlier, we had several reasons for doing so. First, the role of fluid or crystallized cognitive abilities on age differences in MW is not totally clear and may predict some unique variance in MW. Second, given that prior research (e.g., Krawietz et al., 2012) found that interest seems to be strongly predictive of MW (perhaps independent of age), we sought to replicate the effect of this situational factor. We first report age comparisons on the situational features and then submit predictor variables into a series of hierarchical regression analyses for each text genre. Subjective perceptions of interest and difficulty. Ms and SEs for interest and difficulty s can be found in Table 2. A 2 2 ANOVA showed a main effect of such that OAs found the texts to be more interesting, F(1, 66) = 35.07, p <.001, η 2 = All participants tended to find the narrative text more interesting than the expository text, F(1, 66) = 48.86, p <.001, η 2 = The Genre interaction was also significant, F(1, 66) = 11.04, p <.01, η 2 = This interaction indicated that although both age groups found the narrative text more interesting than the expository text (both paired sample t-test ps <.05), the difference in perceived interest between the two genres was greater for the YAs than for the OAs. Although readability scores indicated the two texts were objectively similar, there was a main effect of Genre on subjective perceived difficulty, such that participants reported the expository text to be slightly more difficult to understand, F(1, 66) = 6.64, p <.05, η 2 = However, there were no age differences in perceived difficulty, F(1, 66) = 2.09, ns, and Genre did not interact with, F(1, 66) = 1.10, ns. Additionally, the mean s of difficulty (ranging from 2.4 to 3.1 on a scale from 1 to 5) suggest that overall participants did not find either text to be particularly easy or difficult to understand. (Difficulty was not related to MW tendencies in any of our preliminary analyses, so we dropped it from further analysis.) To address the question of whether age differences in MW are primarily explained by interest (cf. Krawietz et al., 2012) and/or cognitive ability, we examined the relative contributions of interest and cognitive ability in a series of four contrasting hierarchical regressions for each of the texts, with the levels entered as follows: (a) age, then interest, then cognitive abilities; (b) age, then cognitive abilities, then interest; (c) interest, then cognitive abilities, then age; and (d) cognitive abilities, then interest, then age. Predictors of narrative text MW. As shown in Table 5, both age and interest accounted for a significant amount of the variance in overall MW in all four models, regardless of when they were entered into the model. In Models 1 and 2, when age was entered first, cognitive ability did not account for any additional variance in MW, whereas interest did. When entered before age (Models 3 and 4), overall cognitive ability was somewhat predictive of MW; however, standardized coefficients suggest that this was driven primarily by processing speed, and the effect was still smaller than those of interest or age. Predictors of expository text MW. Regression output for expository text MW looked essentially identical to the narrative text. As shown in Table 6, both age and interest accounted for a significant amount of the variance in overall MW in all four models, regardless of when they were entered into the model. In Models 1 and 2, when age was entered first, cognitive ability did not account for any additional variance in MW, whereas interest did. When entered before age (Models 3 and 4), overall cognitive ability was predictive of MW; here, standardized coefficients suggest that this was driven primarily by vocabulary ability. Discussion In this study, we examined whether previously reported age differences in MW during reading could be explained

9 AGE DIFFERENCES IN MIND-WANDERING Page 9 of 13 Table 5. Hierarchical Multiple Regression Analyses Predicting Overall Mind-Wandering While Reading Narrative Text Predictor variable Adjusted R 2 ΔR 2 β F change Model 1 (age, interest, cognitive ability) Step ** 28.95** 0.56** Step ** 0.14** 16.56** 0.41** Step ** Working memory 0.06 Vocabulary 0.11 Processing speed 0.15 Inductive reasoning 0.07 Model 2 (age, cognitive ability, interest) Step ** 28.95** 0.56** Step ** Working memory 0.13 Vocabulary 0.1 Processing speed 0.14 Inductive reasoning 0.12 Step ** 0.13** 14.93** 0.40** Model 3 (interest, cognitive ability, age) Step ** 30.28** 0.56** Step ** 0.12* 3.30* Working memory 0.03 Vocabulary 0.16 Processing speed 0.24* Inductive reasoning 0.11 Step ** 0.04* 4.72* 0.39* Model 4 (cognitive ability, interest, age) Step ** 5.82** Working memory 0.1 Vocabulary 0.28* Processing speed 0.28* Inductive reasoning 0.18 Step ** 0.17** 18.06** 0.44** Step ** 0.04* 4.72* 0.39* Note. *p <.05. **p <.01. by situational factors such as text genre or interest and whether individual differences in cognitive ability were also related to MW. Consistent with previous research (Jackson & Balota, 2012; Krawietz et al., 2012; McVay et al., 2013), we found that OAs tend to mind-wander less overall. However, although some prior studies (Jackson & Balota, 2012) found no age differences in conscious TOs during reading, we found substantially less TOs in older participants (as well as other forms of MW). Furthermore, TOs were the only form of MW for which age differences were sensitive to text genre, such that YAs were less likely to maintain on-task thoughts while reading the expository text compared with the narrative text. In contrast, OAs TO rates did not vary by genre. This suggests that OAs were either more able or more willing to stay consciously engaged with the expository text. Our study also replicated the previous finding by Krawietz and colleagues (2012) that perceived interest in the text is a powerful predictor of MW and extends that finding to two different texts. However, although they found that the age MW relationship disappeared when interest was included as a covariate, our data suggested that age of the participant was a unique and reliable predictor of MW for both texts, above and beyond interest (though its explanatory power was considerably smaller than interest in the text; there are several possible explanations for why Krawietz and colleagues, 2012, found that interest fully explained the age MW relationship and we did not. For example, their older adult sample was somewhat older on average than ours, and there were several differences between the two samples in cognitive ability performance. Another possibility is the texts themselves led to different levels of interest: mean

10 Page 10 of 13 SHAKE ET AL. Table 6. Hierarchical Multiple Regression Analyses Predicting Overall Mind-Wandering While Reading Expository Text Predictor variable Adjusted R 2 ΔR 2 β F change Model 1 (age, interest, cognitive ability) Step ** 42.38** 0.63** Step ** 0.10** 12.54** 0.40** Step ** Working memory 0.04 Vocabulary 0.05 Processing speed 0.23* Inductive reasoning 0.15 Model 2 (age, cognitive ability, interest) Step ** 42.38** 0.63** Step ** Working memory 0.05 Vocabulary 0.06 Processing speed 0.17 Inductive reasoning 0.1 Step ** 0.11** 14.81** 0.45** Model 3 (interest, cognitive ability, age) Step ** 43.17** ** 0.63 Step ** 0.10* 2.96* Working memory 0.08 Vocabulary 0.30** Processing speed 0.15 Inductive reasoning 0.20* Step ** * * 0.41* Model 4 (cognitive ability, interest, age) Step ** 6.85** Working memory 0.01 Vocabulary 0.53** Processing speed 0.01 Inductive reasoning 0.16 Step ** 0.19** 23.05** 0.54** Step ** 0.04* 5.47* 0.41* Note. *p <.05. **p <.01. s of interest for both texts in this study were higher than those reported in Krawietz and colleagues for older adults [and comparable for younger adults]). Our findings thus clearly indicate that perceived interest in the task has a large effect on both older and YAs MW tendencies, although one limitation of both this study and previously cited studies is that the interest s were retrospective; as such, one s own MW tendencies could bias one s subsequent interest s. One way to partially address this could be for future MW research to distinguish between interest in the text and engagement with the text. The relationship between task interest and task engagement is not a new area of psychological inquiry (e.g., Higgins, Cesario, Hagiwara, Spiegel, & Pittman, 2010), but the distinction has not been widely studied in MW research and deserves further emphasis. For example, while researchers have suggested that MW episodes are a good predictor of text comprehension and recall (e.g., Smallwood et al., 2008), Dixon and Bortolussi (2013) recently provided evidence that (a) interest and especially (b) task engagement (i.e., how immersed one feels within the text/story) were better predictors of text memory than simply reporting being on-task (i.e., not MW). Such findings mesh well with our genre effects: even though both our texts were about the same topic and setting, the narrative text was found to be more interesting by both younger and OAs. Moreover, narrative texts lead to a more immersive situation model (Ericsson & Kintsch, 1995; Kintsch, 1988) that likely engages the reader more deeply. In other words, at least with regard to consequences for mindless reading, measures of engagement with the text may be ultimately more reflective of such consequences than existing probe measures of on- or off-task thoughts.

11 AGE DIFFERENCES IN MIND-WANDERING Page 11 of 13 Our experiment also examined the role of individual differences in cognitive abilities, some of which (e.g., WM) have been argued could be related to mindless reading (McVay & Kane, 2010, 2012b; Smallwood, 2011). Although our regressions suggest that overall cognitive ability (as a unitary construct) may predict overall MW tendencies, the effect was no longer significant once interest and age were taken into account. Unlike previous studies, our sample of higher-functioning older participants were equivalent to younger participants in WM, so our large age differences in MW rates cannot be attributed to that fluid cognitive ability. Ours is not the only study with OAs to fail to find a WM MW relationship: Krawietz and colleagues (2012) also found no reliable correlations between WM and MW, in either age group, across two experiments. However, like the aforementioned mindless reading studies (Jackson & Balota, 2012; Krawietz et al., 2012), our sample size was fairly small, and thus definitive conclusions regarding the role of individual differences in cognitive ability on MW cannot yet be reached. In summary, our data are indicative of both text characteristic (genre) and motivational (interest) components in the ability to maintain attentional allocation to the reading task at hand. ingly, age of the participant was also uniquely predictive of MW above and beyond interest, indicating that some other age-related differences may be driving changes in MW tendencies across the life span. Perhaps, as McVay & Kane (2010) have suggested, age-related differences in current concerns are the culprit. For example, one possibility is that OAs have more TRI centered around concerns regarding laboratory/cognitive tasks, which motivates them to engage more with the task at hand. Still, although research suggests TRI is in fact higher in OAs, this does not appear to fully explain the age-related reduction in MW (see McVay et al., 2013), so the question remains open. Another possibility is that differences in prior knowledge play a role. One limitation of this study was that, although our participants had not read the texts previously, it is possible that OAs had more pre-existing knowledge about the topic, which could influence MW tendencies. It should be noted, however, that in the other studies we described earlier, the age-related reduction in MW is found for texts that are unlikely to have pre-existing knowledge differences. Nevertheless, this is an important factor to consider in future research. Future research on age differences in mindless reading should also more closely examine the moment-tomoment online reading patterns that readers of different ages show during mindless reading, especially given age differences in resource allocation to linguistic features, and further evaluate whether age differences in meta-awareness of MW could also be an important factor. Funding This work was supported by Western Kentucky University s Research and Creative Activities Program ( ) to M. C. Shake. Acknowledgments We thank Emily Dickey for her assistance with data collection. Conflict of The authors have no conflicts of interest to declare. Correspondence Correspondence should be addressed to Matthew C. Shake, PhD, Department of Psychological Sciences, Western Kentucky University, 3041 Gary A. Ransdell Hall, Bowling Green, KY matthew. shake@wku.edu. References Barron, E., Riby, L. M., Greer, J., & Smallwood, J. (2011). Absorbed in thought: The effect of mind wandering on the processing of relevant and irrelevant events. Psychological Science, 22, doi: / Britton, B. K. (1994). Understanding expository text: Building mental structures to induce insights. In M. A. Gernsbacher (Eds.), Handbook of psycholinguistics (pp ). San Diego, CA: Academic Press. Carriere, J. S., Seli, P., & Smilek, D. (2013). Wandering in both mind and body: Individual differences in mind wandering and inattention predict fidgeting. Canadian Journal of Experimental Psychology, 67, doi: /a Conway, A., & Kane, M. (2005). Working memory span tasks: A methodological review and user s guide. Psychonomic Bulletin & Review, 12, doi: /bf Dixon, P., & Bortolussi, M. (2013). Construction, integration, and mind wandering in reading. Canadian Journal of Experimental Psychology, 67, doi: /a Ekstrom, R. B. (1976). Kit of factor-referenced cognitive tests. Princeton, NJ: Educational Testing Service. Ericsson, K. A., & Kintsch, W. (1995). Long-term working memory. Psychological Review, 102, doi: / x Foulsham, T., Farley, J., & Kingstone, A. (2013). Mind wandering in sentence reading: Decoupling the link between mind and eye. Canadian Journal of Experimental Psychology, 67, doi: / a Franklin, M., Mrazek, M., Broadway, J., & Schooler, J. (2013). Disentangling decoupling: Comment on Smallwood (2013). Psychological Bulletin, 139, doi: /a Giambra, L. M. (1989). Task-unrelated-thought frequency as a function of age: A laboratory study. Psychology and Aging, 4, doi: / Graesser, A. C., Millis, K. K., & Zwaan, R. A. (1997). Discourse comprehension. Annual Review of Psychology, 48, doi: / annurev.psych Grodsky, A., & Giambra, L. M. (1990). The consistency across vigilance and reading tasks of individual differences in the occurrence of task-unrelated and task-related images and thoughts. Imagination, Cognition and Personality, 10, doi: /6qg5-cxvv-4xur-7p3k Gutman, I., & Berenbaum, M. (1994). Anatomy of the Auschwitz death camp. Bloomington, IN: Published in association with the United States Holocaust Memorial Museum, Washington, D.C. by Indiana University Press. Hayes, M. G., Kelly, A. J., & Smith, A. D. (2013). Working memory and the strategic control of attention in older and younger adults. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 68, doi: /geronb/gbs057 Higgins, E. T., Cesario, J., Hagiwara, N., Spiegel, S., & Pittman, T. (2010). Increasing or decreasing interest in activities: The role of regulatory fit. Journal of Personality and Social Psychology, 98, doi: /a

12 Page 12 of 13 SHAKE ET AL. Jackson, J., & Balota, D. (2012). Mind-wandering in younger and older adults: Converging evidence from the sustained attention to response task and reading for comprehension. Psychology and Aging, 27, doi: /a Johnson, R. E. (2003). Aging and the remembering of text. Developmental Review, 23, doi: /s (03) Kane, M. J., Brown, L. H., McVay, J. C., Silvia, P. J., Myin-Germeys, I., & Kwapil, T. R. (2007). For whom the mind wanders, and when: An experience-sampling study of working memory and executive control in daily life. Psychological Science, 18, doi: /j x Killingsworth, M., & Gilbert, D. (2010). A wandering mind is an unhappy mind. Science (New York, N.Y.), 330, 932. doi: / science Kintsch, W. (1988). The role of knowledge in discourse comprehension: A construction-integration model. Psychological Review, 95, doi: / x Krawietz, S. A., Tamplin, A. K., & Radvansky, G. A. (2012). Aging and mind wandering during text comprehension. Psychology and Aging, 27, doi: /a Levi, P., Woolf, S., & Roth, P. (1996). Survival in Auschwitz: The Nazi assault on humanity. New York, NY: Simon & Schuster. McVay, J. C., & Kane, M. J. (2009). Conducting the train of thought: Working memory capacity, goal neglect, and mind wandering in an executive-control task. Journal of Experimental Psychology: Learning, Memory, and Cognition, 35, doi: / a McVay, J. C., & Kane, M. J. (2010). Does mind wandering reflect executive function or executive failure? Comment on Smallwood and Schooler (2006) and Watkins (2008). Psychological Bulletin, 136, doi: /a McVay, J. C., & Kane, M. J. (2012a). Drifting from slow to D oh! : Working memory capacity and mind wandering predict extreme reaction times and executive control errors. Journal of Experimental Psychology: Learning, Memory, and Cognition, 38, doi: /a McVay, J. C., & Kane, M. J. (2012b). Why does working memory capacity predict variation in reading comprehension? On the influence of mind wandering and executive attention. Journal of Experimental Psychology, 141, doi: /a McVay, J. C., Meier, M. E., Touron, D. R., & Kane, M. J. (2013). Aging ebbs the flow of thought: Adult age differences in mind wandering, executive control, and self-evaluation. Acta Psychologica, 142, doi: /j.actpsy Mooneyham, B. W., & Schooler, J. W. (2013). The costs and benefits of mind-wandering: A review. Canadian Journal of Experimental Psychology, 67, doi: /a Morrow, D. G., Leirer, V. O., & Altieri, P. A. (1992). Aging, expertise, and narrative processing. Psychology and Aging, 7, doi: / Noh, S. R., Shake, M. C., Parisi, J. M., Joncich, A. D., Morrow, D. G., & Stine-Morrow, E. A. L. (2007). differences in learning from text: The effects of content preexposure on reading. International Journal of Behavioral Development, 31, doi: / Payne, B. R., & Stine-Morrow, E. A. L. (2014). Risk for mild cognitive impairment is associated with semantic integration deficits in sentence processing and memory. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences. Advance online publication. doi: /geronb/gbu103 Radvansky, G. A., & Dijkstra, K. (2007). Aging and situation model processing. Psychonomic Bulletin & Review, 14, doi: /bf Radvansky, G. A., Copeland, D. E., Berish, D. E., & Dijkstra, K. (2003). Aging and situation model updating. Aging, Neuropsychology, and Cognition, 10, doi: /anec Radvansky, G. A., Curiel, J. M., Zwaan, R. A., & Copeland, D. E. (2001). Situation models and aging. Psychology and Aging, 16, doi: / Reichle, E. D., Reineberg, A. E., & Schooler, J. W. (2010). Eye movements during mindless reading. Psychological Science, 21, doi: / Sayette, M. A., Reichle, E. D., & Schooler, J. W. (2009). Lost in the sauce: The effects of alcohol on mind wandering. Psychological Science, 20, doi: /j x Sayette, M. A., Schooler, J. W., & Reichle, E. D. (2010). Out for a smoke: The impact of cigarette craving on zoning out during reading. Psychological Science, 21, doi: / Schad, D. J., Nuthmann, A., & Engbert, R. (2012). Your mind wanders weakly, your mind wanders deeply: Objective measures reveal mindless reading at different levels. Cognition, 125, doi: /j.cognition Schooler, J. W., Reichle, E. D., & Halpern, D. V. (2004). Zoning out while reading: Evidence for dissociations between experience and metaconsciousness. In D. T. Levin (Ed.), Thinking and seeing: Visual metacognition in adults and children (pp ). Cambridge, MA: MIT Press. Shake, M. C., Noh, S. R., & Stine-Morrow, E. A. L. (2009). differences in learning from text: Evidence for functionally distinct text processing systems. Applied Cognitive Psychology, 23, doi: /acp Shake, M. C., & Stine-Morrow, E. A. (2011). differences in resolving anaphoric expressions during reading. Neuropsychology, Development, and Cognition, 18, doi: / Smallwood, J. (2011). Mind-wandering while reading: Attentional decoupling, mindless reading and the cascade model of inattention. Language and Linguistics Compass, 5, doi: /j x x Smallwood, J. (2013). Distinguishing how from why the mind wanders: A process-occurrence framework for self-generated mental activity. Psychological Bulletin, 139, doi: / a Smallwood, J., & Schooler, J. W. (2006). The restless mind. Psychological Bulletin, 132, doi: / Smallwood, J., McSpadden, M., & Schooler, J. W. (2008). When attention matters: The curious incident of the wandering mind. Memory & Cognition, 36, doi: /mc Stine-Morrow, E. A., Gagne, D. D., Morrow, D. G., & DeWall, B. H. (2004). differences in rereading. Memory & Cognition, 32, doi: /bf Stine-Morrow, E., Miller, L. M. S., Gagne, D. D., & Hertzog, C. (2008). Self-regulated reading in adulthood. Psychology and Aging, 23, doi: / Stine-Morrow, E. A. L., Miller, L. M. S., & Hertzog, C. (2006). Aging and self-regulated language processing. Psychological Bulletin, 132, doi: / Stine-Morrow, E. A., Noh, S. R., & Shake, M. C. (2010). differences in the effects of conceptual integration training on resource allocation in sentence processing. Quarterly Journal of Experimental Psychology (2006), 63, doi: / Stine-Morrow, E. A., Shake, M. C., Miles, J. R., Lee, K., Gao, X., & McConkie, G. (2010). Pay now or pay later: Aging and the role of boundary salience in self-regulation of conceptual integration in sentence processing. Psychology and Aging, 25, doi: /a Stine-Morrow, E. A., Shake, M. C., Miles, J. R., & Noh, S. R. (2006). Adult age differences in the effects of goals on self-regulated sentence processing. Psychology and Aging, 21, doi: / Sylvain-Roy, S., Lungu, O., & Belleville, S. (2014). Normal aging of the attentional control functions that underlie working