Multiple Text Comprehension for College-Level Readers: Instructions to Enhance Performance

Abstract: The purpose of this study was to determine the optimal method for instructing college-level readers how to use a self-explanation strategy when comprehending multiple texts, which is a difficult task even for advanced readers. Self-explanation has been well documented to enhance single text comprehension. Three experimental conditions tested the degree of explicitness necessary for pre-reading instructions to make an impact: a control condition where readers were instructed to comprehend texts well; a definition only condition where readers were given a definition of a self-explanation strategy and urged to use it during reading; and a definition plus modeling condition where readers were given the self-explanation definition and then a researcher modeled how to use it during reading. The results showed that college-level readers performed better on reading comprehension and written essay tasks when in the definition only condition compared to the control condition. It is concluded that college-level readers need instructions to self-explain to enhance multiple text comprehension but instructions do not need to be overly explicit in terms of how to implement the strategy during reading.

Purpose: Synthesizing and comprehending ideas across multiple text sources is a common task in both secondary and post-secondary education. Theoretically, it is viewed as a complex problem-solving task that requires the reader to monitor multiple task components at once, such as the goal of the task, the source of information, and how each text alters one’s developing mental representation of text information (Rouet & Britt, 2011). Yet, relatively little is known about how to enhance multiple text comprehension. The purpose of this investigation was to determine whether or not a self-explanation strategy may benefit multiple text comprehension given that self-explanation has been shown to be an effective strategy for single text comprehension (e.g., Ainsworth & Loizou, 2003; Ainsworth & Burcham, 2007; Ozuru, Briner, Best, & McNamara, 2010). It could be that self-explaining helps readers focus on how different texts relate to one another and to the readers’ pre-existing knowledge, which allows for proper synthesis among text ideas. A related question was, How explicit do self-explanation instructions need to be in order to benefit the reader? To meet the objectives of the study, college-level students were randomly placed in to one of three instructional conditions: a control condition where readers were instructed to comprehend the texts well; a “definition only” condition where readers were given a definition of a self-explanation strategy and urged to use it during reading; and a “definition plus modeling” condition where readers were given the definition and also were shown by a researcher how to use it during reading a similarly structured text. It was hypothesized that, given the literature on the benefits of modeling reading strategies (e.g., Bereiter & Bird, 2009; Davey, 1983), modeling a self-explanation strategy would be the superior condition in terms of participants' scores on a reading comprehension test and a written essay task that assessed participants’ ability to retain facts and integrate ideas across three texts compared to both the control and definition only conditions.

Method: Sixty-four college students participated in the study to fulfill a course requirement. The average age of the sample was 20.5 years. Participants were randomly assigned to one of three conditions, which resulted in 22 participants in the control condition and 21 participants in each the definition only and the definition plus modeling conditions.

Participants read three expository texts on e1ectrical circuits and how they must form a continuous, closed loop to emit e1ectricity. One text described e1ectrical circuits in general, a second text described lightening as an e1ectrical circuit, and a third text described e1ectrical circuits in the context of batteries. Text order was randomly assigned for each participant. Each participant read the three texts under their randomly assigned instructional condition and then completed the written essay task. The essay task required participants to explain in writing how they would describe e1ectrical circuits to a group of young students, as if they were a teacher. Finally, the participants completed a 10-question short-answer reading comprehension test designed by the researchers to assess their understanding of facts and bigger themes across texts.

The first and second author scored the essay task and the reading comprehension test responses independently. For both the essay and reading test, a pre-established rubric was used to score responses. Inter-rater reliability at or above .90 was achieved on both assessments.

Results:

Data were submitted to a multivariate analysis of variance (MANOVA) to determine the effect of instructions on both the essay and the reading comprehension task, and p = .05 was set as the rejection criterion for all analyses.
The results of the MANOVA showed that pre-reading instructions had an overall effect on both the essay task, F (2, 63) = 3.62, MSe = 16.61, p < .05, partial eta2 = .11, and the scores on the reading comprehension task, F (2, 63) = 3.92, MSe = 2.85, p < .05, partial eta2 = .11. See Table 1 for means and standard errors in each condition.
To break down the main effects further, post-hoc Scheffe tests were performed. For the essay task, the definition only condition yielded more comprehensive essay writing than the control condition, but the significance was marginal, p = .056. No other effects were significant. For the scores on the reading comprehension task, the Scheffe test results showed that the definition only condition yielded significantly more accurate responses than the control condition, p = .027. No other effects were significant.

Table 1

Means and Standard Errors for Essay Task and Reading Comprehension Scores as a Function of Instruction Condition

Instruction condition Essay Task Scores Reading Comprehension Scores
M SE M SE
Control 10.89 .80 6.28 .42
Definition Only 13.95 .89 7.70 .29
Definition + Modeling 11.21 .96 7.17 .37

Discussion

The objective of this study was to optimize multiple text comprehension by providing college-level students with instructions on the use of a self-explanation strategy, where degree of explicitness varied. The results of this experiment showed that the definition only condition, the instruction condition that simply gave readers the definition of a self-explanation reading strategy, yielded higher reading comprehension scores than the control condition. Furthermore, the definition plus modeling condition was statistically similar to both other conditions in terms of readers’ performance and, thus, was not the superior instructional condition as expected. The results also showed that the participants in the definition only condition wrote essays that were more comprehensive overall, albeit marginally, involving greater synthesis of the information contained across the three texts, than either the control or definition plus modeling conditions.

One thought-provoking implication from this study may be that college-level readers benefit most when instructions are abstract. It could be that some degree of ambiguity motivates advanced readers to engage in effortful processing in order to determine how to implement abstract strategies on their own. Related to this interpretation, skilled readers have been shown to perform better when causal connections between text sentences are ambiguous or implicit, requiring the reader to fill in the gaps (Linderholm et al., 2000; also see Duffy, Shinjo, & Myers, 1990). The same process could be at work in this study: Advanced readers’ multiple text comprehension may be maximized if they are left to interpret on their own how to best implement a self-explanation strategy.

From a theoretical standpoint, it appears that strategies that enhance readers’ abilities to monitor comprehension processes during reading are successful given the complexity of multiple text comprehension (Rouet & Britt, 2011). Self-explaining during reading may facilitate the making of connections between and across texts and also may enhance the ability to see errors in one’s understanding of text information. The results of this study are in line with the theory that the process of multiple text comprehension is heavily dependent on one’s ability to closely monitor comprehension.