II.3.3.1 The methodological issue regarding the flow map method
In the recent years, the flow map method has been used to assess cognitive structure in a variety of science education settings. The flow map method was first introduced to science education community in Anderson and Demetrius (1993).
Anderson and Demetrius (1993) argued that, consistent with constructivist models of information analyses, the flow map required minimal intervention by interviewers and low inference for its construction, and could provide a convenient diagram of the sequential and multi-relation thought patterns expressed by the respondent.
In Anderson and Demetrius (1993), students’ information processing was not evaluated. But, in Bischoff and Anderson (1998), a checklist of cognitive
information processing categories were used as the analytical tool for assessing students’ cognitive skills. Students’ statements shown in flow maps were coded for content analyses into the following categories:
(1) Declarative knowledge, including categorical, relational, and systemic statements and principle and generalizations
(2) Cognitive reasoning, including causal reasoning, idiosyncratic reasoning, ascribed reasoning, analogical reasoning, anthropomorphic reasoning, and misconceptions (3) Unidentifiable statements
Similar to Bischoff and Anderson (1998), Tsai (1999) coded each idea presented in the participants’ flow maps into two levels: content level and logic level. The content level mainly deals with the description mode of the recalled information, while the
logic level is concerned with the cognitive reasoning of each idea shown in flow maps.
In Tsai (1999), the content level is divided into four categories (from lowest to highest): specifics, relations, transformations, and generalization; the logic level is divided five categories (from lowest to highest): defining, describing, comparing and contrasting, conditional inferring, and explaining.
In Tsai (2001b), the use of the flow map method coupled with a meta-listening technique was also introduced. A “meta-listening” technique is used to explore the learner’s additional conceptual knowledge (Tsai, 2001). With an original
tape-recorded interview record (i.e., the first phase), researchers will immediately replay the narrative of the learner’s tape-recording to provide an opportunity for him (or her) to recall additional concepts, which has not previously disclosed by
him/herself. The learner’s response in the meta-listening period is also
tape-recorded by a second tape recorder. In this way, the researcher can obtain a more complete picture of the learner’s cognitive structure in a non-directive manner (for the details of meta-listening technique, please refer to Tsai, 2001). More
importantly, the student’s metacognitive ability in reviewing his/ her own ideas can be also assessed (Tsai, 1998a, p. 416). In his study, a new cognitive structure variable, flexibility, was proposed to indicate student ides change as a result of the
meta-listening period.
In sum, the flow map method was proposed by Anderson and Demetrius (1993), and has been refined in the following studies (i.e., Bischoff and Anderson (1998), Tsai (1999)). Consequently, the flow map method is a more useful method to represent learners’ cognitive structures in light of its extent, richness, integratedness, flexibility, availability, and the information processing strategies learners used.
II.3.3.2 Exploring learners’ knowledge construction or development with the flow map method
49
How learners construct their knowledge structures is always highlighted by science education researchers. Tsai and Huang (2001) explored 28 fifth graders’
development of knowledge on biological reproduction. The participants in this study received three weeks of instruction about biological reproduction, and were
interviewed at weekly intervals throughout the instruction (three times) and two months after the instruction. The data collected from tape-recorded interviews were analyzed with the flow map method. According the data collected in their study, Tsai and Huang (2001) suggested three stages of cognitive structure development:
knowledge development, knowledge extension, and knowledge refinement. The findings in this study also imply that the rich connections between concepts and the use of higher-order information processing strategies may facilitate maturation of knowledge reconstruction and refinement.
By using a quasi-experimental research approach, Wu and Tsai (2005a)
conducted a longitudinal study with six instructional units, which explored the effects of long-term constructivist-oriented science instruction on elementary school students’
process of constructing cognitive structures. This study further suggested a four-stage model for students’ process of constructing cognitive structure under the constructivist-oriented science instruction, including “cognitive structure acquisition”,
“metacognition enrichment”, “cognitive structure integration”, and “cognitive structure refinement”. By comparing Tsai and Huang (2001) with Wu and Tsai (2005a), we can find that the three-stage cognitive structure development model proposed by Tsai and Huang (2001) are respectively similar to the first, the third, and the final stages of the model proposed in Wu and Tsai (2005a). Therefore, the model proposed in Wu and Tsai (2005a) could be viewed as a refined one of Tsai and Huang (2001).
To provide potential insights into the mechanisms of learners’ knowledge
construction, with three experiments, Tsai and Chou (2005) investigated the role of the “core concept” and the “anchored concept” within an individual learner’s
knowledge structure in his (or her) knowledge construction. In their study, students’
knowledge structure was represented by using the flow map method. They further defined that the concept connected with the most recurrent linkages was the core concept within one’s knowledge structure regarding a specific topic, while the concept connected with the second most recurrent linkages was the “anchored concept” within one’s knowledge structure regarding a specific topic. The results derived from their study revealed that the importance of the “core” concept and “anchored” concept in one’s knowledge construction.
II.3.3.3 The relationships between students’ performances on the variables obtained from their flow maps
The findings derived from previous studies also reveal the relationships between students’ performances on the variables obtained from their flow maps. These relationships can be summarized as the following assertions:
(1) Sufficient concepts or ideas are necessary for complex cognitive structures as well as the usage of higher-level information processing strategies:
Bischoff and Anderson (1998) interviewed thirteen 9th – and 10th graders four times (once before a unit on ecology and at three weekly intervals thereafter) to determine the pattern of their gain in knowledge and cognitive skills. On each occasion, the students were presented with a hand-drawn diagram of a food web and an aquarium. Then the students were asked to describe the food web and aquarium using a standardized set of questions. From the 13 students’ data, the highest
performing student and the lowest performing student were selected for more detailed analysis. With flow maps and a checklist of cognitive information processing categories were used as analytical tools. The data in their study revealed that the
51
amount and the quality of prior knowledge substantially influence gains in new knowledge and were closely linked to a capacity to apply higher order cognitive thought processes in constructing abstract knowledge. In addition, Bischoff and Anderson (2001) also reported that sufficient concepts or ideas are necessary in order to precede and promote the development of more complex ideational patterns in learners’ cognitive structures.
(2) Connections among existing concepts and higher-level information processing strategies may mutually reinforce one another:
Anderson et al. (2001) reported that the students, who displayed more linkages within their ideational networks tended to use a logical sequential rule of ordering their thoughts. The students’ usage of higher level categorical thought increased with their increasing amount of the linkages within their ideational networks. This study concluded that the linkages within their ideational networks were correlated to their usage of higher level categorical thought. Moreover, Tsai and Huang (2001) also indicated that the use of higher-level information processing strategies and the increase in networking connections among existing concepts may mutually reinforce one another.
(3) Students’ metacognitive engagement, revealed by the flexibility of their cognitive structures, is correlated with their use of higher-level information processing
strategies:
The results in Tsai (2001b) revealed that students who had more extended, richer, and more flexible cognitive structures were more oriented to use higher-level
information processing strategies, suggesting that learners’ usage of higher-level information processing strategies required their well-developed cognitive structures.
II.3.3.4 Learners’ achievement, performances on laboratory work, SEBs and their cognitive structure outcomes derived from flow maps
Some previous studies explored the differences on cognitive structure outcomes shown in flow maps between higher achievers and lower achievers. For example, Bischoff and Anderson (1998) found that the highest-performing student evidenced greater gains in use of higher order cognitive categories than the lowest-performing student. Similarly, Tsai (1998a) reported that students’ science achievement was correlated with many of students’ cognitive structure outcomes. Moreover, Tsai (2001b) further revealed that the students of higher science achievement displayed more extended, richer, more integrated, and more flexible cognitive structures than did their counterparts, and they also tended to have higher information retrieval rate and use higher-level information processing strategies more frequently. By
summarizing the findings above, higher achievers outperform lower achievers in the extent, the richness, the integratedness, the availability, and the flexibility of cognitive structures. Also, they tended to use more higher-order information processing strategies than lower achievers.
With the flow map method, Anderson et al. (2001) further examined the role of 72seventh graders’ ideational networks in laboratory inquiry learning and their knowledge of evolution. The participants’ performances on hand-on laboratory experiments were scored using a standard protocol or rubric, and their knowledge gains were measured by a final unit examination. Their written essays were also obtained, and four categorical thinking were identified accordingly. This study revealed that the network linkages with the students’ ideational networks were
positively correlated with their performance on hand-on laboratory experiments, their knowledge gains, and the extent of correct concepts contained in their essays.
Tsai (1998a) explored the interrelationships between eighth graders’ science achievement, scientific epistemological beliefs and cognitive structure outcomes after learning basic atomic theory. In this study, students’ scientific epistemological
53
beliefs were found to be significantly related to the structure of knowledge recall.
Moreover, students holding more constructivist-oriented views about science tended to recall more information, as well as show more richness, more flexibility and a higher precision of knowledge recall than those having empiricist-aligned epistemological beliefs. However, those students holding more sophisticated epistemological beliefs toward science were prone to display a slower information retrieval rate.
II.3.3.5 The effects of science learning activities on students’ cognitive structure outcomes
Educators have recommended many teaching strategies or models which are based upon the assertions of constructivism to promote students’ science learning (e.g., Novak & Gowin, 1984; Lawson, 2001; White & Gunstone, 1992; Tsai, 2003). In their study, Wu and Tsai (2005b) explored the effects of the constructivist-oriented instruction on fifth graders’ cognitive structures. The participants of this study were assigned to either a constructivist-oriented instruction group or a traditional teaching group. The research treatment was conducted for three weeks, and the interview data were gathered a week later after the instruction about biological reproduction.
The interview narratives were transcribed into the format of flow maps. Wu and Tsai (2005b) revealed that the students in the constructivist-oriented instruction group, in general, attained better learning outcomes about biological reproduction after instruction, both in terms of the extent of concepts and in the richness within their cognitive structures. Furthermore, the students in the constructivist-oriented instruction group were probably better organized and stored their ideas in a
higher-level mode of information processing, especially in the mode of “inferring or explaining.” However, high achievers and low achievers in the
constructivist-oriented instruction attained better usage of information processing
modes in somewhat different ways; only high achievers in the constructivist-oriented instruction group attained significantly better usage of high-order information
processing modes. To explore the effects of long-term constructivist-oriented science instruction on elementary school students’ process of constructing cognitive structures, Wu and Tsai (2005a) also conducted a longitudinal study. The research treatment of this study was conducted for five months, including six instructional units. Wu and Tsai (2005a) revealed that the students in the constructivist-oriented instruction group attained significantly better leaning outcomes in terms of the extent and integration of their cognitive structures, metacognition engagement, and the usage of information processing strategies. Moreover, both high achievers and low
achievers benefited from the constructivist-oriented instructional activities, but in different ways. Both high achievers and low achievers in the constructivist-oriented instruction group attained better usage of information processing strategies than their counterparts in traditional teaching group did; but only high achievers displayed better usage of higher-order information processing modes than their counterparts in
traditional teaching group did. In conclusion, the two aforementioned studies
suggest both higher achievers and lower achievers benefit from constructivist-oriented science instructional activities (no matter short-term or long-term), but
constructivist-oriented science instructional activities benefit higher achievers more.
Tsai (2000b) were conducted to examine the effects of STS instruction on
students’ cognitive structures. This study revealed that STS group student performed better in terms of the extent, richness and connection of cognitive structure outcomes than did traditional group students. Moreover, STS instruction was especially beneficial to students having epistemological beliefs more oriented to constructivist views of science, particularly in the beginning stage of STS instruction. The results
55
derived from Tsai (2000b) imply that learners’ scientific epistemological beliefs may be an important factor mediating the implementation of STS instruction.
II.3.3.6 The role of teachers’ knowledge structures in their ability to diagnose students’ conceptual understanding
Among the aforementioned empirical studies conducted with the flow map method, the flow map method was used to analyze students’ cognitive structures.
With the flow map method, preservice teachers’ knowledge structures were also probed in Bischoff (2006). In this study, 25 preservice teachers’ knowledge structures were explored first. Then, the relationship between their mastery of content knowledge and their ability to diagnose the strengths and weaknesses of a fourth grader’s videotaped explanations of a scientific phenomenon was also investigated. The results of this study revealed a positive correlation between knowledge structure complexity and participants’ ability to diagnose the child’s thinking on a novel molecular kinetics task. Bischoff (2006) further suggested the importance of teachers’ existing knowledge structures regarding a topic in diagnosing the child’s thinking on this topic.
II.4 Internet-based science learning, on-line searching activities, and information commitments