Exploring the effect of scripted roles on cognitive presence in asynchronous online discussions

Context

As researchers have paid attention to effective instructional strategies to promote students’ active knowledge construction, the benefits of using scripted roles in online discussions have been examined by many researchers (De Wever, Van Keer, Schellens, & Valcke, 2010; Wise, Saghafian, & Padmanabhan, 2012). Scripted roles require instructors to assign roles to students; roles also vary in their types, such as summarizer role when students are required to provide the summary of the weekly discussions (Schellens, Van Keer, & Valcke, 2005). Studies have also found that the types of scripted roles have different effect on students’ level of knowledge construction (De Wever et al., 2010; Gasevic, Adesope, Joksimovic, & Kovanovic, 2015). For example, students, who are assigned a moderator role, show increased quantity and diversity of participation in online discussion (Xie, Yu, & Bradshaw, 2014).

Studies have also confirmed other positive effects of scripted roles and, specifically, roles can facilitate students’ high level of cognitive presence (Gasevic et al., 2015). Cognitive presence, a core construct of the Community of Inquiry (CoI) framework, is associated with “the extent to which learners are able to construct and confirm meaning through sustained reflection and discourse in a critical community of inquiry” (Garrison, Anderson, & Archer, 2001, p.11). Specifically, the CoI framework suggests four levels of cognitive presence:  triggering event, exploration, integration, and resolution (Garrison, Anderson, & Archer, 1999). A triggering event, the initial level, of cognitive presence indicates “a state of dissonance” or “feeling of unease resulting from an experience” (Garrison et al., 1999, p. 98). The exploration level focuses on searching for new information, knowledge, and alternatives to address a problem. The integration level pushes students to integrate and combine the information learned. The resolution level focuses on implementation of the proposed solutions and application of newly created knowledge (Garrison et al, 1999).

Studies have reported that exploration level (44%) was the most coded indicator while the rate of integration (13%) and resolution (4%) was relatively few (Garrison et al., 2001; Rourke & Kanuka, 2009). It seems that students are not likely to reach higher level of cognitive presence (Kanuka, Rourke, & Laflamme, 2007). However, Swan, Garrison, and Richardson (2009) proposed that integration and resolution stages might actually be more likely to appear in   final assignments rather than in online discussions as students are demonstrating their  knowledge construction from the entire course. Therefore, considering that computer-mediated communication presents advantages for active knowledge construction, it is important to conduct an in-depth examination of how cognitive presence is expressed in relations to scripted roles. To fulfil this gap, the following research questions have been investigated: (1) what is the effect of scripted roles on the level of cognitive presence? and (2) how is cognitive presence expressed across different types of scripted roles when students participate in asynchronous online discussion?

Methods

A mixed method research design was used (Creswell & Plano Clark, 2011). Through the purposeful sampling, all participants (n=76) were undergraduate students enrolled in a six-week online course “Introduction to Nutrition” at a large eastern public university in summer 2014. The participants were gathered from two course sections, section one (n=39) and section two (n=37). All students were assigned to smaller groups with five-six students in each. Out of 76 enrolled in this course, 72 students participated in all four required online discussions. This study results report on the findings generated by the 72 students. To facilitate discussions, all three required roles (starter, wrapper, and skeptic) were assigned randomly at the beginning of the semester, via a table provided to students. However, there were discussions when the same students didn’t play any roles. Specific tasks were associated with each role: the starter was responsible to post a preliminary response to get the discussion started in response to the discussion prompt. Starters also suggested a structure to the discussion; the skeptic challenged points made by other students and keeping the discussion balanced by bringing up other points of view; and the wrapper summarized the points that were made during the discussion with regard to a particular reading.

Qualitative data were collected from students’ discussion postings. All students’ postings (n= 1271) were coded and categorized into the four levels provided in Garrison et al.’s model which consists of: (1) triggering event, (2) exploration, (3) integration, and (4) resolution. For the next step, qualitative data were transformed into quantitative data (frequencies) to run descriptive and non-parametric statistical analysis. Descriptive statistics was used to analyze the frequencies of codes with the role and without role by four levels of cognitive presence across all four discussions in the course. Then, a non-parametric statistical analysis of a chi-square test for independence was applied to analyze associations between each type of scripted roles and the levels of cognitive presence. To establish validity and reliability of the data, two researchers independently coded each online discussion. Results were then compared and consensus building allowed for 95 % inter-rater reliability.

Results

RQ1: what is the effect of scripted roles on the level of cognitive presence?

In order to identify the effect of scripted roles on the level of cognitive presence, the percentages and frequencies of the codes generated based on when students were assigned a role (with the role) and when the same students were not assigned  the role (without the role)   were calculated. As the result, integration level was the most frequently coded category across both conditions (67.25% with the role and 72.17% without the role). No frequencies were found at the resolution level with the role. The resolution level was very low (0.43%) for those students, who didn’t play role during some online discussions but participated. The results of a chi-square test for independence indicated there was a significant association between the roles and the levels of cognitive presence (χ² (1, n=1036) = 13.38, p = .004).

The percentages and frequencies of the codes generated by each type of roles (starter, skeptic, and wrapper) were examined to explore the effect of each type on the level of cognitive presence. The majority of the postings occurred at the integration level across all three types of roles (61.83% for starter; 69.92% for skeptic; and 70.03% for wrapper) while triggering was the least frequent level (4.58% for starter; 4.22% for skeptic; and 5.54% for wrapper). No frequencies were found at the resolution level for any of the roles. The results of the chi-square test for independence did not indicate any significant associations between each type of the   role and each level of cognitive presence (χ² (1, n=806) = 6.75, p = .15).

RQ2: how is cognitive presence expressed across different types of scripted roles when students participate in asynchronous online discussion?

For the second question, percentages and frequencies generated by each type of scripted roles across the four online discussions were examined. To examine whether there was an association between the types of roles and the online discussions at each level of cognitive presence, a chi-square test for independence was run. The results of a chi-square test indicated significant association between the types of scripted roles across all four online discussions at the triggering event level (χ² (1, n=39) = 22.29, p = .001, V=.53), the exploration level (χ² (1, n=225) = 66.93, p = .000, V=.34), and the integration level (χ² (1, n=542) = 146.55, p = .000, V=.33). The effect size was large at the triggering event level .53, the exploration level .34, and the integration level .33 showing a strong association between the types of scripted roles across all four online discussions.

Discussion

The results of this research corresponds to the findings of previous research that scripted roles can be an effective strategy to improve both learning processes and outcomes (Strijibos & Weinberger, 2010). Even though previous studies have reported that the exploration level was the most coded level of cognitive presence (Kanuka et al., 2007), this study reports the opposite findings echoing the results reported by Akyol and Garrison in 2008 and 2011. Like the Akyol and Garrison’s studies (2008, 2011), the great majority of the students’ discussion postings in this study stayed at the higher levels of cognitive presence – integration. One of the reasonable explanations for this positive effect could be that all the students in this study had an opportunity to play each of the assigned roles at least once during the online discussions. Another explanation could be the type of questions. The questions in each online discussion required a collaborative decision making as a group. For example, in the third online discussion, which had the highest percentages across all three roles (average 74%), the discussion questions encouraged students to find examples of marketing techniques as a team. Then, as a team, they were required to identify each of the eleven marketing techniques on the collected nutritional products’ websites. This study contributed to the field by revealing the effectiveness of scripted roles in asynchronous threaded discussions. The researchers will share practical implications to help practitioners and educators design effective online discussions.