Introduction
The view that language should be divided into four skills – reading, writing, listening, speaking – has dominated second language (L2) teaching and learning for some time; perspectives have emerged that conceptualize language as holistic or skills as integrated. This construct of language is evident in language assessments, with an increase in ‘integrated skills’ tasks appearing in tests of writing or speaking. Integrated skills assessments have been defined in several ways (Knoch & Sitajalabhorn Reference Knoch and Sitajalabhorn2013) such as reading-into-writing performance with the reading content included in the writing or as tasks that use input, textual or visual, as stimulus for writing under the integrated umbrella (Plakans Reference Plakans2013). Even with these multiple interpretations questions arise: ‘Is that all?’ ‘Is there more to integration?’
Missing from L2 integrated skills assessment research and practice are tasks that start with writing, i.e writing-into-reading. In L2 reading instruction, this sequence is common as pre-reading activities often include writing as a way to prepare students to read about a certain topic or in a genre of text (Hirvela Reference Hirvela2016). Zamel (Reference Zamel1992) proposed using writing to engage students with reading and to help them see the two skills as shared acts of composing meaning, rather than viewing reading as simply static receptive absorbing of information. In first language (L1) literacy research, the complimentary relationship of writing to reading has been documented. In a meta-analysis of over 90 studies, Graham & Herbert (Reference Graham and Herbert2011) explored whether: (1) writing about reading improves comprehension, (2) learning to write supports learning to read, and (3) if writing more improves reading. They were able to confirm all of these areas of impact, concluding that writing is a means to improve reading ability.
The support for connecting reading and writing has spurred researchers in L1 and L2 literacy to study how writing can be integrated with reading and the outcomes of such synthesis. The research is mostly focused on writing during or after reading, such as notetaking, journaling, concept mapping, question response writing, and summary writing. However, a few studies have looked at the impact of writing-into-reading in the L2 context. For example, Wickramaarachichi (Reference Wickramaarachichi2014) focused on a writing task's effect on the sub-skill of inferential reading for 50 first-year, undergraduate English as a foreign language students in Sri Lanka. The results showed that the pre-reading writing had a positive impact on scores.
Given Graham & Herbert's (Reference Graham and Herbert2011) meta-analysis, with L1 research and the few studies with L2 readers, writing before reading seems to benefit comprehension. To develop assessments that elicit the best possible performance from test-takers, we argue for including writing before reading. However, there is a dearth of research in this area. To explore writing-into-reading integration, we created an integrated task that included writing before and after reading and conducted a study to explore the following research questions:
1. What processes are elicited by the writing-into-reading prompt?
2. Does writing before reading facilitate reading performance?
Methods
To consider the impact of writing before reading in an integrated assessment task, we conducted a verbal protocol (think aloud) study with follow-up interviews (these data are part of a larger study of shared processes in L2 reading and writing). Data were analyzed qualitatively for the two research questions, looking for emergent patterns across writers with the use of scores.
Eleven L2 writers who were preparing for, or enrolled in, academic programs at a North American university were recruited to participate in the study. The sample participants possessed the following L1s: six Chinese, two Arabic, two Kannada, and one Korean. Self-reported English proficiency test scores ranged from a TOEFL of 64 to 106. Given the exploratory nature of the qualitative study, the variation across participants was an asset.
The task designed for the study consisted of four sections: (1) free writing on the task topic, (2) reading two passages on the topic, (3) short-answer and multiple-choice questions about the passages that targeted comprehension, identifying main ideas, making inferences, and understanding vocabulary-in-context, and (4) a persuasive essay on the topic using the reading passages as support.
The topic, cell phone use while driving, was selected after piloting several options. The reading passages were screened to be accessible with a general understanding of the topic and at a difficulty level appropriate for the target participant level. A scoring scheme was developed for the reading and the final writing; two raters scored each test.
While completing the task, participants engaged in a concurrent think aloud followed by a short questionnaire and an interview. The think-aloud protocols entailed a training session to practice voicing thoughts audibly while completing a reading and writing task. The session was recorded in a computer classroom with dividers between participants and with a headset/microphone. Time was not restricted; however, most participants finished one to one-and-a-half hours. The last step was a recorded oral interview in which participants were asked more in-depth questions about their impressions of the task. The data from the think aloud and interviews were transcribed by one researcher and checked for accuracy by another.
The transcriptions were segmented into idea units for further coding. These data, along with test performances and scores, were used to answer each research question. To explain the process during writing before reading (RQ1), transcripts from the think aloud were coded for processes related to reading and writing (Saldaña Reference Saldaña2015) and then compared to arrive at coding categories. With this final coding scheme, all transcripts were coded then analyzed for the first research question. To investigate the effect of this initial writing on reading (RQ2), similar analyses were taken with the interview data in conjunction with reading section scores and the coding of the think aloud data.
Results
Processes elicited in writing before reading (RQ1)
The writers’ processes prior to reading revealed a good deal of time spent eliciting background knowledge about the topic of cell phone use while driving. Most writers considered if they had experience with the topic, ‘Ok. . .um, yes, I did use cell phones while driving’, said Ran.Footnote 1 Chan thought, ‘I'm not sure because I don't drive’. Writers also considered their current opinions on cell phone use. Rebecca thought aloud, ‘So, I don't think it's a good idea to speak on the cell phone while driving’. General knowledge was also considered during this writing; Rebecca elaborated on her knowledge about the dangers of cell phone use, ‘cell phones, they bring brain damages, means headaches and all’. The process of generating background knowledge was very prevalent in this initial writing stage.
The other major sub-process appearing in the writing-into-reading think aloud was composing written text. This composing occurred as it might for an essay writing task. First, the writers did some prewriting, which extended from their background knowledge thinking. For example, they fleshed out ideas, considered rationale to support their opinion, and articulated their theses. A few writers organized ideas before they wrote. Then the majority of the time was spent writing, frequently by first composing a sentence in their heads and then writing it fully. Half of the think-aloud segments in this task section were devoted to this composing process. Following writing, and in a few cases during, the writers would reread what they wrote, edit their writing, and, rewrite sentences. Even though this task was free writing, writers’ approaches were similar to an essay or in-class writing, but, some aspects of the think aloud reflected test-taking strategies, such as re-reading the prompts and checking for the completeness of the writing as per instructions. The analysis revealed that the major processes during writing-into-reading were elicitation of background knowledge and composing written text.
The effect of writing before reading (RQ2)
To explore the impact of this initial writing-on-reading, interviews, reading scores, and think-aloud and interview data were synthesized. When asked directly about the effect this initial writing had on their reading, only half of the participants (five) agreed that writing first helped them in reading the two passages. For these participants, the writing allowed them to think about the topic in detail before reading. Saleh described this, ‘Because I was brainstorming some ideas about, about the test, when I read some of, some I already write about.’ Anaya explained that writing first explicitly improved her comprehension of the readings, again due to writing on the topic first. Several writers pointed out that establishing their opinion on the topic through writing first was useful because, as they read the opinions in the reading passages, they could compare their argument with the authors’. For example, Jie said, ‘Ok, I think my point of view is kind of similar with the reading two, so when I read, I just would like to find something similar with my opinions when I read the reading one’.
One participant mentioned that writing first allowed her to read the passages in a ‘more critical way’. Essentially, the effect of the writing on reading was felt, by half the participants, as a means to bring forward their ideas on the topic, which (1) impacted meaning construction in their reading, and (2) supported their thinking about the argument regarding cell phone use while driving.
We compared reading scores to see if there was a visible impact for those who felt the writing facilitated their reading and those who did not. Mean scores were equivalent for the two groups. The mode was higher for those who felt the effect, while the median was lower. The numbers suggest that the impact felt by the participants did not directly translate into a higher score. However, given the sample size of 11, more research would be needed for meaningful conclusions.
The percentage of think-aloud segments in each coded sub-process in writing before reading reveals fairly similar patterns for high scoring and low scoring test-takers. The one difference of note may be in the attention at the word level, with lower scoring readers spending more time thinking on the topic at this level as they wrote and higher scoring readers focused slightly more at the sentence-level. For example, Jung, who scored 71% on the reading section, ‘moreover, moreover, hmmm, moreover...um. . . keep concentrate...no, different word’ while Jie, with a score of 100% thought, through the sentence, ‘Um. . .what's else. . .um. . .um (laugh), so maybe um, it depends on the, yeah, it depends on the individual differences’. This difference may show that writing first affects reading or it may reflect lower proficiency L2 readers’ use of bottom-up processes (Koda Reference Koda2004).
Concluding discussion
This exploratory study analyzed data from L2 writers completing a writing-reading-writing prompt to explore writing before reading in integrated assessment. The results revealed that the initial writing process elicited two main sub-processes: generating background knowledge on the topic and composing sentences. This finding reflects the potential for such writing to emulate a ‘pre-reading’ task, to prime test-takers to better comprehend and make meaning from the reading passages. The positive impact of writing on reading has been studied with L1 reader-writers (Graham & Herbert Reference Graham and Herbert2011) and our study suggests this may transfer to L2 skill integration. Writing before reading may also initiate the shaping of the test-taker's argument for persuasive writing after reading. In fact, all 11 participants stated in their interviews that the initial writing section impacted their writing in the final section. This perception was not as unanimous for the reading. About half of the writers felt that writing first supported their reading of the two passages. The other half did not feel this impact, and, while these writers might lack metacognitive awareness, further exploration by grouping the think-aloud data by perceived impact did not reveal substantial evidence for the effect of writing-into-reading. Further investigation of this relationship is recommended with a larger sample using quantitative methods.
In conclusion, we see potential of writing-into-reading to extend skill integration in L2 assessment. It gave writers the opportunity to think about the topic, establish a nascent position on an issue, and to warm up for later writing. All of these processes potentially allow test-takers to display L2 literacy abilities at their highest level, a goal that serves accuracy in assessing language. These benefits should be seen in balance with the challenge of adding more time to task completion. More research in this area is needed to build evidence for the usefulness and feasibility of this new kind of integrated skills assessment task.
