How Interest Effects Comprehension of Reading Passages of Various Difficulities
Education Psychology Fall 2000 index
Contents:
1. Abstract 2. Introduction 3. Methods 4. Results 5. Discussion 5. Works Cited 6. Acknowledgements
Abstract
Reading comprehension is influenced by many factors. This study focuses on two factors that may influence reading comprehension, interest for the topic of the reading material and the difficulty level of the reading material. The reading comprehension of each subject was tested on four different reading passages: interesting instructional, boring instructional, interesting mastery, and boring mastery. After defining "comprehension," the number of students who comprehended each type of passage was found in order to uncover the influences that interest and difficulty level have on reading comprehension. The results revealed that interest exerted a greater influence than difficulty level on reading comprehension. Further exploration into this topic is needed to determine the signifigance of these findings.
Introduction
Previous studies (Wade, 1992; Hidi & Anderson, 1992) have addressed similar issues of the influence of interest on reading. From these studies, we can make some predictions about what we might expect from our study.
Suzanne Wade (1992) discusses the detrimental effects of "seductive details" on learning important information from learning. These seductive details are by definition facts that are interesting to the reader. Drawing from this oversimplified version of Wade's findings, one might hypothesize for our study that students would comprehend less of the interesting passages than the boring passages. But upon closer inspection, important differences between Wade's findings and our study become apparent, and a clear hypothesis from Wade's findings becomes more difficult to make.
First, there's the issue of different kinds of interest. Wade's "seductive details" are elements purposely sought out because of they are interesting to most people. These elements includes certain topics, (danger, death, money, etc) as well as certain categories of events (surprise, personal relatedness, etc.)(Wade, 257-8). In short, these "seductive details" sound a lot like items of situational interest. The interest used in the 5/6 study is altogether another type of interest. By asking students to rate their interest in terms of knowledge, frequency of activity, and liking, the level of individual interest for certain activities was sought out. Thus, students in our study read passages involving topics for which they had a well-developed individual interest in, whereas Wade's subjects read passages embedded with elements that sparked situational interest. Given this rather crucial difference in type of interest involved (situational versus individual interest), it may be misleading to draw hypotheses from Wade's results.
That Wade's study involved interesting elements embedded within a larger passage about something else, whereas our study involved passages written explicitly about a topic of interest to the subject, stands as another important difference between Wade's study and our study. Wade asserts that interest enhances learning and recall. She further asserted that when interest was directed at unimportant and unrelated details that didn't contribute to the main idea, i.e., "seductive details," learning was interfered. However, when this same interest was directed at important details that did support the main idea, learning was enhanced. Thus, interest always makes for better recall. The factor that makes or breaks learning is whether a "seductive detail" or an important relevant detail is the piece of information that is, through the presence of interest, better recalled. In narrative texts, interestingness of details correlated highly with importance of details (Hidi & Anderson, 1992:225). Analogously, given that our interesting passages were written about a topic of well-developed interest for the subject, what was interesting in our passages were also important to the main idea. Therefore, even though both Wade's and our studies dealt with reading passages with elements of interest, whereas Wade's interesting elements were unimportant and unrelated to the main idea, our interesting elements were important and directly related to the main idea. This crucial difference makes it difficult for us to draw on Wade's results for our study.
One of Wade's side findings, however, can prove instrumental in pointing us toward a hypothesis. As mentioned earlier, in narrative passages (as opposed to other types of passages), what was interesting and what was important came into alignment. Narrative passages may be easier to understand because of this congruence between interest and importance passages (Hidi & Anderson, 1992). Our interesting passages were also characterized by a match between interest and importance. Therefore, we can draw the hypothesis that our interesting passages would produce higher rates of comprehension when compared to our boring passages.
Another difference between Wade's and our experiment is rather theoretical, but nonetheless worthwhile to point out. Wade's ultimate goal was to see how much information kids could learn from reading passages. The dependent variable they talked about was "recall," also referred to as "what is learned." Thus, the reading component itself takes a back seat, whereas focus is on the subject matter of the passage. This study, however, had for its ultimate goal the reading comprehension itself. How much of what the kid read did she understand? The subject matter of the reading passage, then, was only conveniently used as a dependent measure to gauge this understanding. Thus, while Wade focused on learning information, the 5/6 study focused on reading comprehension. The difference may exist only in the theoretical framework, however, because in the end, the dependent measures ("recall" for Wade and "comprehension" for the 5/6 study) taken in the two studies are very similar.
One particular study in Wade (Renninger & Wozniak, 1985) lends itself to providing some answers for our question. They found that three-year-olds exhibited high recall for an interesting object, even when the recall task itself was difficult. These results suggest that task interest has the ability to outweigh task difficulty in terms of influence on task performance. Drawing from these results, then, we can hypothesize that interestingness of passage has a greater impact on reading comprehension than does reading difficulty level.
Methods
This study was conducted as a subset of a larger study, entitled the 5/6 study, conducted by Ann Renninger, Program in Education, Swarthmore College. The 5/6 study was conducted at a suburban Philadelphia elementary school. All of the fifth and sixth graders in this school participated in this study. 379 students participated in this study, 197 fifth graders, and 182 sixth graders. Female students made up 192 of the students (103 fifth graders and 89 sixth graders) and male students constituted 187 of the participants (94 fifth graders and 93 sixth graders). The student body is largely white.
This study focuses on the interest and reading comprehension data collected in the 5/6 study. At the start of the 5/6 study, students filled out questionaires which asked the students to rate their interests in topics and their degree of involvement in those topics or activities. Having processed these questionaires, interest and boring reading passages were composed for each of the students. Students were judged to be interested in something if they had rated that activity high on their questionaire and if they were fairly highly involved in that activity. Students were also tested using neutral passages to determine the students' reading ability. Students were then given four reading passages as part of their normal school work. These four passages were composed of a interest mastery, boring mastery, interest instructional, and boring instructional passage. Mastery was defined as the reading level at which the student was currently at, and instructional was the reading level just above the student's current reading ability. The students responses to various questions based on the readings were then coded. Statistical analyses based upon these data were then run using the statistical analysis program, SPSS.
The first step of this study was to define comprehension. Focusing on points ("specific peices of information"), gists ("general ideas"), points across (either combining "information from two or more paragraphs" or "abstracting information that applied to two or more paragraphs"), and overall gists (general ideas about the entire passage), we decided that various combinations of these codes could function as legitimate comprehension (EdPsych Manual, 5/6 Study, 66-68). We originally defined 5 types of comprehension. Comprehension 1 consisted of all three paragraphs represented by either points of gists. We defined Comprehension 2 as two paragraphs represented by points or gists, and one overall gist. Comprehension 3 consisted of 2 paragraphs represented by points or gists, and two paragraphs represented at a higher level of difficulty. Comprehension 4 was defined as 2 paragraphs represented by points or gists, and one paragraph represented at a higher level of difficulty by eiher points of gists. Finally, Comprehension 5 consisted of two paragraphs represented by points or gists and one point across.
We decided not to use Comprehensions 3 and 4, as the data neccesary for those two comprehensions was not represented in the 5/6 study. At this point we decided to run frequencies of each of the types of comprehension (Comp. 1, 2, and 5) to find the number of kids who were comprehending in each way. The frequency of use of each type of comprehension was found for each of the passages used in the study: boring mastery, interest mastery, interest intstructional, and boring instructional. We also found the total number of children comprehending for each of the passages.
The next step we took was to determine the signifigance of these findings. We constructed a crosstabs table and ran a c2 analysis. However, our coding was shown to be wrong. These analyses need to be run again.
Results
The frequency runs showed the total number students who comprehended each passage was 153 for the Boring Mastery passages, 200 for the Interest Mastery passages, 181 for the Interest Instructional passages, and 130 for the Boring Instructional passages. Comprehension 2 was the most frequently appearing comprehension that we defined. These results and the results of the frequency runs for each type of comprehension are summarized in Table 1 and Figure 1, below.
Boring Mastery
Interest Mastery
Interest
Intstructional Boring
Intructional # Using Comprehension 1 # Using Comprehension 2 # Using Comprehension 5 Total # of Comprehenders
Table 1: Frequencies of occurance of
each type of comprehension, as well as total number of
students comprehending, for each of the four passages: boring
mastery, interest mastery, interest instructional, and boring
instructional.
Note that the numbers of students for each type of "comprehension" do not add up to the total number of students who comprehended. This is because many students fell into two or more categories of comprehension, and these students were only counted once in the "total comprehension" category.
Figure 1: Bar graph showing the number of students who comprehend each passage using the different comprehensions and the total number of students who comprehended each passage. Comprehension 1=red; Comprehension 2=blue, Comprehension 5=green, Total comprehension-white.
Figure 2: Graphs illustrating the change in the number of students comprehending due to changes in difficulty level (left graph) and interest level (right graph). The difference between the lines representing students comprehending the mastery and instructional passages is about 20 students. The difference between the lines representing students comprehending the boring and interest passages is about 50 students.
Discussion
Results I
Our first set of results showed that interest is more influential than difficulty level in determining reading comprehension. Consequently more students comprehended the interesting instructional passages than students who comprehended boring mastery passages. These results fit neatly with our hypothesis drawn from previous literature. In our "interesting" passages, interesting details were usually also important details related directly to the main idea &emdash; similar to Wade's narrative passages. And like the narrative passages, our interesting passages were often comprehended, in comparison with the boring passages.
Our results make sense in light of Renninger and Wozniak's (1985) study with three-year-olds. Like the three-year-olds who recalled interesting objects even when the task was difficult, the students in our study were more likely to comprehend an interesting passage even when it was difficult than a boring passage that was easier.
These results also speak to Nina's tutoring experiences at Rose Valley. Nina's tutee does not like to write because it is very difficult for him. The results of this study, applied more generally, suggests that even relatively difficult tasks can be accomplished, given sufficient interest. Nina's tutee was very much interested in his flying machine project. Thus, when it was necessary that he wrote as a component of his flying machine project, he wrote (sometimes). His very high interest (building a flying machine) allowed him to perform a very difficult task (writing) (again, sometimes), as these results of our study would suggest.
Jill had a similar experience with her tutee. Jill's tutee had great difficulty with reading and writing. Whenever they needed to write something down, Jill's tutee had Jill do the writing. In order to overcome his inhibtions about writing, Jill brought her laptop computer to one of her tutoring sessions. Her tutee immediately began to explore and play with the computer. When they needed to write, Jill's tutee remained at the computer and typed. Her tutee's interest and excitement about the computer enabled him to complete a task that without the interest he would not even attempt. As out results suggest, interest often acts as an enabler, allowing people to overcome difficulties they face.
Results II
We attempted to construct a crosstabs table and run a c2 analysis on the number of students who comprehended each kind of passage. Unfortunately, the code was written incorrectly and yielded misleading results. Future correct analysis must be performed in order to confirm the findings from Results I.
One reason our analysis was only partially successful could be due to the multi-dimensional nature of "comprehension." In our study, children "comprehended" a passage when they wrote down any of a number of combinations of gists and points. As mentioned already, this "comprehension" is very similar to the "recall" variable used in Wade's analysis. In addition, when we speak of comprehension, are we speaking of a binomial (e.g. comprehension/no comprehension) or a continuous (e.g. excellent comprehension/good comprehension/ok comprehension) concept? In our study, comprehension was treated as a binomial concept. Thus, comprehension is a complicated concept not only because it is difficult to define, but also because it is difficult to conceptualize for measuring purposes.
Implications for Practice
In many remedial reading programs, students are given easier reading passages to read. As this study points out, passages that are very easy can also be quite boring. According to Results I, interestingness influences reading comprehension significantly more than difficulty level, thus the tactic used by remedial reading programs of decreasing interest and decreasing difficulty seems like a terrible combination. While the decreased difficulty level may increase reading comprehension by a small amount, the decreased interest level will decrease reading comprehension by significantly more than that gain. Such a combination could actually decrease reading comprehension! A much better combination, then, would be to increase interest and keep the difficulty level at a medium-to-challenge level. With such a combination, students' comprehension would increase a lot with the increased interest and decrease only a little from the higher difficulty level. The increased difficulty level also provides a challenge which can push the students to learn and expand further their reading skills. The potential results of our study speak for a stance against decreased difficulty levels and for increased interest levels in remedial reading programs.
What is useful in a remedial reading programs could also be useful in a mainstream reading program. Providing students with interesting reading passages allows reading difficulty levels to be increased while still achieving the same level of comprehension; and the increased difficulty levels would provide the challenge for students to improve still more. An upward spiral seems to ensue when individual student interests are catered to. Why not, then, at the beginning of the year, take a survey to discover each student's individual interests? Then, whenever possible, assign reading or homework that is related to each individual student's personal interest. Finding ways to cater to students' interests may prove to be a great way to push forward students' reading abilities.
Works Cited
Hidi, S. & Anderson, V. (1992). Situational interest and its impact on reading and expository writing. In K. A. Renninger, S. Hidi, and A. Krapp (Eds.), The role of interest in learning and development (pp. 215-238). Hillsdale, NJ: Lawrence Erlbaum Associates.
Renninger, K. A. (2000). Individual interest and its implications for understanding intrinsic motivation. In C. Sansone & J. M. Harakiewicz (Eds.), Intrinsic and extrinsic motivation: The search for optimal motivation and performance (pp. 373-404). New York: Academic Press.
Wade, S. (1992). How Interest Affects Learning From Text. In K. A. Renninger, S. Hidi, and A. Krapp (Eds.), The Role of Interest in Learning and Development (pp. 255-278. Hillsdale, NJ: Lawrence Erlbaum Associates.
Acknowledgements
Extensive help on the design of this study, as well as analyses of data, was provided by Kristina Lasher and Ann Renninger. Eric Martin assisted in the design of some analyses. Additional support was provided by Jen Jarson, Claire Feldman-Riordan, Vanessa A. Gorman, and Kim Pickney.