In knowledge-construction content analysis data from discussion groups, the largest share of variance lies at the message level, with only small between-group and between-student variance.¶

Claim · All claims
Evidence · 1 study (2 entries) · causal r2 · q3
Currency · newest study 2007 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i? In the three-level null model, 94.53% of total variance in knowledge construction is between messages, 2.84% between students within groups, and 2.63% between groups. → De Wever 2007 q3 i? In the four-level null model, 91% of variance arises from differences between messages, 7% at theme level, and 2% at group level, with no variance at the student level. → De Wever 2007 (2)

Evidence¶

De Wever 2007¶

De Wever, B., Van Keer, H., Schellens, T., & Valcke, M. (2007). Applying multilevel modelling to content analysis data: Methodological issues in the study of role assignment in asynchronous discussion groups. Learning and Instruction, 17, 436-447. https://doi.org/10.1016/j.learninstruc.2007.04.001

q3 · i? · causal · r2

Random intercept null model from the three-level analysis (messages within students within groups) decomposing total variance; the group-level variance component was not significantly different from zero (χ² = 3.415, df = 1, p = .065).

"The null model shows that respectively 2.63%, 2.84% , and 94.53% of the total variance in students' level of knowledge construction is lin ked to differences between groups, between students within groups, and between students' messages."

De Wever 2007 (2)¶

De Wever, B., Van Keer, H., Schellens, T., & Valcke, M. (2007). Applying multilevel modelling to content analysis data: Methodological issues in the study of role assignment in asynchronous discussion groups. Learning and Instruction, 17, 436-447. https://doi.org/10.1016/j.learninstruc.2007.04.001

q3 · i? · causal · r2

Four-level random intercepts null model (messages within themes within students within groups) for the analysis of all four discussion themes; the article reports no part of the total variance assigned to the individual student level.

"91% of the variance arises from differences between messages within measurement occasions (χ² = 1440.268, df = 1, p < .001)."

Discussion¶

Related Claims¶

• Messages mentioning or discussing theory, containing minor or extensive summaries, or including content moderating reflect significantly higher levels of knowledge construction. — related
• Messages merely mentioning sources, containing organisational moderating, or introducing new points reflect significantly lower levels of knowledge construction, while discussing sources shows no significant difference. — related
• Students assigned the summariser role post messages reflecting a significantly higher level of knowledge construction than students without roles, while starter, moderator, theoretician, and source searcher roles do not differ significantly. — related
• Knowledge construction does not increase gradually across successive discussion themes: messages in theme 2 reflect significantly lower levels, theme 3 significantly higher, and theme 4 no significant difference from theme 1. — related