Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Mar 1.
Published in final edited form as: Neuropsychol Dev Cogn B Aging Neuropsychol Cogn. 2021 Nov 24;30(2):181–200. doi: 10.1080/13825585.2021.2006598

So You Think You Can Read? Generalized Metacomprehension in Younger and Older Adults

Erika K Fulton 1
PMCID: PMC9127002  NIHMSID: NIHMS1766976  PMID: 34818140

Abstract

This study was an exploration of whether age differences in task-specific metacomprehension accuracy could be partly explained by age differences in generalized metacomprehension (GM) or the use of GM as a task-specific judgment anchor. GM was measured before and after a summarization and metacomprehension judgment task and then correlated with prediction judgment magnitude to assess anchoring and correlated with comprehension and task-specific metacomprehension accuracy to assess GM accuracy. Age differences in all of these relationships were then tested. GM was related to judgment magnitude but despite age differences in GM ratings, age did not moderate anchoring or GM accuracy. Age differences in task-specific metacomprehension accuracy do not seem to be explained by age differences in GM accuracy or its use as a judgment anchor. However, results are the first to show that older adults anchor task-specific metacomprehension judgments on their GM, providing unique evidence for the Anchoring and Adjustment Model of Metacomprehension in advanced age.

Keywords: metacomprehension, metacognitive accuracy, comprehension, anchoring and adjustment, reading

Metacomprehension is the ability to monitor and control one’s text comprehension (Maki & Berry, 1984). It improves throughout youth but, like comprehension itself (e.g., Hultsch, Hertzog, Dixon, & Small, 1998; Kintsch, 1998; Stine-Morrow, Soederberg Miller, Gagne, & Hertzog, 2008), may still be challenged in adulthood (Commander, Zhao, Li, Zabrucky, & Agler, 2014; Dunlosky & Lipko, 2007; Maki, 1998; Ozuru, Kurby, & McNamara, 2012; Pilegard & Mayer, 2015; Reid, Morrison, & Bol, 2017; Schraw & Moshman, 1995; Zabrucky, Agler, & Moore, 2009). Most metacomprehension research to date has focused on younger adults in educational contexts (Dunlosky & Lipko, 2007; Dunlosky, Rawson, Marsh, Nathan, & Willingham, 2013; Thiede, Anderson, & Therriault, 2003), but some cross-sectional aging studies provide evidence that metacomprehension worsens in older adulthood (Baker, Dunlosky, & Hertzog, 2010; Fulton, 2019; Miles & Stine-Morrow, 2004). This decline can be problematic because accurate metacomprehension is needed for news articles, as well as to prevent ill-informed political decisions and the social exchange of inaccurate information. It is also critical to comprehending financial and health documents, a skill that can worsen in older age (Finucane, Mertz, Slovic, & Schmidt, 2005; Finucane et al., 2002; Liu, Kemper, & Bovaird, 2009). The present study was designed to test whether metacomprehension monitoring deficits in older age might be partly explained by age differences in the accuracy of general beliefs about one’s reading ability and how these beliefs influence metacomprehension judgments. Although the results did not support the hypotheses, they provide unique evidence that older adults anchor their task-specific metacomprehension judgments on these general beliefs, despite their low accuracy, and the first evidence for the Anchoring and Adjustment Model of Metacomprehension (Zhao & Linderholm, 2008) in advanced age.

The importance of metacomprehension can be illustrated through The Nelson and Naren’s (1990) model of metacognition. According to the model, metacognition is an iterative process of monitoring object-level cognitions (e.g., memory or comprehension) and using knowledge about the processes of those cognitions to control one’s behavior. Metacomprehension monitoring can be evaluated by asking people to judge whether and when they have comprehended written texts, whereas metacomprehension control is the behavioral response to the monitoring assessment. For example, readers may notice that their attention has drifted or that they do not understand a referent in a prose passage (monitoring) and may pause to reread the passage to resolve a comprehension failure (control). Accurate awareness of comprehension failures and subsequent adjustment of reading can help bring readers closer to their comprehension goals. As such, accurate metacomprehension is essential to accurate comprehension (Maki, Jonas, & Kallod, 1994; Moore, Zabrucky, & Commander, 1997a; Thiede et al., 2003) and self-regulated learning (Stine-Morrow, Miller, & Hertzog, 2006; Winne & Perry, 2000).

Most metacomprehension research to date has focused on task-specific monitoring, which involves monitoring comprehension of a particular text’s content during or immediately after reading. Task-specific metacomprehension monitoring accuracy is defined as the correspondence between comprehension of specific texts and one’s judgments about their comprehension of those texts, and includes the example described above. There are two types of task-specific monitoring accuracy, relative and absolute, which differ in their real-world relevancies. Absolute accuracy is a measure of over- or under-confidence (the average difference between object-level performance and judgments) on a specific task and can affect whether one stops reading prematurely or spends more time rereading than is necessary. In contrast, relative accuracy is the ability to discriminate between more and less-well understood texts and can affect whether one rereads only parts one did not understand, rather than wasting time rereading the whole text.

Evidence for age differences in task-specific metacomprehension (Baker et al., 2010; Dunlosky, Baker, Rawson, & Hertzog, 2006; Fulton, 2019; Lin, Zabrucky, & Moore, 2002; Olin & Zelinski, 1997) is inconsistent. When age differences were found they generally involved both absolute and relative accuracy (Baker et al., 2010; Fulton, 2019; Miles & Stine-Morrow, 2004), but when age equivalence was found it was primarily for relative accuracy only (Dunlosky et al., 2006; Lin, Zabrucky, and Moore, 2002; Olin and Zelinski, 1997; Stine-Morrow et al., 2008 Experiment 1). This set of findings is difficult to interpret because the methodology varied so widely across a small number of studies. They differed in the number and type of texts used, the types of metacognitive judgments solicited, the structure of the comprehension test, and the type of monitoring accuracy that was reported; only three studies reported both absolute and relative accuracy (Baker et al., 2010; Fulton, 2019; Miles & Stine-Morrow, 2004). Importantly, except for Fulton (2019) and Lin et al. (2002), most “metacomprehension” studies are arguably studies of metamemory. That is, they used memory tests instead of comprehension tests and solicited judgments of learning of specific details rather than judgments of comprehension of a text’s situation model (i.e., getting the gist of a text; Kintsch & van Dijk, 1978), which includes aspects not explicitly stated in the prose.

Age differences in task-specific monitoring accuracy have been explained in a few different ways, with no converging evidence for one account over another. Miles & Stine-Morrow (2004) posited that older adults performed worse because the response scales were less intuitive to them. Baker et al. (2010) suggested that older adults based their judgments more on self-views of their general reading ability rather than task-specific cues. Most recently, Fulton (2019) compared age differences in task-specific monitoring measured with the traditional approach versus with summarization. She showed that age deficits in task-specific metacomprehension monitoring accuracy are much smaller when measured with summarization than when measured with a multiple-choice test. More specifically, the age difference was smaller when measured with overall summary quality, non-existent when measured with main ideas and important details, and even favored older adults when measured with themes. Fulton concluded that older adults’ monitoring accuracy suffers when measured with multiple choice tests because this type of assessment is less familiar and recently practiced for them (Jackson & Kemper, 1993), and because it is inconsistent with their social goals (Carstensen, 1993). Because goals affect how texts are processed (Kintsch, 1994), older adults may process texts in ways that allow for better metacomprehension performance with summaries, which have more social utility. But Fulton also speculated that older adults may have based their judgments more on past experiences and self-views, which were likely more accurate for judging summarization than multiple-choice tests. Baker and colleagues (2010) were the first to suggest that age differences in self-views may account for age differences in task-specific monitoring accuracy, but they did not directly test this relationship. Fortunately, the project underlying Fulton (2019) included a measure of these self-views, hereto called generalized metacomprehension (GM), which allowed for a statistical test of Baker et al.’s. hypothesis1.

To understand why task-specific judgments might be based on GM, it is important to understand the nature of GM first. GM includes reading self-efficacy, knowledge, skills, and experiences that accumulate over time and make up one’s reading self-concept (Efklides, 2009; Pintrich, 2002; Winne, 1996; Zimmerman, 1998). The difference between task-specific metacomprehension monitoring and GM can be illustrated by considering someone who fails to accurately monitor how well they understood a passage in “To Kill a Mockingbird” (inaccurate task-specific metacomprehension) but is generally a skilled reader and knows that they are above average in comprehension ability (accurate GM). Unfortunately, GM may also include implicit theories and subjective beliefs about oneself that may not be objectively true, as has been found in memory studies (Frank & Kuhlmann, 2017; Koriat, 1997; Koriat, Bjork, Sheffer, & Bar, 2004; Mueller & Dunlosky, 2017). GM accuracy is important because past reading experiences and beliefs have been shown to influence monitoring accuracy even more than current task experiences (Hacker et al., 2000; Moore, Lin-Agler, & Zabrucky, 2005; Mueller, Dunlosky, Tauber, & Rhodes, 2014; Mueller, Tauber, & Dunlosky, 2013). However, GM accuracy is much less studied than task-specific metacomprehension accuracy and it is unclear whether GM accuracy declines in older adulthood or whether GM influences older adults’ task-specific metacomprehension judgments.

Whether age differences in GM accuracy emerge may depend on how it is defined and can vary according to when GM ratings are solicited. GM accuracy was defined in two main ways for this study. The first, GM-Comprehension (GM-Comp) accuracy, is the correlation between GM ratings and objective comprehension performance. Only two studies so far have measured GM-Comp accuracy in older adults. When GM ratings (self-report) were solicited before a comprehension task, age differences were mostly absent (Moore et al., 1997a), but when GM ratings were solicited after the comprehension task, an age deficit emerged (Lin et al., 2000). Notably, in Lin et al., GM Capacity was correlated with comprehension only for younger adults. The second measure of GM accuracy, GM-Metacomprehension (GM-Meta) accuracy, is the relationship of GM ratings to task-specific metacomprehension monitoring accuracy (predictions or postdiction accuracy). There are no published studies of GM-Meta accuracy in older adults. Studies in younger adults have shown that those who are more overconfident in their task-specific monitoring report worse monitoring (Schraw, 1994, 1997) and have GM that differs most from their objective comprehension (Kwon & Linderholm, 2014). Because older adults tend to be more overconfident in task-specific metacomprehension than younger adults, they may also have less accurate GM. To assess GM accuracy more fully than in past studies, both GM-Comp accuracy and GM-Meta accuracy were calculated in the present study.

The Anchoring and Adjustment model of Metacomprehension Judgment (AAMJ; Zhao & Linderholm, 2008) was used as a framework to explore the research questions—whether age differences in task-specific metacomprehension monitoring accuracy could be explained by (1) age differences in GM accuracy, or (2) the extent to which GM serves as a basis for task-specific monitoring judgments. The AAMJ posits that metacomprehension monitoring accuracy suffers because we use low validity judgment anchors and/or insufficiently adjust from those anchors. For example, people may anchor their comprehension judgments on the belief that they are good readers (i.e., high GM) and then downgrade the judgments (i.e., reduce their confidence) when reading particularly difficult prose. However, a high level of GM may limit how much of a downward adjustment is made to metacomprehension judgments. Even if the anchor is not strongly activated or weighted, it can still influence and bias judgments. This would be especially likely if one’s GM anchor is wildly inaccurate, as in the case when one believes they are a very good reader but are quite unskilled. Even if the reader appropriately considers a task-specific feature when making judgments, adjustment from the GM anchor could be dramatically insufficient. The power of metacomprehension anchors is highlighted in two studies by Linderholm and colleagues. In one, they demonstrated that people’s comprehension predictions varied as a function of text genre and test type even when they did not read any actual texts (Linderholm, Zhao, Therriault, & Cordell-Mcnulty, 2008). In another, they showed that initial anchors affect prospective judgments (made before a test) as well as retrospective judgments (made after or at the time of test), albeit to a lesser degree (Zhao & Linderholm, 2011). But both studies included young adults only; there is no published evidence of AAMJ in older adults.

Using the AAMJ as a framework, it was reasoned that older adults may weight their GM when making task-specific judgments more than younger adults do, and that this would reduce the accuracy of their metacomprehension monitoring if their GM was also less accurate than younger adults’ GM. A decline in older adult GM accuracy would be worrisome because young adults’ GM is already not highly consistent with actual comprehension (Falchikov & Boud, 1989). Although GM has been shown to account for a significant amount of variance in actual comprehension, this amount is well below 50%, even when considering a full-scale measure of GM (Lin, Moore, & Zabrucky, 2000; Moore et al., 2005; Moore, Zabrucky, & Commander, 1997a; Moore et al., 1997b; Zabrucky, Moore, Agler, & Cummings, 2015).

To further guide hypotheses about GM accuracy, age differences in each of the component constructs (GM ratings, comprehension, and task-specific monitoring accuracy) were considered. Older adults’ GM ratings are sometimes similar to younger adults’ (Lin et al., 2000; Moore et al., 1997a) but other times lower (De Beni, Borella, & Carretti, 2007; Moore et al., 1997a). When older adults had lower GM ratings, they reported less use of regulation strategies, placing lower value on good comprehension skills (Moore et al., 1997a), and less knowledge of effective reading strategies (De Beni et al., 2007; Moore et al., 1997a). The inconsistent age differences in GM ratings are also reflected in age comparisons of objective comprehension (Caplan, DeDe, Waters, Michaud, & Tripodis, 2011; Fulton, 2019; Liu et al., 2009; Radvansky, Zwaan, Curiel, & Copeland, 2001; Shake & Stine-Morrow, 2017) and in task-specific metacomprehension (Baker et al., 2010; Dunlosky et al., 2006; Fulton, 2019; Lin et al., 2002; Olin & Zelinski, 1997). Thus, age differences in the constructs that contribute to GM accuracy suggest that age differences may exist in GM accuracy itself.

The assessment of GM accuracy requires additional considerations of the timing of GM rating solicitation. Measuring GM ratings before the main task (pretest GM) allows an assessment of age differences in the use of GM as a judgment anchor. Measuring GM ratings after the main task (posttest GM) is necessary because metacognitive task experiences can influence subsequent metacognitive knowledge and behaviors (Hacker, Bol, Horgan, & Rakow, 2000; Hertzog, Dixon, & Hultsch, 1990; Hertzog, Price, & Dunlosky, 2008; Hunter-Blanks, Ghatala, Pressley, & Levin, 1988; Sitzmann & Ely, 2010). Older adults do not always update strategy knowledge as well as younger adults (Price, Hertzog, & Dunlosky, 2008). Thus, age differences in GM accuracy may occur only when GM ratings are solicited after the main task, as previously shown for GM-Comp accuracy (Lin et al., 2000). Furthermore, because age differences in task-specific metacomprehension monitoring accuracy can be smaller at posttest (Baker et al., 2010), exactly how GM-Meta accuracy is calculated can have important bearings on results and conclusions. Pretest GM and posttest GM differ in their temporal proximity to task-specific monitoring judgments, so pretest GM is more likely to be related to task specific prediction accuracy, whereas posttest GM is more likely to be related to task-specific postdiction accuracy. With this in mind, GM-Meta accuracy was calculated by comparing pretest GM to absolute and relative prediction accuracy separately and then by comparing posttest GM to absolute and relative postdiction accuracy separately.

In summary, the AAMJ model was used as a framework for exploring the possibility that age differences in task-specific metacomprehension monitoring accuracy may be partly due to age differences in GM accuracy and/or age differences in the extent to which GM is used as a judgment anchor. It was hypothesized that 1) pretest GM ratings would be related to prediction magnitude (anchoring effect), but more so for older adults because of evidence that their predictions may be more biased by self-perceptions (Baker et al., 2010); 2) for GM-Comp Accuracy, age differences would be seen in post-accuracy (Lin et al., 2000) but not pre-accuracy (Moore et al., 1997a); and 3) for GM-Meta accuracy, age differences could occur either before or after the main task, given the age differences in GM ratings (De Beni et al., 2007; Moore et al., 1997a) and task-specific monitoring were previously found both before (Fulton, 2019) and after (Baker et al., 2010; De Beni et al., 2007; Fulton, 2019) the main tasks.

Method

In accordance with approval by the Institutional Review Board of the Office of Research Integrity Assurance at Georgia Institute of Technology (IRB #H13128), informed consent was obtained from all participants.

Participants

Participants included 141 younger adults (ages 18–30, M = 19.40, SD = 1.67, 54 women) from a large, southeastern university in the U.S. and 138 older adults (ages 60–80, M = 69.38, SD = 5.60, 84 women) recruited from a lab database. Younger adults earned course credit and older adults earned $10 per hour of their participation. Younger adults reported an average of 14.26 (SD = .99) years of education. Older adults reported an average of 17.36 (SD = 1.95) years of education.

Design

There were two between-subjects factors, Age Group (younger adult, older adult) and Reading Goal (to summarize for a professor/boss vs. acquaintance). Participants in each age group were randomly assigned to one of the two reading goal conditions. Reading goal effects were an exploratory part of the study and are discussed in the online supplement of Fulton (2019). All analyses were performed by collapsing across reading goal condition.

Materials

Metacomprehension Scale (MCS)

The MCS is a validated scale for measuring GM (Moore, Zabrucky, & Commander, 1997a). The MCS is composed of 22 items measuring seven subscales: Anxiety (“stress related to comprehension performance”), Achievement (perceived “importance of good comprehension skills”, Strategy (knowledge of “techniques to improve comprehension”), Capacity (“perception of [one’s own] comprehension abilities”), Task (“knowledge of basic comprehension processes”), Locus (belief in one’s “control of reading skills”), and Regulation (knowledge and use of “methods of resolving comprehension failures). Response to each item is on a 5-point Likert scale ranging from never (1) to always (5). In the sample, Cronbach alphas were .606 and .658 for the pretest and posttest MCS, respectively, which is within the range of published internal consistency (.57 to .87) for the scale (Moore et al., 1997b).

Texts

There were six expository texts, adapted from the Scholastic Aptitude Test (The College Board, 1997), were borrowed from Experiment 3 of Rawson and Dunlosky (2002). They contained 370 words on average and were written at a Flesch-Kincaid grade-level of 9.8–12.0 (M = 11.6) with a Flesch readability score of 22.1–62.2 (M = 44.3), and were presented in 18-point, Courier New bold font on personal desktop computers. Text titles included: Television Newscast, Precision of Science, Women in the Workplace, Zoo Habitats, American Indians, and Real vs. Fake Art.

Comprehension test

A five-alternative forced-choice test (from the same source as the texts, Rawson & Dunlosky, 2002), was used to assess text comprehension. The test consisted of eight questions per text, four about information explicitly stated in the text and four that required inferencing.

Procedure

After informed consent, participants completed a demographic survey and the MCS (the measure of pre-GM). Participants then read the six texts and made predictions about their ability to summarize the texts and answer multiple-choice questions about them. Next, they completed a measure of working memory capacity, attempted to summarize the texts, and made postdictions about the quality of each of their summaries. They then took a multiple-choice comprehension test on the texts, made postdictions about their performance after each set of questions, took a vocabulary test, and completed the MCS again (the measure of posttest GM).

Please see Fulton (2019) for full details of the materials and procedure. However, because the task-specific metacomprehension task was novel, it is further described here as well. Judgments about various aspects of summary quality were uniquely solicited both before and after they summarized each text. Predictions and postdictions about summary quality were made on a 7-point Likert scale (with 1 = not well and 7 = very well) in response to the following questions: “How well do you think you could generate/generated the theme to the text you just read?; “How well do you think you will be able to summarize/summarized this text?; “What percentage of the main ideas in this text do you think your summary will include/included?”; “What percentage of the important details in this text do you think your summary will include/included?”. To mirror the typical procedure used in task-specific metacomprehension accuracy studies, judgments about multiple-choice comprehension test, were also solicited before and after the test, with these questions: “How many questions out of eight do you think you will answer correctly/answered correctly about this text?”. The task-specific monitoring accuracy reported in the results below refers to only these judgments (not the judgments about summary quality) as they are where the largest age differences were seen in Fulton (2019) and are akin to the method used in past studies of GM.

Results

Analytic Approach

Analyses are reported with GM defined as only the Capacity score on the MCS. This is for focus and simplicity and because the Capacity subscale reflects the most likely anchor for judgments (e.g., I have good reading comprehension (Capacity subscale) and therefore am confident I will do well on this test). The main conclusions, most notably regarding age differences, do not differ when GM is operationally defined using the full MCS scale.

SAS (Littell, Milliken, Stroup, & Wolfinger, 2000) was used for all analyses. Independent t-tests comparing age groups on GM ratings, for pretest GM Capacity and posttest GM Capacity, separately, were run first. The other analyses involve running two models in Proc GLM for each dependent variable. Model 1 included GM Capacity and age group as predictors and Model 2 added the interaction of GM Capacity and age group as a predictor. Type III Sums of Squares F-tests are reported.

When null effects were found, data was reanalyzed with JASP (JASP Team, 2021) to generate Bayes factors for regression. Results were interpreted using Lee & Wagenmakers’ (2013) scheme for BF10, which measures the likelihood of the alternative hypothesis (age differences) compared to the null hypothesis (no age differences). Age interactions were assessed by dividing the BF10 for the model that included the age interaction term by BF10 for the model with just the main effects.

GM Ratings

Age differences in GM ratings are reported first, as a connection to prior literature and to show support for some of the premises in the introduction. Older adults were more confident than younger adults in their comprehension abilities at pretest, t(273) = −3.03, p < .05, d =.35, and older were also more confident than younger adults at posttest, tpost (275) = −2.54, p < .05, d = .31. See top of Table 1 for means. Table 1 also includes descriptive statistics for all other main variables.

Table 1.

Variable Means by Age Group

Young Old
Pre-GM ratings 4.24 (0.06) 4.47 (0.05)
Post-GM ratings 3.90 (0.06) 4.12 (0.06)
Task specific predictions 6.07 (0.08) 5.71 (0.10)
Comprehension 4.87 (0.10) 4.11 (0.10)
Pre-GM absolute accuracy 1.20 (0.11) 1.60 (0.11)
Pre-GM relative accuracy 0.10 (0.03) 0.05 (0.03)
Post-GM absolute accuracy 0.11 (0.10) 0.64 (0.10)
Post-GM relative accuracy 0.19 (0.03) 0.17 (0.03)
Open in a new tab

Note. Means with standard errors in parentheses.

GM Accuracy and Age Differences Therein

GM-Comp Accuracy

Pretest GM Capacity accounted for 5.4% the variance in comprehension, F(1, 273) = 15.78, p < .001, with no age interaction, F(3, 271) = .01, p = .93, . Posttest GM Capacity accounted for 8% of the variance in comprehension, F(1, 273) = 24.99, p < .001, with no age interaction, F(3, 271) = 0.00, p = .97. See Table 2 for regression coefficients. Bayesian analyses provided moderate evidence for null age interactions at both pretest, BF10 = .16, and post-test, BF10 = .15. In summary, at both pretest and posttest, those who reported more comprehension capacity also performed better on the comprehension test, with no apparent age differences therein.

Table 2.

Relationship of GM-Capacity to Comprehension

Variable B β SE R 2 ΔR2
Pretest Step 1 .18 .18***
 Constant 1.58 0.48**
 GM-Capacity 0.57 0.30 0.11***
 Age Group 0.89 0.36 0.14***
Step 2 .18 0.0
 Constant 1.62 0.70*
 GM-Capacity 0.56 0.29 0.16*** .05***
 Age Group 0.81 0.33 0.94
 GM-Capacity × Age Group 0.02 0.03 0.21
Posttest Step 1 .21 .21***
 Constant 1.70 0.39***
 GM-Capacity 0.58 0.34 0.09***
 Age group 0.90 0.37 0.13***
Step 2 .22 .01
 Constant 1.68 0.56**
 GM-Capacity 0.59 0.35 0.13*** .08***
 Age group 0.93 0.38 0.75
 GM-Capacity × Age Group −0.01 −0.01 0.18
Open in a new tab

Note.

*

p < .05

**

p < .01

***

p < .001.

GM-Meta Accuracy

As a reminder, GM Meta-accuracy is the relationship between GM and task-specific monitoring accuracy, defined separately as the relationship between GM and absolute accuracy and between GM and relative accuracy. Absolute accuracy is a measure of over or under-confidence and was calculated as the average difference between judgments and performance. Relative accuracy is a measure of the ability to distinguish between more and less-well understood passages and was calculated with a gamma correlation (See Fulton (2019) for age differences in absolute and relative accuracy).

Pretest GM Capacity was unrelated to pretest absolute accuracy, F(1, 273) = 1.46, p = .23, and there was no interaction with age, F(3, 271) = .12, p = .73. There was moderate Bayesian evidence for the null main effect, BF10 = .27, and for the null age interaction, BF10 = .25. Posttest GM Capacity was unrelated to posttest absolute accuracy, F(1, 275) = .26, p = .61, with no age interaction, F(3, 271) = .11, p = .74. Bayesian analyses provided moderate evidence for the null main effect, BF10 = .15, as well as for the null age effect, BF10 = .23. See Table 3 for regression coefficients.

Table 3.

Relationship of GM-Capacity to Task-Specific Monitoring Absolute Accuracy

Variable B β SE R 2 ΔR2
Pretest Step 1 .02 .02*
 Constant 1.16 .56
 GM-Capacity 0.10 0.05 0.12
 Age Group −0.37 −0.14 0.16*
Step 2 .025 .005
 Constant 1.37 0.82
 GM-Capacity 0.05 0.03 0.18 .005
 Age Group −0.74 −0.28 1.09
 GM-Capacity × Age Group 0.08 0.14 0.25
Posttest Step 1 .0476 .0476**
 Constant 0.66 0.43
 GM-Capacity −0.003 −0.002 0.10
 Age Group −0.53 −0.22 0.15***
Step 2 .0480 .0004
 Constant 0.81 0.62
 GM-Capacity −0.04 −0.02 0.15 .00
 Age Group −0.81 −0.33 0.83
 GM-Capacity × Age Group 0.07 0.11 0.20
Open in a new tab

Note.

*

p < .05

**

p < .01

***

p < .001.

Pretest GM Capacity was also unrelated to pretest relative accuracy, F(1, 250) = 1.56, p = .21, with no age interaction, F(3, 248) = 1.11, p = .29. Bayesian analyses provided moderate evidence for the null main effect, BF10 = .28, and for the null age interaction, BF10 = .33. Similarly, posttest GM Capacity was unrelated to posttest relative accuracy, F(1, 260) = .14, p = .71, with no interaction with age, F(3, 258) = .01, p = .91. Bayesian analysis provided moderate evidence for the null main effect, BF10 = .14, and slightly anecdotal evidence for the null age interaction, BF10 = .39. See Table 4 for regression coefficients. In summary, at both pretest and posttest, there was no relationship between GM Capacity and either measure of task-specific monitoring accuracy (absolute or relative) and no apparent age differences in these relationships.

Table 4.

Relationship of GM-Capacity to Task-Specific Monitoring Relative Accuracy

Variable B β SE R 2 ΔR2
Pretest Step 1 .01 .01
 Constant 0.23 0.18
 GM-Capacity −0.04 −0.07 0.04
 Age Group 0.04 0.05 0.05
Step 2 .01 .00
 Constant 0.03 0.26
 GM-Capacity 0.00 0.01 0.06 .006
 Age Group 0.40 0.51 0.35
 GM-Capacity × Age Group −0.08 −0.45 0.08
Posttest Step 1 .00 .00
 Constant 0.21 0.14
 GM-Capacity −0.01 −0.02 0.03 .00
 Age Group 0.01 0.02 0.05
Step 2 .00 .00
 Constant 0.20 0.21
 GM-Capacity −0.01 −0.01 0.05
 Age Group 0.05 0.06 0.27
 GM-Capacity × Age Group −0.01 −0.04 0.07
Open in a new tab

Note.

*

p < .05

**

p < .01

***

p < .001.

GM Anchoring and Age Differences Therein

GM Capacity accounted for 12% of the variance in prediction magnitude, F(1, 273) = 36.76, p <.0001. Those who reported greater reading comprehension capacity had more confident predictions, t(273) = 6.06, p < .0001. However, age did not moderate this relationship, F(3, 271) = .28, p = .60, despite younger adults having more confident predictions (M = 6.07, SE = .08) than older adults (M = 5.71 , SE = .10), t(277) = 2.78, p < .01. A Bayesian analysis provided moderate evidence for the null age interaction, BF10 = .19. See Table 5 for regression coefficients. In summary, younger and older adults seemed to base their judgments partly on GM, with no apparent age difference therein.

Table 5.

Relationship of Pre GM-Capacity to Prediction Magnitude

Variable B β SE R 2 ΔR2
Pretest Step 1 .17 .17***
 Constant 3.78 0.43
 GM-Capacity 0.66 0.39 0.10***
 Age Group −0.51 −0.23 0.12***
Step 2 .17 .00
 Constant 2.99 0.64***
 GM-Capacity 0.61 0.35 0.14*** .12***
 Age Group 0.07 0.03 0.85
 GM-Capacity × Age Group 0.10 0.20 0.19
Open in a new tab

Note.

***

p < .0001

Despite null age effects for anchoring, correlations among all main variables for each age group separately includes unique evidence that older adults anchor their task-specific metacomprehension judgments on GM and that GM accuracy is low even in older age. See Table 6.

Table 6.

Pearson Correlations Among All Main Variables

1 2 3 4 5 6 7 8 9
Pretest GM 1 .31* .27* .24* .02 .01 −.07 .06 .71**
Predictions .47** 1 .36** .57** .54** −.04 .12 .02 .41**
Comprehension .35** .33*** 1 .33** −.59** −.06 −.67** .05 .32*
Postdictions .32** .55** .53** 1 .20* −.15 .47** −.10 .36**
Pre Absolute Accuracy .07 .53** −.62** −.03 1 .01 .72** −.03 .06
Pre Relative Accuracy −.13 −.02 −.16 .05 .13 1 −.06 .18 .05
Post Absolute Accuracy −.07 .17* −.56** .40** .64** .22* 1 −.11 −.02
Post Relative Accuracy −.04 −.00 .02 .07 −.02 .10 .05 1 −.01
Posttest GM .74** .50** .40** .47** .06 −.11 .02 −.03 1
Open in a new tab

Note. Younger adult correlations are on bottom of table, older adult correlations are on top.

*

p < .05

**

p < .001.

***

p < .0001.

Discussion

Younger and older adults’ GM was similarly accurate and related to a similar extent to metacomprehension judgment magnitude. As such, age differences in GM accuracy and anchoring do not seem to help explain age differences in task-specific metacomprehension monitoring accuracy, but the moderate support for this evidence should be considered preliminary. Although the hypotheses were not supported, the results show for the first time that older adults anchor their task-specific metacomprehension judgments on their GM and that this anchor is a low-validity cue even in older age. These findings have implications for self-regulated learning in older adulthood as well as metacognitive theories more generally.

Anchoring in Older Adults

Despite Baker et al.’s (2010) suggestion that older adults might base their judgments more on self-views of general reading ability, there was no age difference in the relationship between GM and prediction magnitude. As such, age differences in GM anchoring do not seem to help explain age differences in task-specific metacomprehension monitoring. Still, the results provide novel and important evidence that older adults use GM as an anchor in their metacomprehension judgments and thus the first evidence for the AAMJ in advanced age. Given that GM was not less accurate in older adults in the present study, any task-specific monitoring deficit in older age may result from an age difference in adjustment from anchors . Relatedly, Epley & Gilovich (2006) found that younger adults who adjust less from anchors (for trivia questions) are those who are less motivated or under more cognitive load. Given that older adults tend to be differently motivated (Carstensen, 1993) and many have lower working memory capacity than younger adults (Verhaeghen & Salthouse, 1997), they may adjust less from anchors than younger adults. The motivational account, particularly, is consistent with Fulton (2019) in which it was reasoned that age deficits in task-specific metacomprehension monitoring are smaller when measured with summaries because older adults are more motivated in those contexts than with multiple-choice tests.

If age differences in anchoring and adjustment do partially account for age differences in task-specific metacomprehension monitory accuracy, the present results would be the first to provide evidence in support of this explanation. The only other evidence for older adult anchoring is in metamemory research. Connor et al. (Connor, Dunlosky, & Hertzog, 1997) showed that both younger and older adults anchored their judgments of learning on the midpoint of the range of possible recall performance (i.e., 50%). Notably, this is a very different type of anchor from GM. Thus, the current results are the first demonstration not only of metacomprehension anchoring in older adults but specifically for a self-view/trait-based anchor. In younger adult metamemory research, there are several demonstrations that anchoring affects monitoring accuracy (Finn & Metcalfe, 2007; Scheck, Meeter, & Nelson, 2004; Scheck & Nelson, 2005) and control (Yang, Sun, & Shanks, 2018), but more exploration of anchoring in older adults is still needed, in both metamemory and metacomprehension. Of course, the possibility that age differences in anchoring on other types of self-views, or on GM measured in a different way, affect age differences in monitoring accuracy cannot be ruled out.

The apparent anchoring on GM, a trait-based cue, deserves additional theoretical interpretation. According to the cue-utilization hypothesis (Koriat, 1997), a seminal theory in the field, metacognitive cues fall into one of three classes: intrinsic (characteristics of study materials), extrinsic (learning conditions and encoding operations), and mnemonic (subjective cues internal to the learner). The theory does not include a category for trait-based or self-view-based cues, such as GM, so the current finding suggests a possible need to revise the cue utilization hypothesis to include such a cue category. However, although the cue utilization hypothesis has been invoked within metacomprehension research, it is first and foremost a metamemory theory and more evidence that traits or general-self views act as cues may be warranted before revising theory.

GM Accuracy

The seeming age parity in GM-Comp accuracy might be interpreted in the context of similar studies that differed in methodology. The age similarity in pretest GM-Comp accuracy is consistent with the age parity reported by Moore et al., (1997a), but the age similarity in posttest GM-Comp accuracy is counter to the age deficits previously reported Lin et al. (2000). This discrepancy might be due to the different lengths of the comprehension tests used in Lin et al and the current study. Because Lin et al.’s test was considerably longer, it may have been more challenging for older adults who are more sensitive to memory interference and accessibility (Radvansky, Zacks, & Hasher, 2005), leading to an age deficit in GM-Comp accuracy in their sample. Thus, age differences in GM accuracy may only emerge when task demands exceed the ability and/or motivation of older adults. Alternatively, it cannot be ruled out that the metacognitive monitoring task in the present study, which Lin et al. did not have, changed GM in ways that allowed for comparable GM-Comp accuracy between age groups. The age parity in the other type of GM accuracy, GM-Meta accuracy, is harder to interpret because there are no published studies comparing GM ratings to task-specific metacomprehension monitoring accuracy, let alone age differences therein. Age differences in GM-Meta accuracy were hypothesized because of age differences in GM ratings (Moore et al., 1997) and age differences in metacomprehension monitoring accuracy (Baker et al., 2010; De Beni et al., 2007; Fulton, 2019), but age differences in these component constructs may not translate to an age difference in the relationship between them. Perhaps it is unreasonable to expect an age difference in a non-relationship between GM ratings and task-specific metacomprehension monitoring accuracy. Importantly, although Bayesian analyses provided moderate support for null age interactions, there may be true age differences that were undetected in this study, and strong conclusions should await further study.

The age parity in GM accuracy may be somewhat surprising given that there were age differences in GM ratings. Age differences in GM ratings were found on all subscales, but previously GM ratings were found to be similar between younger and older adults (Lin et al., 2000; Moore et al., 1997a) or only found for some aspects of GM, not in the GM-Capacity scale used in the present study. Furthermore, prior age differences were only found in the valuing of comprehension skills and knowledge and regulations of reading strategies (De Beni et al., 2007; Moore et al., 1997a), with higher ratings in older adults, the opposite direction to those herein. These inconsistent findings might be explained by sample differences. The younger adult cohort in the present study was at a particularly competitive and demanding university that is primarily known for its students’ quantitative aptitude, where students may have a higher standard for comparison and/or may not have perceived their verbal skills as their strength. Altogether, it seems that age comparisons GM accuracy may be particularly sensitive to methodological nuances and sample characteristics. Strong conclusions should await further study, especially given that the present study is the first to measure GM-Meta accuracy (including age differences therein) and only the third to measure age differences in GM-Comp accuracy.

Although it may be reassuring that GM accuracy did not seem to be worse for older adults (unlike other cognitions), it is also somewhat disappointing that they did not seem to benefit from the additional reading experiences that come with age. Despite years of accumulating knowledge, skills, and experiences with reading (Efklides, 2009; Pintrich, 2002; Winne, 1996; Zimmerman, 1998), GM may still reflect implicit theories and subjective beliefs about one’s reading abilities and task experience that are not entirely true (Frank & Kuhlmann, 2017; Koriat, 1997; Koriat et al., 2004; Mueller & Dunlosky, 2017). Whether this reflects limits on GM updating over the lifespan or not, it has negative implications for self-regulated learning because GM seems to be used as a judgment anchor. The results suggest a need for interventions to improve GM accuracy, for both younger and older adults, because accurate metacomprehension is essential to accurate comprehension (Maki, Jonas, & Kallod, 1994; Moore, Zabrucky, & Commander, 1997a; Thiede et al., 2003) and self-regulated learning (Stine-Morrow, Miller, & Hertzog, 2006; Winne & Perry, 2000). Such interventions might include addressing deficits in self-awareness and/or inaccurate integration and interpretation of reading experiences.

Age Differences in Task-Specific Metacomprehension Monitoring

Because the hypotheses were unsupported, reasons for the age differences in task-specific metacomprehension monitoring found in Fulton (2019) are still unclear. Fulton concluded that the deficit seems to be an issue of measurement approach because the deficit was absent or significantly smaller when measured in a way that is more consistent with older adults’ goals (Carstensen, 1993) and their more recent experiences (i.e., with summarization, Jackson & Kemper, 1993). This conclusion is still plausible but needs more supporting evidence. It is also possible that older adults’ task-specific metacomprehension monitoring is less accurate than younger adults’ because of stereotype threat. Stereotype threat experienced by older adults has been found to be related to age deficits in both memory (Chasteen, Bhattacharyya, Horhota, Tam, & Hasher, 2005) and metamemory (Fourquet et al., 2020). Another possible explanation is that younger and older adults use different task-specific cues to form their judgments, and/or weight cues differently. In support of this account, older adults’ GM ratings were higher than younger adults (i.e., they were more confident generally in their reading), but their task-specific predictions were lower in magnitude and their task-specific postdictions did not differ from younger adults’ (see Table 1), despite having worse comprehension and similarly accurate posttest GM. Cue selection, cue weighting, and even cue integration may differ between younger and older adults. Although the integration of multiple cues has been demonstrated for metamemory judgments in younger and older adults (Hines, Hertzog, & Touron, 2015; Undorf, Söllner, & Bröder, 2018), there is minimal evidence for age differences in cue use in metacomprehension (Dunlosky et al., 2006), so this will need to be explored further.

Limitations

As with any study, the present one is not without limitations. First, self-reports can be unreliable and self-reports of reading ability may vary across reading contexts (Hadwin, Winne, Stockley, Nesbit, & Woszczyna, 2001). GM ratings are presumably based on one’s overall history of reading texts that vary along a number of dimensions. The texts used herein may have been different from the types with which participants have the most experience and thus be limited in what they can tell us about their GM more generally, across all types of prose. Second, although the sample size was sufficient to detect several age-moderated effects, as reported in Fulton (2019), it may have been insufficient to detect particularly small interactions. As such, null effects herein may be Type II errors. Bayesian analyses provided moderate evidence for null age interactions, but stronger evidence is still needed for stronger conclusions. Lastly, because the analyses were correlational, it is not clear whether comprehension and task-specific metacomprehension experiences shape GM or vice versa, although any causal relationship is likely to be bidirectional in nature (Efklides, 2009). Third, neither anchors nor adjustment processes were experimentally manipulated in this study. GM, the anchor tested, cannot be manipulated in the lab as it is a set of beliefs that accumulate over years. Thus, a future direction could be to experimentally manipulate shorter-term or less stable anchors, as has been done in memory research (Mueller & Dunlosky, 2017), to directly assess whether age differences in those anchors may help explain age differences in task-specific metacomprehension monitoring. Such studies may lead to a revised Anchoring and Adjustment model of Metacomprehension Judgment (Zhao & Linderholm, 2008) that depends on age.

Conclusion

The age differences in task-specific metacomprehension monitoring accuracy found in Fulton (2019) do not seem to be explained by age differences in GM accuracy or anchoring, although these results should be considered preliminary. Although this suggests a relatively rare area of preserved cognition in older adults, it still leaves questions about why and when older adults are worse at task-specific metacomprehension monitoring accuracy. Nonetheless, a novel methodology was employed and a new finding emerged. This is the first reported measure of GM accuracy by comparing GM with task-specific metacomprehension monitoring and the first to show that older adults anchor their judgments on their GM. Older adults’ GM anchoring was similar to younger adults despite the low validity of the GM cue and a lifetime of experience that should, theoretically, result in more accurate GM with advanced age. Age differences in cue use, weighting, and adjustment from anchors should be further explored to help understand age comparisons of metacomprehension monitoring accuracy, and interventions should be developed to improve older adults’ GM accuracy because it is related to their task-specific judgments.

Acknowledgements

I would like to thank Jason Chan, Kevin Pan, and Raine Hayes for help with data collection and Christopher Hertzog and Erin Madison for feedback on initial drafts.

Funding

This work was supported by the National Institute of Aging under the Ruth L. Kirschstein Training Grant (NIA T32 AG00175); an APA Dissertation Award; and the APF/CogDop William C. Howell Scholarship.

Footnotes

Declaration of Interest

I declare no conflict of interest.

1

The GM data was not reported in Fulton (2019) because it was a peripheral part of the study and would have detracted from the focus of an already long and complex set of analyses.

References

  1. Baker J, Dunlosky J, & Hertzog C (2010). How Accurately Can Older Adults Evaluate the Quality of Their Text Recall? The Effect of Providing Standards on Judgment Accuracy. Appl Cogn Psychol, 24(1), 134. 10.1002/acp.1553 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Board TC (1997). 10 real SATs. New York: College Entrance Examination Board. [Google Scholar]
  3. Caplan D, DeDe G, Waters G, Michaud J, & Tripodis Y (2011). Effects of age, speed of processing, and working memory on comprehension of sentences with relative clauses. Psychology and Aging, 26(2), 439–450. 10.1037/a0021837 [DOI] [PubMed] [Google Scholar]
  4. Carstensen LL (1993). Motivation for social contact across the life span: A theory of socioemotional selectivity. In Jacobs J (Ed.), Nebraska Symposium on Motivation (pp. 209–254). Lincoln, NE: University of Nebraska Press. [PubMed] [Google Scholar]
  5. Chasteen AL, Bhattacharyya S, Horhota M, Tam R, & Hasher L (2005). How feelings of stereotype threat influence older adults’ memory performance. Experimental Aging Research, 31(3), 235–260. 10.1080/03610730590948177 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Commander NE, Zhao Y, Li H, Zabrucky KM, & Agler L-ML (2014). American and chinese students’ calibration of comprehension and performance. Current Psychology: A Journal for Diverse Perspectives on Diverse Psychological Issues. 10.1007/s12144-014-9235-4 [DOI] [Google Scholar]
  7. Connor LT, Dunlosky J, & Hertzog C (1997). Age-related differences in absolute but not relative metamemory accuracy. Psychology and Aging, 12(1), 50–71. 10.1037/0882-7974.12.1.50 [DOI] [PubMed] [Google Scholar]
  8. De Beni R, Borella E, & Carretti B (2007). Reading comprehension in aging: the role of working memory and metacomprehension. Neuropsychology, Development, and Cognition. Section B, Aging, Neuropsychology and Cognition, 14(2), 189–212. 10.1080/13825580500229213 [DOI] [PubMed] [Google Scholar]
  9. Dunlosky J, Baker JMC, Rawson KA, & Hertzog C (2006). Does aging influence people’s metacomprehension? Effects of processing ease on judgments of text learning. Psychology and Aging, 21(2), 390–400. 10.1037/0882-7974.21.2.390 [DOI] [PubMed] [Google Scholar]
  10. Dunlosky J, & Lipko AR (2007). Metacomprehension: A brief history and how to improve its accuracy. Current Directions in Psychological Science, 16(4), 228–232. 10.1111/j.1467-8721.2007.00509.x [DOI] [Google Scholar]
  11. Dunlosky J, Rawson KA, Marsh EJ, Nathan MJ, & Willingham DT (2013). Improving students’ learning with effective learning techniques: Promising directions from cognitive and educational psychology. Psychological Science in the Public Interest, 14(1), 4–58. 10.1177/1529100612453266 [DOI] [PubMed] [Google Scholar]
  12. Efklides A (2009). The role of metacognitive experiences in the learning process. Psichothema, 21, 76–82. [PubMed] [Google Scholar]
  13. Epley N, & Gilovich T (2006). The anchoring-and-adjustment heuristic : Why the adjustments are insufficient. Psychological Science, 17(4), 311–318. 10.1111/j.1467-9280.2006.01704.x [DOI] [PubMed] [Google Scholar]
  14. Falchikov N, & Boud D (1989). Student Self-Assessment in Higher Education : A Meta-Analysis, 59(4), 395–430. 10.3102/00346543059004395 [DOI] [Google Scholar]
  15. Finn B, & Metcalfe J (2007). The role of memory for past test in the underconfidence with practice effect. Journal of Experimental Psychology: Learning, Memory, and Cognition. Metcalfe, Janet: Department of Psychology, Columbia University, 401B Schermerhorn Hall, New York, NY, US, 10027, jm348@columbia.edu: American Psychological Association. 10.1037/0278-7393.33.1.238 [DOI] [PubMed] [Google Scholar]
  16. Finucane ML, Mertz CK, Slovic P, & Schmidt ES (2005). Task Complexity and Older Adults’ Decision-Making Competence. Psychology and Aging, 20(1), 71–84. 10.1037/0882-7974.20.1.71 10.1037/0882-7974.20.1.71 [DOI] [PubMed] [Google Scholar]
  17. Finucane ML, Slovic P, Hibbard JH, Peters E, Mertz CK, & MacGregor DG (2002). Aging and decision-making competence: An analysis of comprehension and consistency skills in older versus younger adults considering health-plan options. Journal of Behavioral Decision Making, 15(2), 141–164. 10.1002/bdm.407 [DOI] [Google Scholar]
  18. Fourquet NY, Patterson TK, Li C, Castel AD, & Knowlton BJ (2020). Effects of Age-Related Stereotype Threat on Metacognition. Frontiers in Psychology, 11(December). 10.3389/fpsyg.2020.604978 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Frank DJ, & Kuhlmann BG (2017). More than just beliefs: Experience and beliefs jointly contribute to volume effects on metacognitive judgments. Journal of Experimental Psychology: Learning Memory and Cognition, 43(5), 680–693. 10.1037/xlm0000332 [DOI] [PubMed] [Google Scholar]
  20. Fulton EK (2019). How well do you think you summarize? Metacomprehension accuracy in younger and older adults. Journal of Gerontology Series B: Psychological Sciences and Social Sciences, Advance Online Publication. 10.1093/geronb/gbz142 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hacker DJ, Bol L, Horgan DD, & Rakow EA (2000). Test prediction and performance in a classroom context. Journal of Educational Psychology, 92(1), 160–170. 10.1037//0022-0663.92.1.160 [DOI] [Google Scholar]
  22. Hadwin AF, Winne PH, Stockley DB, Nesbit JC, & Woszczyna C (2001). Context moderates students’ self-reports about how they study. Journal of Educational Psychology, 93(3), 477–487. 10.1037/0022-0663.93.3.477 [DOI] [Google Scholar]
  23. Hertzog C, Dixon RA, & Hultsch DF (1990). Metamemory in adulthood: Differentiating knowledge, belief, and behavior. In Hess TM (Ed.), Aging and cognition: Knowledge organization and utilization. (pp. 161–212). Elsevier Science. [Google Scholar]
  24. Hertzog C, Price J, & Dunlosky J (2008). How is knowledge generated about memory encoding strategy effectiveness? Learning and Individual Differences, 18(4), 430–445. 10.1016/j.lindif.2007.12.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Hines JC, Hertzog C, & Touron DR (2015). Younger and older adults weigh multiple cues in a similar manner to generate judgments of learning. Aging, Neuropsychology, and Cognition, 22(6). 10.1080/13825585.2015.1028884 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Hultsch DF, Hertzog C, Dixon RA, & Small BJ (1998). Memory change in the aged. Memory change in the aged. New York, NY US: Cambridge University Press. [Google Scholar]
  27. Hunter-Blanks P, Ghatala ES, Pressley M, & Levin JR (1988). Comparison of monitoring during study and during testing on a sentence-learning task. Journal of Educational Psychology, 80(3), 279–283. 10.1037//0022-0663.80.3.279 [DOI] [Google Scholar]
  28. Jackson JD, & Kemper S (1993). Age differences in summarizing descriptive and procedural texts. Experimental Aging Research, 19(1), 39–51. 10.1080/03610739308253922 [DOI] [PubMed] [Google Scholar]
  29. JASP Team. (2021). JASP.
  30. Kintsch W (1994). Text comprehension, memory, and learning. American Psychologist, 49(4), 294–303. 10.1037/0003-066x.49.4.294 [DOI] [PubMed] [Google Scholar]
  31. Kintsch W (1998). The representation of knowledge in minds and machines. International Journal of Psychology, 33(6), 411–420. 10.1080/002075998400169 [DOI] [Google Scholar]
  32. Kintsch W, & van Dijk TA (1978). Toward a model of text comprehension and production. Psychological Review, 85(5), 363–394. 10.1037/0033-295X.85.5.363 [DOI] [Google Scholar]
  33. Koriat A (1997). Monitoring one’s own knowledge during study: A cue-utilization approach to judgments of learning. Journal of Experimental Psychology: General, 126(4), 349–370. 10.1080/095414496383077 [DOI] [Google Scholar]
  34. Koriat A, Bjork RA, Sheffer L, & Bar SK (2004). Predicting one’s own forgetting: The role of experience-based and theory-based processes. Journal of Experimental Psychology: General, 133(4), 643–656. 10.1037/0096-3445.133.4.643 [DOI] [PubMed] [Google Scholar]
  35. Kwon H, & Linderholm T (2014). Effects of self-perception of reading skill on absolute accuracy of metacomprehension judgements. Current Psychology: A Journal for Diverse Perspectives on Diverse Psychological Issues, 33(1), 73–88. 10.1007/s12144-013-9198-x [DOI] [Google Scholar]
  36. Lee MD, & Wagenmakers E-J (2013). Bayesian cognitive modeling: A practical course. Bayesian cognitive modeling: A practical course. New York, NY, US: Cambridge University Press. 10.1017/CBO9781139087759 [DOI] [Google Scholar]
  37. Lin L-M, Moore D, & Zabrucky KM (2000). Metacomprehension knowledge and comprehension of expository and narrative texts among younger and older adults. Educational Gerontology, 26, 737–749. 10.1080/036012700300001395 [DOI] [Google Scholar]
  38. Lin L-M, Zabrucky KM, & Moore D (2002). Effects of text difficulty and adults’ age on relative calibration of comprehension. The American Journal of Psychology, 115(2), 187–198. 10.2307/1423434 [DOI] [PubMed] [Google Scholar]
  39. Linderholm T, Zhao Q, Therriault DJ, & Cordell-Mcnulty K (2008). Metacomprehension effects situated within an anchoring and adjustment framework. Metacognition and Learning, 3(3), 175–188. 10.1007/s11409-008-9025-1 [DOI] [Google Scholar]
  40. Littell RC, Milliken GA, Stroup WW, & Wolfinger RD (2000). SAS system for mixed models (4th ed.). Cary, NC: SAS Institute, Inc. [Google Scholar]
  41. Liu C, Kemper S, & Bovaird JA (2009). Comprehension of health-related written materials by older adults. Educational Gerontology, 35(7), 653–668. 10.1080/03601270902885504 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Maki RH (1998). Test predictions over text material. In Hacker DJ, Dunlosky J, & Graesser AC (Eds.), The eductional psychology series: Metacognition in educational theory and practice (pp. 117–144). Lawrence Erlbaum Associates Publishers. [Google Scholar]
  43. Maki RH, & Berry SL (1984). Metacomprehension of text material. Journal of Experimental Psychology: Learning, Memory, and Cognition, 10(4), 663–679. 10.1037/0278-7393.10.4.663 [DOI] [PubMed] [Google Scholar]
  44. Maki RH, Jonas D, & Kallod M (1994). The relationship between comprehension and metacomprehension ability. Psychonomic Bulletin & Review, 1(1), 126–129. 10.3758/bf03200769 [DOI] [PubMed] [Google Scholar]
  45. Miles JR, & Stine-Morrow EAL (2004). Adult age differences in self-regulated learning from reading sentences. Psychol Aging, 19(4), 626–636. 10.1037/0882-7974.19.4.626 [DOI] [PubMed] [Google Scholar]
  46. Moore D, Lin-Agler L-MM, & Zabrucky KM (2005). A Source of Metacomprehension Inaccuracy. Reading Psychology, 26(3), 251–265. 10.1080/02702710590962578 [DOI] [Google Scholar]
  47. Moore D, Zabrucky KM, & Commander NE (1997a). Metacomprehension and Comprehension Performance in Younger and Older Adults. Educational Gerontology. 10.1080/0360126970230506 [DOI] [Google Scholar]
  48. Moore D, Zabrucky KM, & Commander NE (1997b). Validation of the Metacomprehension Scale. Contemporary Educational Psychology, 22(4), 457–471. 10.1006/ceps.1997.0946 [DOI] [PubMed] [Google Scholar]
  49. Mueller ML, & Dunlosky J (2017). How beliefs can impact judgments of learning: Evaluating analytic processing theory with beliefs about fluency. Journal of Memory and Language, 93. 10.1016/j.jml.2016.10.008 [DOI] [Google Scholar]
  50. Mueller ML, Dunlosky J, Tauber SK, & Rhodes MG (2014). The font-size effect on judgments of learning: Does it exemplify fluency effects or reflect people’s beliefs about memory? Journal of Memory and Language, 70(1), 1–12. 10.1016/j.jml.2013.09.007 [DOI] [Google Scholar]
  51. Mueller ML, Tauber SK, & Dunlosky J (2013). Contributions of beliefs and processing fluency to the effect of relatedness on judgments of learning. Psychonomic Bulletin and Review, 20(2), 378–384. 10.3758/s13423-012-0343-6 [DOI] [PubMed] [Google Scholar]
  52. Nelson TO, & Narens L (1990). Metamemory: A Theoretical Framework and New Findings. The Psychology of Learning and Motivation, 26, 125–173. 10.1016/S0079-7421(08)60053-5 [DOI] [Google Scholar]
  53. Olin J, & Zelinski EM (1997). Age Differences in Calibration of Comprehension. Educational Gerontology. 10.1080/0360127970230106 [DOI] [Google Scholar]
  54. Ozuru Y, Kurby CA, & McNamara DS (2012). The effect of metacomprehension judgment task on comprehension monitoring and metacognitive accuracy. Metacognition and Learning, 7(2), 113–131. 10.1007/s11409-012-9087-y [DOI] [Google Scholar]
  55. Pilegard C, & Mayer RE (2015). Within-subject and between-subject conceptions of metacomprehension accuracy. Learning and Individual Differences, 41, 54–61. 10.1016/j.lindif.2015.07.003 [DOI] [Google Scholar]
  56. Pintrich PR (2002). The Role of Metacognitive Knowledge in Learning, Teaching, and Assessing. Theory Into Practice (Vol. 41). 10.1207/s15430421tip4104_3 [DOI] [Google Scholar]
  57. Price J, Hertzog C, & Dunlosky J (2008). Age-related differences in strategy knowledge updating: Blocked testing produces greater improvement in metacognitive accuracy for younger than older adults. Neuropsychology, Development, and Cognition. Section B, Aging, Neuropsychology and Cognition, 15(5), 601–626. 10.1080/13825580801956225 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Radvansky GA, Zacks RT, & Hasher L (2005). Age and inhibition: the retrieval of situation models. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences, 60(5), P276–P278. https://doi.org/60/5/P276 [pii] [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Radvansky GA, Zwaan RA, Curiel JM, & Copeland DE (2001). Situation models and aging. Psychology and Aging, 16(1), 145–160. 10.1037/0882-7974.16.1.145 [DOI] [PubMed] [Google Scholar]
  60. Rawson KA, & Dunlosky J (2002). Are performance predictions for text based on ease of processing? Journal of Experimental Psychology: Learning, Memory, and Cognition, 28(1), 69–80. 10.1037/0278-7393.28.1.69 [DOI] [PubMed] [Google Scholar]
  61. Reid AJ, Morrison GR, & Bol L (2017). Knowing what you know: Improving metacomprehension and calibration accuracy in digital text. Educational Technology Research and Development, 65(1), 29–45. 10.1007/s11423-016-9454-5 [DOI] [Google Scholar]
  62. Scheck P, Meeter M, & Nelson TO (2004). Anchoring effects in the absolute accuracy of immediate versus delayed judgments of learning. Journal of Memory and Language, 51(1), 71–79. 10.1016/j.jml.2004.03.004 [DOI] [Google Scholar]
  63. Scheck P, & Nelson TO (2005). Lack of Pervasiveness of the Underconfidence-With-Practice Effect: Boundary Conditions and an Explanation Via Anchoring. Journal of Experimental Psychology: General. Scheck, Petra: Psychology Department, University of Maryland, College Park, MD, US, 20742, petrabradley@yahoo.com: American Psychological Association. 10.1037/0096-3445.134.1.124 [DOI] [PubMed] [Google Scholar]
  64. Schraw G (1994). The effect of metacognitive knowledge on local and global monitoring. Contemporary Educational Psychology, 19, 143–154. 10.1006/ceps.1994.1013 [DOI] [Google Scholar]
  65. Schraw G (1997). The effect of generalized metacognitive knowledge on test performance and confidence judgments. Journal of Experimental Education, 65(2), 135–146. 10.1080/00220973.1997.9943788 [DOI] [Google Scholar]
  66. Schraw G, & Moshman D (1995). Metacognitive Theories. Educational Psychology Review, 7(4), 351–371. 10.1007/BF02212307 [DOI] [Google Scholar]
  67. Shake MC, & Stine-Morrow EAL (2017). Language and aging. Reference Module in Neuroscience and Biobehavioral Psychology, 1–6. 10.1017/CBO9781107415324.004 [DOI] [Google Scholar]
  68. Sitzmann T, & Ely K (2010). Sometimes you need a reminder: The effects of prompting self-regulation on regulatory processes, learning, and attrition. Journal of Applied Psychology, 95(1), 132–144. 10.1037/a0018080 [DOI] [PubMed] [Google Scholar]
  69. Stine-Morrow EAL, Miller LMS, & Hertzog C (2006). Aging and self-regulated language processing. Psychological Bulletin, 132(4), 582–606. 10.1037/0033-2909.132.4.582 [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Stine-Morrow EAL, Soederberg Miller LM, Gagne DD, & Hertzog C (2008). Self-regulated reading in adulthood. Psychology and Aging, 23(1), 131–153. 10.1037/0882-7974.23.1.131 [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Thiede KW, Anderson MCM, & Therriault D (2003). Accuracy of metacognitive monitoring affects learning of texts. Journal of Educational Psychology, 95(1), 66–73. 10.1037/0022-0663.95.1.66 [DOI] [Google Scholar]
  72. Undorf M, Söllner A, & Bröder A (2018). Simultaneous utilization of multiple cues in judgments of learning. Memory and Cognition, 46(4). 10.3758/s13421-017-0780-6 [DOI] [PubMed] [Google Scholar]
  73. Verhaeghen P, & Salthouse TA (1997). Meta-analyses of age–cognition relations in adulthood: Estimates of linear and nonlinear age effects and structural models. Psychological Bulletin, 122(3), 231–249. 10.1037/0033-2909.122.3.231 [DOI] [PubMed] [Google Scholar]
  74. Winne PH (1996). A metacognitive view of individual differences in self-regulated learning. Learning and Individual Differences, 8(4), 327–353. 10.1016/S1041-6080(96)90022-9 [DOI] [Google Scholar]
  75. Winne PH, & Perry N (2000). Measuring self-regulated learning. In Boekaerts M, Pintrich PR, & Zeidner M (Eds.), Handbook of Self-Regulation (pp. 531–566). San Diego, CA: Academic Press. [Google Scholar]
  76. Yang C, Sun B, & Shanks DR (2018). The anchoring effect in metamemory monitoring. Memory and Cognition, 46(3), 384–397. 10.3758/s13421-017-0772-6 [DOI] [PubMed] [Google Scholar]
  77. Zabrucky KM, Agler LML, & Moore D (2009). Metacognition in Taiwan: Students’ calibration of comprehension and performance. International Journal of Psychology, 44(4), 305–312. 10.1080/00207590802315409 [DOI] [PubMed] [Google Scholar]
  78. Zabrucky KM, Moore D, Agler L-ML, & Cummings AM (2015). Students’ Metacomprehension Knowledge: Components that Predict Comprehension Performance. Reading Psychology, 36(7), 627–642. 10.1080/02702711.2014.950536 [DOI] [Google Scholar]
  79. Zhao Q, & Linderholm T (2008). Adult metacomprehension: Judgment processes and accuracy constraints. Educational Psychology Review, 20(2), 191–206. 10.1007/s10648-008-9073-8 [DOI] [Google Scholar]
  80. Zhao Q, & Linderholm T (2011). Anchoring effects on prospective and retrospective metacomprehension judgments as a function of peer performance information. Metacognition and Learning, 6(1), 25–43. 10.1007/s11409-010-9065-1 [DOI] [Google Scholar]
  81. Zimmerman BJ (1998). Academic studing and the development of personal skill: A self-regulatory perspective. Educational Psychologist, 33(2), 73–86. [DOI] [Google Scholar]

RESOURCES