Reading and Writing

, Volume 25, Issue 6, pp 1385–1402

Exploring the syntactic skills of struggling adult readers

Authors

    • Department of Educational Psychology and Special EducationGeorgia State University
  • Daphne Greenberg
    • Department of Educational Psychology and Special EducationGeorgia State University
  • Jacqueline Laures-Gore
    • Department of Educational Psychology and Special EducationGeorgia State University
  • Justin C. Wise
    • Department of PsychologyOglethorpe University
Article

DOI: 10.1007/s11145-011-9324-9

Cite this article as:
Taylor, N.A., Greenberg, D., Laures-Gore, J. et al. Read Writ (2012) 25: 1385. doi:10.1007/s11145-011-9324-9

Abstract

This study investigated the syntactic ability of 82 struggling adult readers who recognize words between the third and fifth grade levels. Analysis of the adults’ performance on the TOLD-I:3 indicated that they were deficient on the syntactic task. Correlations found the struggling adult readers’ oral language skills, written language skills, and reading comprehension skills to be related. A regression analyses indicated that the adults’ syntactic knowledge did not individually predict reading comprehension, however their other oral language skills did. The findings of this study suggest that the adults performed similar to children who are either learning to read or considered poor readers. This study also contributes to the adult literacy field by providing exploratory information on an area (syntax and struggling adult readers) that is lacking.

Keywords

Syntactic abilityStruggling adult readersOral language

Introduction

Little is known about the relationship between oral and written language skills in the adult struggling reader population. Because of this, investigators often rely on what is known about typically developing children to study oral and written language processes with adult struggling readers (e.g., Greenberg, Ehri, & Perin, 1997). For children, oral language skills have been described as a foundation for reading skills (Hoff, 2009) with reading skills typically only developing after considerable proficiency in oral language is attained (Muter, Hulme, Snowling, & Stevenson, 2004). Within this paper, we discuss the relationship between oral language skills and written language skills in the adult struggling reader population. Specifically, we are contrasting oral language skills (phonology, expressive and receptive vocabulary, syntax) with written language skills (decoding and fluency) as predictors of reading comprehension.

Oral language is a complex system which relates sound to meaning. It consists of skills in phonology, expressive and receptive vocabulary, and syntax. Phonology refers to the sound system in oral language. Language consists of individual sounds called phonemes. Individual phonemes when sequenced together create words that we are able to understand and utilize to communicate with others (e.g. the phonemes /m/, /a/, /d/ blended together create the word mad) (Torgesen, Wagner, & Rashotte, 1994). The National Reading Panel describes the differences between expressive and receptive vocabulary as: “Receptive vocabulary is the vocabulary that we can understand when it is presented to us in text, or as we listen to others speak, while productive (expressive) vocabulary is the vocabulary we use in writing or when speaking to others” (National Institute of Child Health, Human Development, 2000 p. 2). Syntax is the way in which units of meaning are combined with one another and governs the ordering of sentences (Tomasello & Brooks, 1999). While a few studies have looked at the phonological and oral vocabulary skills of adults, none have looked at syntax. For this reason, the present study focuses on the oral syntactic skills of adult struggling readers.

Typical development of syntactic skill acquisition

Syntax is a primary tool for expressing complex thought thus allowing words to be combined to create unique combinations of meaning (Ferreira & Engelhardt, 2006). How a person decides to join words together to make a sentence is of interest to researchers of syntax (e.g., Hartsuiker & Westenberg, 2000). There is controversy in the field whether the underlying mechanisms for syntax construction in speaking and writing are the same, with Cleland and Pickering (2006) proposing that “syntax is accessed in the same way in spoken and written production” (p. 195).

Typical syntactic development includes the acquisition of the ability to form sentences from words and other smaller structural units (Simms & Crump, 1983; Tomasello & Brooks, 1999). The capability to manipulate syntactic structures of oral language is most commonly demonstrated by word order, a necessary cue to sentence interpretation. Typically developing children appear to learn basic word order very quickly (Akhtar, 1999; Nation & Snowling, 2000), demonstrating a sensitivity to the meaning carried by word order before producing word combinations (Tomasello & Brooks, 1999). The grammatical complexity of children’s speech usually continues to mature with age and develop as sentences become longer and structurally more complex. Later syntactic development is not primarily a matter of acquiring new grammatical structures. Rather, it seems to be a process of learning how to use existing structures to express thoughts with greater efficiency and skillfulness (Nippold, Hesketh, Duthie, & Mansfield, 2005).

Researchers have focused on syntactic development in children and adolescents, with very few investigators studying adults. Nippold et al. (2005) studied syntactic development in adults ages 20–49 years of age and found that syntactic development continues into early adulthood (20–29 years) and remains stable into the middle ages (40–49 years) (Nippold et al., 2005). Specifically, they defined these findings in terms of participant’s growth in mean length of T-unit (an utterance containing one main clause and possibly one or more subordinate clauses) and clausal density (total number of main and subordinate clauses that occur per T-unit). In another study, (Nippold, Mansfield, & Billow, 2007) clear differences were found between the syntactic complexity of children, adolescents, and adults. Compared to the children and adolescents, adult utterances were more complex and they produced longer T-units with greater amounts of subordinate clauses.

The relationship between syntactic knowledge and reading comprehension

Comprehension is the goal of reading (Nation, 2004), and therefore it is of interest to understand the relationship between syntactic ability and reading comprehension ability. For children, syntactical knowledge appears to have a relationship to their reading comprehension ability (Bentin, Deutsch, & Liberman, 1990; Gillon & Dodd, 1995; Mokhtari & Thompson, 2006; Nation & Snowling, 2000). Due to the paucity of research on adults who have difficulty reading, it is unknown whether a similar relationship exists for adults who struggle with their reading. This literature review turns to a brief overview of studies that discuss the relationship between syntax and reading comprehension with children. It is possible that this literature can be informative for the study of adult struggling readers. As Greenberg et al. (1997) found, the same underlying processes of reading predicted word recognition skills for children and adults matched for reading age. It is possible that if it is known how syntax relates to reading with children, hypotheses can be generated for adults who struggle with their reading.

Most studies that were reviewed for this paper indicate a relationship between syntactic and reading comprehension skills. However, two studies were found which do not lend strong support for this type of relationship. A brief review of the studies which indicate a relationship will first be described, followed by a brief review of the two studies which question this relationship. Gillon and Dodd (1995) investigated the effect of training syntactic skills in spoken language on the reading comprehension performance of students between the ages of 10 and 12. Results suggested that remediation of the student’s syntactic skills had a positive impact on their reading comprehension performance. Similar to Gillon and Dodd, Nation and Snowling’s (2000) findings support the view that children’s syntactic skills are related to their reading comprehension skills. Nation and Snowling used a word order correction task to assess syntactic skill of children considered good or poor comprehenders, between the ages of 7 and 10. Good comprehenders had at least average for age comprehension skills while poor comprehenders performed almost 3 years below the expected level for typical readers. Results indicated that compared to the good comprehenders, the poor comprehenders had deficient performances on the word order correction task. In another study, Mokhtari and Thompson (2006) found significant relationships between fifth grade students’ syntactic skill and their reading comprehension and reading fluency performances. Syntactic skill was measured through a Syntax Quotient obtained from performance on the Sentence Combining, Word Ordering, and Grammatical Competence subtests of the Test of Language Development-Intermediate [TOLD II:3] (Hammill & Newcomer, 1997). Results indicated that lower levels of syntactic skill corresponded to poor performance in reading comprehension and fluency. On the other hand, higher levels of syntactic skill contributed to higher levels of reading comprehension among the fifth grade students.

There are two studies which question the relationship between syntax and reading comprehension. Layton, Robinson, and Lawson’s (1998) research with children (8–10 years old) who received training in syntactic awareness (six 30-min lessons) indicated that improved proficiency in syntactic skill was not related to improvement in the children’s comprehension skills. In this study, syntactic skills were measured by the Grammatic Comprehension and Sentence Combining subtests of the Test of Language Development-Intermediate [TOLD II]. The authors acknowledge that their results differed from previous researchers and suggest possible reasons for their different findings, including that either the training period was not long enough for an instructional impact to emerge on comprehension, or alternatively, that the reading measure (a comprehension subtest of a standardized Australian Achievement Test) was not sensitive enough to capture the participants’ comprehension gains.

In a more recent study, Cain (2007) investigated the associations between reading comprehension, and two measures of syntactic awareness (a grammatical correction task and a word order correction task) in children aged 7–10 years. Performances on both measures of syntactic awareness were correlated with reading comprehension scores for the older age group (9–10 year olds). However, the younger age group’s (7–8 year olds) comprehension scores were only significantly correlated with their word order correction scores, and not with their grammatical correction scores. Cain explains that this finding indicates that not all syntactic measures are necessarily equivalent.

Altogether, most of the studies examining the relationship in children between syntactic awareness and reading comprehension have found these skills to be related (Gillon & Dodd, 1995; Nation & Snowling, 2000; Mokhtari & Thompson, 2006). Children who are considered good comprehenders are thought to possess better syntactic skills than those considered poor comprehenders. Contrarily, two studies have indicated that this relationship may not always be evident (Cain, 2007; Layton et al., 1998). Based on the differences found within studies assessing the relationship between syntax and reading comprehension, it may be suggested that perhaps the relationship between syntax and reading comprehension differs based on the reading level of the participant, and on the type of syntax test employed.

Oral language processes in adults who struggle with reading

Although there is a lack of research in the area of adult literacy, studies have indicated that adults who struggle with written language also struggle with oral language. Research will be reviewed in the areas of phonology, oral vocabulary, and syntax to examine the trends found with the adult struggling reader population.

Phonology

In the area of phonology, a pattern is emerging which indicates that adult struggling readers have deficiencies in this area. Dietrich and Brady (2001) investigated phonological representations of less skilled adult readers. Three groups of subjects participated in the study, adult skilled readers, adult less skilled readers, and reading-age matched children in the seventh and eighth grades. When compared to the other groups, results indicated deficient phonological skills for the struggling adult readers in psuedoword repetition and in spelling. In addition, Greenberg, Ehri, and Perin (2002) found that struggling adult readers compared to reading matched children at the third-fifth grade levels were less apt to use phonological knowledge to help them decode nonsense words and spell real words. Finally, when reading words, Binder and Borecki (2007) found that adults reading below the sixth grade level use phonological skills less efficiently during word recognition and during activation of word meanings compared to adult skilled readers. This trend was also apparent within Sabatini, Sawaki, Shore, and Scarborough’s (2010) adult sample (between first-seventh grade reading level) as they seemingly relied on their sight word knowledge during a task of decoding.

Vocabulary

Studies of expressive and receptive vocabulary with struggling adult readers have demonstrated that they possess poor vocabulary skills (Byrne, Crowe, Hale, Meek, & Epps, 1996; Davidson & Strucker, 2002; Greenberg, Pae, Morris, Calhoon, & Nanda, 2009; Greenberg et al., 1997). For example, Byrne et al. (1996) examined the overall oral language skills of adults who read at an average reading level of fifth grade, and found that their receptive and expressive vocabulary skills were deficient. The adults’ deficiencies in their oral language skills were demonstrated by scores at the fourth and fifth grade level on subtests such as word definitions and synonyms. In other words, their oral language skills were more similar to their reading ages than to their chronological ages.

Davidson and Strucker’s (2002) research is consistent with Byrne et al.’s (1996) findings. Davidson and Strucker found that their adult struggling reader participants who identified words between the fourth and sixth grade levels possessed receptive vocabulary skills below the 10th percentile of the norming adult population. Greenberg et al. (1997) similarly found that her sample of adult learners (reading between the third and fifth grade levels) possessed very poor receptive vocabulary skills, with the adults scoring in the 1st percentile of the norming adult population. In a recent study, Greenberg et al. (2011) found adults reading between the third and fifth grade levels performed two standard deviations below the mean on a receptive vocabulary measure. Finally, Sabatini et al. (2010) reported that their adults demonstrated vocabulary skills similar to their reading ages (first to seventh grade) as opposed to their chronological ages which ranged between 16 and 76 years.

Syntax

Not much is known about the syntactic use of struggling adult readers. According to Nippold et al. (2005) typically, children gain an awareness of syntax at early ages and this awareness continuously matures into adulthood. Typical adults are generally assumed to have maturity in syntactic use. Unfortunately, no studies were found that examined the syntactic abilities of struggling adult readers. The closest study that was found focused on second year psychology students. Cupples and Holmes (1992) divided their students into two groups: those who “performed best” (fewer than 15% of errors on a reading comprehension test), and those who “performed worse” (more than 36% of errors). Syntactic skills were measured by asking participants to determine whether pairs of words expressed similar grammatical functional roles (for example, determining whether a pair of words represent nouns, verbs, adjectives, adverbs). Results indicated that the better comprehenders performed significantly more accurately on this syntactic judgment measure than the participants in the poorer comprehender group.

Aims of the present study

In this study, the syntactic ability of struggling adult readers was measured in order to answer the following questions: (1) what are the oral language abilities (with a special focus on syntax) of adults who have difficulties reading; (2) what do their syntactic error patterns look like; (3) is age related to their syntactic knowledge; (4) is their syntactic knowledge correlated to other oral language and written language skills; (5) do their oral language skills (with a particular focus on syntactic knowledge) account for significant variance in their reading comprehension abilities?

Research indicates that struggling adult readers have poor phonological and vocabulary skills (e.g., Binder & Borecki, 2007; Greenberg et al., 2011). Therefore it was predicted that the struggling adult readers in this study would also demonstrate low syntactic abilities. Examination of the error patterns of struggling adult readers was chosen as a purely exploratory analysis because descriptions of syntactic errors in this population are lacking. By examining the types of syntactic errors we hope to better understand the types of errors made by struggling adult readers. Such information may eventually lead to development of a theory regarding syntactic production in this population. The third question regarding age, is similarly exploratory, and is based on what is known about typical development; that syntactic development continues into early adulthood (20–29 years) and remains stable into the middle ages (40–49 years) and beyond (Nippold et al., 2005). The fourth and fifth questions regarding the relationship between syntax and oral and written language skills are exploratory, but it is hypothesized that the struggling adult readers’ syntactic skill performances will be related to their other oral language and written language skills.

Method

Participants

Participants included 82 struggling adult readers enrolled in an adult literacy program, located in the southeastern United States. Seventy percent of the students were female and the student’s ages ranged between 16 and 63 years (M = 35, SD = 15.49). The majority of the students were African American (88%), followed by Caucasian (7%), Other/Mixed ethnicity (4%), and Hispanic (1%). All students were native English speakers. The data in this study are from a larger study1 in which the researchers were interested in examining intervention effects on the reading skills of struggling adult readers. To qualify for this intervention study, the participants had to read words on the Woodcock-Johnson III (WJ III) (Woodcock, McGrew, & Mather, 2001) Letter-Word Identification subtest, between the third and fifth grade reading equivalency levels.

Materials

As part of the larger study (See foot note 1), all participants took a battery of tests. This study focuses on a subset of the tests.

To assess oral language skills, the following assessments were used:

Syntactic skill

To assess the syntactic skills, the Test of Language Development- Intermediate (TOLD-I:3; Hammill & Newcomer, 1997) was administered. According to the test manual, the TOLD I:3, is based on the theoretical work of a variety of linguists and psycholinguists (i.e., Bloom, Chomsky, Vygotsky). Therefore, the underlying theory considers syntax to be the structure of language where knowledge of the syntactic system enables the speaker to construct and understand novel sentences, by correctly sequencing words and by the appropriate use of inflections.

The TOLD-I:3 has been normed on individuals ages 8–12.11 and is used to assess understanding and use of different aspects of syntactic ability. The test battery consists of six subtests, with three focusing on syntax (Grammar Comprehension, Sentence Combining, and Word Ordering). In the larger intervention study, only the Word Ordering subtest was administered, and therefore is the syntax measure analyzed in this paper. The Word Ordering subtest controls for the influence of vocabulary and memory. According to the test manual, knowledge of vocabulary is controlled by using words that are below the third grade level. The influence of memory is minimized by limiting the number of words in the word sequences to seven. The word sequences chosen allow the participant to construct a variety of different sentences including simple, complex, and compound sentences and to display competence in using embedded phrases, forming questions, and transforming verbs. To administer the subtest, the examiner orally presents a series of randomly ordered words and directs the participant to repeat the words by reordering them to form complete, correct sentences. This subtest is untimed, however if the participant does not respond within 10 s, the tester proceeds to the next item. Before items are presented to participants, there are two practice items. In this study, all participants indicated understanding of the task, by being able to respond correctly to the practice items.

Expressive and receptive vocabulary

To assess expressive vocabulary, the Boston Naming Test (BNT; Kaplan, Goodglass, & Weintraub, 2001) was administered. This unstandardized assessment measures expressive vocabulary by presenting drawings in order of decreasing frequency and asking participants to verbally label the drawings.

To assess receptive vocabulary, the Peabody Picture Vocabulary Test, 3rd edition (PPVT-III; Dunn & Dunn, 1998) was administered. The PPVT assesses the extent of students’ knowledge of word meanings. It was normed on people ages 2–90+. Designed for use as a measure of receptive vocabulary, this test requires participants to select one of four pictures that best tells about the target word.

Phonological processing

To assess phonological processing skills, the Comprehensive Test of Phonological Processing (CTOPP; Wagner, Torgesen, & Rashotte, 1999) was administered. The CTOPP is designed to assess the phonological processing skills of individuals. For this study two core subtests were used: (a) the Elision subtest and (b) the Blending Words subtest. Both subtests were normed on people ages 5–24. The Elision subtest, requires the examinee to repeat aloud a target word then identify elements of it on demand (e.g., after correctly repeating the target word ‘bold,’ the task is to report the sound produced when the /b/ sound is dropped). The Blending Words subtest involves listening to sounds and combining these sounds into words.

To assess written language skills the following assessments were used:

Word identification

The Letter-Word Identification subtest from the WJ-III (Woodcock et al., 2001) was administered. This test has been normed on people ages 5–80+. This test requires participants to identify words of increasing difficulty.

Decoding

The Word Attack subtest from the WJ-III (Woodcock et al., 2001) was administered. This test has been normed on people ages 4–80+. Participants were given pseudo words arranged in sets of increasing difficulty (from consonant- vowel-consonant words to polysyllabic strings) to read.

Reading fluency

The Fluency subtest from the WJ-III (Woodcock et al., 2001) was administered. This test has been normed on people ages 6–80+. This measure assessed the participants reading speed and rate within a 3-min time limit. The task required the participants to silently read sentences and circle yes or no to indicate whether the sentences are true or false. During test administration, the difficulty level of the sentences gradually increases.

To assess reading comprehension, the following assessment was used:

Reading comprehension

The Passage Comprehension subtest from the WJ-III (Woodcock et al., 2001) was administered. This test has been normed on people ages 2–80+. This test required the participant to read short passages (usually two to three lines) and identify a key word missing from the passages.

To assess working memory, the following assessment was used.

Working memory

The Digit Span subtest from the Wechsler Adult Intelligence Scale, 3rd Edition (WAIS-III; The Psychological Corporation, 1997) was administered. This test has been normed on people ages 16–89. This test requires participants to repeat presented number sequences in the same order as presented (Forward Item) or in the reverse order (Digits Backward).

Procedure

Testing was administered one-on-one by graduate research assistants who received intensive training in adult literacy sensitivity, as well as training in general and specific testing techniques. Before testing participants, they practiced with a psychometrist and tested their first few adults with the psychometrist present.

Data analysis

Raw scores were used in the descriptive, correlational, and regression analyses. All of the analyses are reflective of a group of 82 struggling adult readers, with the exception of the correlation between working memory (Digit Span) and the word ordering subtest (TOLD-I:3), which is representative of 51 participants. For the error analysis, the types of errors participants produced on the TOLD-I:3 were identified and coded into five different categories: grammar, key word, novel word, position, and word order. Grammar errors included changes in subject verb agreement, changes in verb tense, fragments, or ellipsis. Key word errors included a change or deletion of words. Novel word errors included unnecessary words added to the sentences. Word order errors included a change in sentence type (e.g., declarative to imperative). Position errors included instances when the intended subject was turned into the object, the object turned into the subject, or both the subject and object were inverted. Table 1 provides examples of the error analysis for the TOLD-I:3.
Table 1

Example of TOLD error analysis

Category

Stimuli

Book, read, he, the

Target response

He read the book.

Participant response

He likes to read his books.

Grammar

Correct

Key word

“The” (omitted), “book” (changed)

Novel word

“Likes, to, his” all novel words

Word order

Correct

Position

Correct

Results

Oral language abilities

What are the oral language abilities (with a special focus on syntax) of adults who have difficulties reading? As indicated in Table 2, similar to previous studies, (e.g., Binder & Borecki, 2007; Greenberg et al., 2011) the adults performed very poorly on the phonological and vocabulary tests. Of specific interest to this study, performance on the TOLD-I:3 was also very low, with the adults on average, only completing half of the assessment before reaching failure. Additionally, on average the adults’ performances were considered equivalent to children’s performance on the syntactic task, with mean age equivalency levels of less than 9 years of age.
Table 2

Raw score and standard score performance of struggling adult readers on oral language and reading related tests

Test

Raw score

Grade (a)/Age (b) Equivalency

Standard score

n

Range

Mean

SD

Mean

Mean (SD)

TOLD-I:3

82

0–19

10.67

4.55

8.89 (b)

n/a

BNT

82

1–51

36.36

9.33

n/a

n/a

PPVT-III

82

84–173

138.88

18.21

11.52 (b)

74.44 (9.52)

CTOPP-Elision

82

0–18

5.87

2.50

1.51 (a)

n/a

CTOPP-blending

82

0–13

5.13

2.98

.98 (a)

n/a

WJ- word attack

82

2–26

11.56

5.43

2.61 (a)

73.25 (10.34)

WJ- fluency

82

18–68

38.09

8.65

4.16 (a)

79.14 (6.53)

WJ-comprehension

82

12–36

26.24

4.44

3.62 (a)

77.70 (11.36)

WJ-Word ID

82

42–57

48.93

4.31

4.10 (a)

78.62 (9.42)

Standard scores are unavailable for the TOLD-I:3, BNT, Elision, and Blending, since these measures do not provide norming information appropriate for this study

TOLD-I:3 Test of Language Development, BNT Boston Naming Test, PPVT-III Peabody Picture Vocabulary Test, CTOPP Comprehensive Test of Phonological Processing, WJ Woodcock Johnson

Since all items were verbally administered, it could be argued that the participants’ low performance on the test was more indicative of working memory issues rather than syntax knowledge. To investigate this possibility, scores on the Digit Memory Span Test were obtained for 51 of the participants. Results of the correlation analyses of the Word Ordering performance and the forward, backward, and forward + backward scores indicated that word ordering performance was not associated with working memory performance, at least as measured by these tests (forward, r = .14, p > .05; backward, r = .02, p > .05; forward +backward, r = .08, p > .05).

To further understand the syntax abilities of adults who have difficulties with reading, a regression analysis was performed to analyze which other oral language factors account for variance in the syntactic knowledge of the participants. Results (see Table 3) indicated that the adults’ oral expressive vocabulary (BNT) accounted for the most variance (13%) followed by phonological processing (blending) (8%). Oral receptive vocabulary (PPVT) and elision performances did not significantly account for variance in syntactic performance.
Table 3

Hierarchical regression assessing prediction of syntactic skills (TOLD)

Step and predictor

Syntactic skill (TOLD)

F change

r² change

β

1. BNT

11.87

.13*

.35*

2. PPVT

.038

.00

−.03

3. Elision

7.79

.08*

.29*

4. Blending

.412

.00

.07

* p < .05

Syntactic error patterns

What do the adults’ syntactic error patterns look like? TOLD-I:3 errors were classified into five categories: grammar, key word, novel word, position, and word order. Prior to analysis for category differences, error scores were computed for every participant: The numbers of errors in each category were divided by the total number of errors, producing percentage scores. As indicated in Fig. 1, analysis of the group’s error patterns indicated that most of the participants’ errors were found in key words (40%), followed by novel words (28%), grammar (13%), word order (12%), and position (7%).2 In other words, a dominant pattern reveals that the adults changed or omitted words that were essential to the order of sentences. They also had difficulty solely using the provided words to make sentences.
https://static-content.springer.com/image/art%3A10.1007%2Fs11145-011-9324-9/MediaObjects/11145_2011_9324_Fig1_HTML.gif
Fig. 1

Comparison among error patterns on word ordering task (bars represent standard error of the mean)

A repeated measures analysis of variance (ANOVA) was conducted to determine if the percentage of error types were significantly different from one another. Results indicated that sphericity was violated, therefore Greenhouse-Geisser is reported. The repeated measures ANOVA with types of errors (grammar vs. key word vs. novel word vs. word order vs. position) as the within- subjects factor revealed a significant error difference, F (3.26, 160.92) = 97.73, p < .001, η² = .54. Post hoc Tukey tests indicated significant differences between all the error types with the exception of the differences between grammar and word order errors (M = 1.19, SD = 1.26 vs. M = 1.14, SD = .97) and word order and position errors (M = 1.14, SD = .97 vs. M = .68, SD = .90).

Additional analyses were conducted to analyze whether different groups of our struggling adult reader sample (based on age or performance on the administered measures testing oral and print based skills) showed different error type patterns. Significant correlations were not found between age and error type. However, with the oral and written skills that were tested, one relationship was found, with grammar errors and fluency being related (r = −.233; p < .05) and grammar errors and elision being related (r = −.245; p < .05). Although the correlations are of a small strength, it does indicate that there is a trend for participants who had higher fluency and elision skills to produce fewer grammar error types of responses.

Syntactic knowledge and age

Is age related to the adults’ syntactic knowledge? A significant correlation between performance on the TOLD-1:3 and age was not found (r = −.14; p > .05).

The relationship between syntactic knowledge, reading, and other language tests

Is the adults’ syntactic knowledge correlated to other language and reading skills? A correlation analyses was conducted to determine whether syntactic ability could be associated with additional oral language and reading skills. As indicated in Table 4, positive correlations were found between syntactic performance and oral vocabulary, elision, reading comprehension, and reading fluency. Specifically, the adults’ oral syntax had small to moderate correlations with expressive vocabulary (r = .36; p < .01), elision (r = .33; p < .01), and receptive vocabulary (r = .27; p < .01). A small to moderate relationship was also found between syntactic skill and reading comprehension (r = .37; p < .01) and reading fluency (r = .22; p < .05).
Table 4

Correlations among syntactic skill, oral language measures, written language measures, and reading comprehension

 

1.

2.

3.

4.

5.

6.

7.

8.

9.

1. BNT

        

2. TOLD

.36**

       

3. Elision

.14

.33**

      

4. Blending words

.14

.20

.38**

     

5. PPVT

.79**

.27*

−.04

.01

    

6. WJ fluency

−.03

.22*

.22*

−.04

−.02

   

7. Word attack

.00

.01

.50**

.36**

−.18

.25*

  

8. WJ comprehension

.39**

.37**

.33**

.07

.28**

.33**

.29**

 

9. WJ word ID

.26*

.09

.31**

.02

.19

.27*

.45**

.40**

** Correlation is significant at the .01 level (2 tailed)

* Correlation is significant at the .05 level (2 tailed)

Do the adults’ oral language skills (with a particular focus on syntactic knowledge) account for significant variance in their reading comprehension abilities? To examine the unique contribution of the adults’ syntactical knowledge on their reading comprehension performance, hierarchical regression analyses were conducted. In the first regression analysis, oral vocabulary (PPVT and BNT) were entered in the first step, followed by syntax (TOLD) in the second step. The adults’ oral vocabulary accounted for the largest amount of variance (15%) followed by syntax which contributed 6%. In the second regression analysis, Word Attack, Word ID, and Fluency were entered in the first step, followed by oral vocabulary (PPVT, BNT) in the second step), and syntax (TOLD) in the third step. The adults’ reading skills accounted for the largest amount of the variance for reading comprehension (22%) followed by oral vocabulary which contributed 12%. Syntactic ability added no significant predictive value to the regression equation. These findings are demonstrated in Tables 5 and 6.
Table 5

Hierarchical regression assessing prediction of reading comprehension

Step and predictor

Reading comprehension

F change

r² change

β

1. PPVT

 BNT

7.21

.15*

.44*

 

−.68

2. TOLD

6.24

.06*

.27*

* p < .05

Table 6

Hierarchical regression assessing prediction of reading comprehension

Step and predictor

Reading comprehension

F change

r² change

β

1. Word attack

 Word ID

 Fluency

7.32

.22*

.10

.28*

.23*

2. PPVT

6.62

.12*

.01

 BNT

  

.35*

3. TOLD

3.77

.03

.20

* p < .05

Discussion

This study examined the syntactic ability of struggling adult readers who recognize words between the third and fifth grade levels. It was conducted because adults’ syntactic ability is an important aspect of oral language, which is not well documented, in previous research of struggling adult readers. Results indicated that this sample of struggling adult readers performed very deficiently on the Word Ordering subtest of the TOLD-I:3. Typically, adults are expected to exhibit maturity and complexity within syntactic tasks (Nippold et al., 2005, 2007). This was not the case with these adult participants, who performed similarly to what may be expected of children who read at an elementary level (e.g., Bentin et al., 1990; Gillon & Dodd, 1995). The participants’ syntactic knowledge was not correlated to their ages or memory, suggesting that regardless of age or memory, struggling adult readers have weaknesses in syntactic knowledge.

As a result of their skill performance, the extent of the struggling adult readers’ error patterns on the TOLD-I:3 were not surprising. On the word ordering task, the adults used key words incorrectly and added unnecessary words in the sentences. Grammatical errors (e.g. subject- verb agreement errors) and word order errors (the target sentence type is changed to a response sentence type) were also apparent. Research has found that typically developing children appear to learn basic word order quickly. It is believed they are able to make word patterns that are sensitive to the meaning carried by word order as they continue on to create longer and more complex sentences (Barrett, 1999; Hoff, 2009). With the adults in this study, their errors demonstrated that their syntactic skills were low, and more similar to children who are considered poor readers.

Further research is warranted to explore whether the errors exhibited by the adult participants in this study are unique to struggling adult readers, or whether they are similar to other populations, such as those who are developmentally delayed or learning disabled. Additional research is also warranted to explore whether there are different profiles for struggling adult readers, such that some readers are more likely to exhibit more error types than others.

The struggling adult readers’ syntactic skills were found to be related to reading comprehension and fluency. This relationship is similar to studies with children, which have found syntactic ability to be closely related to reading performance (e.g., Gillon & Dodd, 1995; Mokhtari & Thompson, 2006; Nation & Snowling, 2000). However, despite the similar findings, the correlations in the present study were not as strong as reported in other studies. Low correlation strengths are reported in other studies of struggling adult readers who read at the third-fifth grade levels, and may be indicative of poor integration of skills (e.g., Greenberg et al., 1997, 2011). Further research is clearly warranted to further explore why weak and/or no relationships are found between various oral and written literacy measures for adults who read at these levels.

Similar to other studies, the adults’ other oral language skills (expressive and receptive vocabulary and phonological processing) were closely related to their written language skills (e.g., Dietrich & Brady, 2001; Greenberg et al., 1997; Hoover & Gough, 1990; Nation, 2004). These findings support our hypothesis that the struggling adult readers’ syntactic skill performances are related to their other oral and written language skills.

The adults’ oral vocabulary and syntactic skills individually predicted reading comprehension, independent of their written language skills. However, when taking into account the adults’ written language skills (decoding, word identification, fluency) the adults’ syntactic skills did not individually predict reading comprehension, yet their other oral language skills such as oral vocabulary did. One major limitation to this study, is the fact that the TOLD-I:3 is not standardized on adults. A major problem in the area of adult literacy research is the paucity of tests that struggling adult readers can take (i.e., items start at low enough levels) to measure constructs of interest. Until such tests are developed, researchers need to rely on tests, which may or may not be valid and reliable measures of adult struggling readers’ skills (Greenberg et al., 2009).

Altogether, the present study found weaknesses in the syntactic ability of struggling adult readers. These findings suggest that struggling adult readers perform similar to children who are either learning to read or are considered poor readers. The findings of this study have relevance for the adult literacy field by providing exploratory information on an area (syntax and struggling adult readers) that is lacking. However, more research with adults is necessary in order to make any claims about the relationships between syntax and their reading abilities.

In this study, the participants’ deficient performance on an isolated syntactic task, adds to an accumulating literature base that demonstrates that struggling adult readers’ performance levels on a variety of oral language skills tests are commensurate with their reading levels, or at least lower than their predicted adult levels (e.g., MacArthur, Greenberg, Mellard, & Sabatini, 2010; Nanda, Greenberg, & Morris, 2010). Future research is warranted to try to uncover the underlying reasons for these deficits. Do these struggling readers have developmental language deficiencies from childhood that manifest themselves in both oral and written language skills? Are their poor language skills a consequence of poor educational opportunities (at home, at school, and at work) where advanced language skills are not in demand and therefore not fully developed? These types of questions may be considered and necessitate longitudinal studies (i.e. preschool to adult years) to understand the origins of the pervasive deficits that are being found with struggling adult readers.

Future research may also want to investigate the relationship between struggling adult readers’ oral syntax production and their written syntax production. As indicated by Cleland and Pickering (2006) there is controversy in the field whether the underlying mechanisms for syntax construction in speaking and writing are the same. Therefore, in addition to further research on struggling adult readers and their syntax abilities, much more research is warranted with non struggling adult readers. As Nippold et al. (2005) write: “Little is known about language development in adults, because this has not been widely investigated.” (p. 1051).

This study focused on adults who read at elementary school aged levels and only utilized one type of syntax test. It would be interesting to replicate this study with various syntax tests (such as Grammar Comprehension, Sentence Combining and Word Ordering) with a range of adult readers, from those who can barely read at all through the expert adult reading level. Perhaps the relationship between syntax and reading comprehension differs based on the reading level of the participant, and on the type of syntax test employed. For example, in Cain’s (2007) study, performances on a grammar correction task and a word order correction task were correlated with reading comprehension scores for 9–10 year olds, but for 7–8 year olds comprehension scores were only significantly correlated with their word order correction scores, and not with their grammatical correction scores. This indicates that not all syntactic measures are necessarily equivalent for every age group.

Footnotes
1

This paper represents part of a larger study funded by the Eunice Kennedy Shriver National Institute of Child Health and Human Development, the National Institute for Literacy, and the U.S. Department of Education, Office of Vocational and Adult Education* grant # R01HD43801.

 
2

Analyses were also conducted using raw scores, with identical results found. .

 

Copyright information

© Springer Science+Business Media B.V. 2011