Precision in the measurement of lexical expertise: the selection of optimal items for a spelling assessment.

Skilled adult readers vary in many skills related to visual word form processing such as phonological processing, vocabulary size, comprehension skill, and spelling skill (Kuperman & Van Dyke, 2011). Spelling skill in particular has received much attention because low- and high-skill spellers show different patterns of lexical processing as measured through eye movement behavior, reaction times, and word learning (Eskenazi et al., 2018; Veldre & Andrews, 2014). Researchers commonly use a spelling dictation task to measure lexical expertise; however, there is limited evidence for its psychometric properties and room for improvement in item selection (Andrews et al., 2020). The purpose of this study was to assess the precision of 110 words as measures of lexical expertise, to compare various subsets of words in a spelling dictation task, and to provide a set of words that more precisely measure lexical expertise. In Study 1, a spelling dictation task with 110 words was administered to 682 participants. In Study 2, that same task and measures of vocabulary and comprehension were administered to 786 participants. Results indicated that the set of 110 words contains many words that are imprecise measures of spelling skill. Through an iterative process of removing words with high error variance, a set of 20 words was selected that minimizes measurement error and demonstrates discriminant validity from vocabulary and comprehension ability. We recommend this set of words as a more precise measure of spelling skill, which will provide more power to detect moderating effects of lexical expertise on reading processes.

Readers accurately monitor their incidental word learning success during silent reading.

College-age adults who are skilled at reading use sentence context and morphemic information to learn definitions for new words they encounter in print (e.g., Tong et al., 2014). Often, the definitions that adult readers are learning are for word forms that already exist in their mental lexicon (Hulme et al., 2019). Readers demonstrate discrepancies between confidence in what they are retaining during reading and what they actually recall after reading (e.g., Sperling et al., 2004). Thus, metacognitive awareness is undoubtedly important for vocabulary learning. Few studies have investigated the role of metacognition in adults' word learning during reading. In the current study, we examined readers' metacognitive awareness and learning for novel compound pseudowords. Participants read sentences containing semantically opaque or transparent novel compound words in informative sentence contexts. Reading times were recorded, and readers judged how well they derived and would remember the novel words' correct definitions. After reading, participants took a surprise vocabulary test. Sentences containing opaque novel compounds were difficult, indicated by longer reading times. Participants were also less confident in their abilities deriving definitions for opaque than transparent words, and were less accurate on the vocabulary test for opaque words. Participants judged making an inference to define opaque compounds as more difficult than for transparent compounds, suggesting awareness of word learning difficulty. Overall, metacognitive judgments were mostly accurate. However, individual readers' accuracy varied greatly. These results provide a start to understanding how incidental vocabulary learning, and monitoring for this, proceeds in skilled (college-level) adults.

Learning to Spell Novel Words: The Relationship Between Orthographic and Semantic Representations During Incidental Learning.

The current study investigated whether semantic (meaning) knowledge benefits learning orthography (spelling). Adult readers read 14 novel non-words embedded in sentences with informative or uninformative context. Orthographic and semantic posttests assessed learning. In E1, results indicated that the relationship between context and orthographic accuracy was moderated by spelling frequency. In E2, all novel words had low-frequency spelling bodies. The results did not show a main effect of access to meaning on learning spelling, but they did reveal a strong association between learning spelling and meaning. In E3, participants received fewer exposures to increase the task difficulty. There was no main effect of access to words' meaning on learning spellings, but there were strong associations between orthographic and semantic posttest accuracy. These findings indicate that teaching words' spellings and meanings independently of one another may not be the most beneficial means of learning new words.

Are there Individual Differences in Learning Homophones During Silent Reading?

The present study evaluated if lexical expertise, defined as the quality and quantity of a reader's word representations, influenced college students' ability to learn novel homophones while reading. In two experiments novel homophones (e.g. 'brale') and novel nonhomophones (e.g. 'gloobs') were embedded in sentences. In Experiment 1, novel homophones had low-frequency familiar word mates, and in Experiment 2 they had high-frequency familiar word mates. Learning was assessed with meaning and spelling recognition post-tests. Although eye movements during reading did not differ between the word types, participants had more difficulty learning the spellings of the novel homophones compared to the novel nonhomophones in Experiments 1 and 2. In contrast, participants only had difficulty learning the meaning of novel homophones when it had a low-frequency mate. Higher levels of lexical expertise were related to higher learning rates of novel homophone spellings only when the novel homophones had a high-frequency mate. Phonology is activated when novel words are encountered and can interfere with learning under certain circumstances.

Word skipping during the lexical acquisition process.

During reading, most words are identified in the fovea through a direct fixation; however, readers also identify some words in the parafovea without directly fixating them. This word skipping process is influenced by many lexical and visual factors including word length, launch position, frequency, and predictability. Although these factors are well understood, there is some disagreement about the process that leads to word skipping and the degree to which skipped words are processed. The purpose of this study was to investigate the word skipping process when readers are exposed to novel words in an incidental lexical acquisition paradigm. Participants read 18 three-letter novel words (i.e., pru, cho) in three different informative contexts each while their eye movements were monitored. They then completed a surprise test of their orthographic and semantic acquisition and a spelling skill assessment. Mixed-effects models indicated that participants learned spellings and meanings of words at the same rate regardless of the number of times that they were skipped. However, word skipping rates increased across the three exposures and reading times decreased. Results indicate that readers appear to process skipped words to the same degree as fixated words. However, this may be due to a more cautious skipping process used during lexical acquisition of unfamiliar words compared to processing of already known words.

Orthographic texture effects during spelling are due to variations in representational strength.

In this study, we investigated the source of the orthographic texture effect during familiar word spelling. Orthographic texture refers to the differential strength that individual letters in a word may be activated for output. Prior work indicates that strongly activated letters are more accurately produced than weakly activated ones (Jones, Folk, & Rapp, 2009, All Letters are not Equal: Sub-Graphemic Texture in Orthographic Working Memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 35, 1389-1402. DOI: 10.1037/a0017042). According to an orthographic lexical strength account, differences in activation of individual letters within words are the result of learning; more easily learned letters are more strongly represented and, thus, activated for output. However, an online-competition hypothesis indicates that the contribution of sublexical assembly processes during familiar word spelling may account for the effect. Current results support the orthographic lexical strength account; the lexical orthographic representations used for spelling are themselves textured.

Eye-tracking to Distinguish Comprehension-based and Oculomotor-based Regressive Eye Movements During Reading.

Regressive eye movements are eye movements that move backwards through the text and comprise approximately 10-25% of eye movements during reading. As such, understanding the causes and mechanisms of regressions plays an important role in understanding eye movement behavior. Inhibition of return (IOR) is an oculomotor effect that results in increased latency to return attention to a previously attended target versus a target that was not previously attended. Thus, IOR may affect regressions. This paper describes how to design materials to distinguish between regressions caused by comprehension-related and oculomotor processes; the latter is subject to IOR. The method allows researchers to identify IOR and control the causes of regressions. While the method requires tightly controlled materials and large numbers of participants and materials, it allows researchers to distinguish and control the types of regressions that occur in their reading studies.

Uninformative contexts support word learning for high-skill spellers.

The current study investigated how high-skill spellers and low-skill spellers incidentally learn words during reading. The purpose of the study was to determine whether readers can use uninformative contexts to support word learning after forming a lexical representation for a novel word, consistent with instance-based resonance processes. Previous research has found that uninformative contexts damage word learning; however, there may have been insufficient exposure to informative contexts (only one) prior to exposure to uninformative contexts (Webb, 2007; Webb, 2008). In Experiment 1, participants read sentences with one novel word (i.e., blaph, clurge) embedded in them in three different conditions: Informative (six informative contexts to support word learning), Mixed (three informative contexts followed by three uninformative contexts), and Uninformative (six uninformative contexts). Experiment 2 added a new condition with only three informative contexts to further clarify the conclusions of Experiment 1. Results indicated that uninformative contexts can support word learning, but only for high-skill spellers. Further, when participants learned the spelling of the novel word, they were more likely to learn the meaning of that word. This effect was much larger for high-skill spellers than for low-skill spellers. Results are consistent with the Lexical Quality Hypothesis (LQH) in that high-skill spellers form stronger orthographic representations which support word learning (Perfetti, 2007). Results also support an instance-based resonance process of word learning in that prior informative contexts can be reactivated to support word learning in future contexts (Bolger, Balass, Landen, & Perfetti, 2008; Balass, Nelson, & Perfetti, 2010; Reichle & Perfetti, 2003). (PsycINFO Database Record (c) 2018 APA, all rights reserved).

Regressions during reading: The cost depends on the cause.

The direction and duration of eye movements during reading is predominantly determined by cognitive and linguistic processing, but some low-level oculomotor effects also influence the duration and direction of eye movements. One such effect is inhibition of return (IOR), which results in an increased latency to return attention to a target that has been previously attended (Posner & Cohen, Attention and Performance X: Control of Language Processes, 32, 531-556, 1984). Although this is a low level effect, it has also been found in the complex task of reading (Henderson & Luke, Psychonomic Bulletin & Review, 19(6), 1101-1107, 2012; Rayner, Juhasz, Ashby, & Clifton, Vision Research, 43(9), 1027-1034, 2003). The purpose of the current study was to isolate the potentially different causes of regressive eye movements: to adjust for oculomotor error and to assist with comprehension difficulties. We found that readers demonstrated an IOR effect when regressions were caused by oculomotor error, but not when regressions were caused by comprehension difficulties. The results suggest that IOR is primarily associated with low-level oculomotor control of eye movements, and that regressive eye movements that are controlled by comprehension processes are not subject to IOR effects. The results have implications for understanding the relationship between oculomotor and cognitive control of eye movements and for models of eye movement control.

Reading skill and word skipping: Implications for visual and linguistic accounts of word skipping.

We investigated whether high-skill readers skip more words than low-skill readers as a result of parafoveal processing differences based on reading skill. We manipulated foveal load and word length, two variables that strongly influence word skipping, and measured reading skill using the Nelson-Denny Reading Test. We found that reading skill did not influence the probability of skipping five-letter words, but low-skill readers were less likely to skip three-letter words when foveal load was high. Thus, reading skill is likely to influence word skipping when the amount of information in the parafovea falls within the word identification span. We interpret the data in the context of visual-based (extended optimal viewing position model) and linguistic based (E-Z Reader model) accounts of word skipping. The models make different predictions about how and why a word and skipped; however, the data indicate that both models should take into account the fact that different factors influence skipping rates for high- and low-skill readers.

Skipped words and fixated words are processed differently during reading.

The purpose of this study was to investigate whether words are processed differently when they are fixated during silent reading than when they are skipped. According to a serial processing model of eye movement control (e.g., EZ Reader) skipped words are fully processed (Reichle, Rayner, Pollatsek, Behavioral and Brain Sciences, 26(04):445-476, 2003), whereas in a parallel processing model (e.g., SWIFT) skipped words do not need to be fully processed (Engbert, Nuthmann, Richter, Kliegl, Psychological Review, 112(4):777-813, 2005). Participants read 34 sentences with target words embedded in them while their eye movements were recorded. All target words were three-letter, low-frequency, and unpredictable nouns. After the reading session, participants completed a repetition priming lexical decision task with the target words from the reading session included as the repetition prime targets, with presentation of those same words during the reading task acting as the prime. When participants skipped a word during the reading session, their reaction times on the lexical decision task were significantly longer (M = 656.42 ms) than when they fixated the word (M = 614.43 ms). This result provides evidence that skipped words are sometimes not processed to the same degree as fixated words during reading.

Mrs. Malaprop's Neighborhood: Using Word Errors to Reveal Neighborhood Structure.

Many theories of language production and perception assume that in the normal course of processing a word, additional non-target words (lexical neighbors) become active. The properties of these neighbors can provide insight into the structure of representations and processing mechanisms in the language processing system. To infer the properties of neighbors, we examined the non-semantic errors produced in both spoken and written word production by four individuals who suffered neurological injury. Using converging evidence from multiple language tasks, we first demonstrate that the errors originate in disruption to the processes involved in the retrieval of word form representations from long-term memory. The targets and errors produced were then examined for their similarity along a number of dimensions. A novel statistical simulation procedure was developed to determine the significance of the observed similarities between targets and errors relative to multiple chance baselines. The results reveal that in addition to position-specific form overlap (the only consistent claim of traditional definitions of neighborhood structure) the dimensions of lexical frequency, grammatical category, target length and initial segment independently contribute to the activation of non-target words in both spoken and written production. Additional analyses confirm the relevance of these dimensions for word production showing that, in both written and spoken modalities, the retrieval of a target word is facilitated by increasing neighborhood density, as defined by the results of the target-error analyses.

All letters are not equal: subgraphemic texture in orthographic working memory.

A central issue in the study of reading and spelling has been to understand how the consistency or frequency of letter-sound relationships affects written language processing. We present, for the first time, evidence that the sound-spelling frequency of subgraphemic elements of words (letters within digraphs) contributes to the accuracy with which these letters are produced in spelling. We report findings from 2 studies that demonstrate that letters within digraphs display differential susceptibility to error under conditions of disruption to orthographic working memory (O-WM). In the 1st, O-WM was disrupted as a result of neurological damage; in the 2nd, O-WM disruption was produced in neurologically intact, skilled spellers under dual task conditions with a shadowing task carried out during spelling. In both studies, segments with low versus high levels of sound-letter convergence, a measure of the frequency of sublexical mappings, were more vulnerable to disruption even when factors such as letter position, consonant-vowel context, and phoneme-to-grapheme mapping probability of the digraphs were controlled. These results contribute to our understanding of the internal texture of orthographic representations, providing evidence that individual letters differ in their activation strength and, as a result, in their susceptibility to error.

The purpose of lexical/sublexical interaction during spelling: further evidence from dysgraphia and articulatory suppression.

We investigated how the lexical and sublexical processes interact in spelling using an articulatory suppression task to disrupt the sublexical process in a dysgraphic patient (JDO). Using a similar task, Folk et al. (2002) found evidence that the sublexical process interacts with the lexical process by strengthening a target word's graphemes. We replicated the findings of Folk et al. in a patient with a more severe deficit to the lexical process. We compared the error patterns produced under normal spelling conditions versus spelling during articulatory suppression and found an increase in lexical substitution errors ("thaw"-->T-H-O-U-G-H) under articulatory suppression. These findings indicate that by strengthening a target word's graphemes, the sublexical process helps to create an advantage for a target word over form-related word neighbours that compete with it for output.

Effects of syntactic category assignment on lexical ambiguity resolution in reading: an eye movement analysis.

Readers' eye movements were monitored as they read sentences containing lexically ambiguous words whose meanings share a single syntactic category (e.g., calf), lexically ambiguous words whose meanings belong to different syntactic categories (e.g., duck), or unambiguous control words. Information provided prior to the target always unambiguously specified the context-appropriate syntactic-category assignment for the target. Fixation times were longer on ambiguous words whose meanings share a single syntactic category than on controls, both when prior context was semantically consistent with the subordinate interpretation of a biased ambiguous word (Experiment 1) and when prior context was semantically neutral as to the intended interpretation of a balanced ambiguous word (Experiment 2). These ambiguity effects, which resulted from differences in difficulty with meaning resolution, were not found when the ambiguity crossed syntactic categories. These data indicate that, in the absence of syntactic ambiguity, syntactic-category information mediates the semantic-resolution process.

The interaction of lexical and sublexical information in spelling: What's the point?

Most theories of spelling propose two major processes for translating between orthography and phonology: a lexical process for retrieving the spellings of familiar words and a sublexical process for assembling the spellings of unfamiliar letter strings based on knowledge of the systematic correspondences between phonemes and graphemes. We investigated how the lexical and sublexical processes function and interact in spelling by selectively interfering with the sublexical process in a dysgraphic individual. By comparing spelling performance under normal conditions and under conditions of sublexical disruption we were able to gain insight into the functioning and the unique contributions of the sublexical process. The results support the hypothesis that the sublexical process serves to strengthen a target word and provide it with a competitive advantage over orthographically and phonologically similar word neighbours that are in competition with the target for selection.