Differences and Relations Among Teacher Confidence, Use, and Performance in Data-based Decision-Making

Differences and Relations Among Teacher Confidence, Use, and Performance in Data-based Decision-Making

First Author: Eric Oslund -- Middle Tennessee State University
Additional authors/chairs: 
Amy Elleman
Keywords: Assessment, Teacher Knowledge, Teacher perceptions, Measurement, Response to Intervention
Abstract / Summary: 

Purpose: Teacher data-based decision-making is a critical component of RTI/MTSS models. This poster presents findings on the relations among teachers use, confidence, and performance on data-literacy tasks.

Method: Teacher data from 450 elementary teachers was analyzed using ANOVA and regression on measures of teacher confidence, use of progress monitoring (PM) data, and their performance on 14 graph literacy items. Items included graphically presented student data with questions about slope, intercept, response, and decision making.

Results: ANOVA’s indicated statistically significant differences on five intervally-scaled confidence ratings (e.g., confidence in data interpretation, etc.) among teachers with different amounts of PM data use (i.e., weekly, twice-monthly, monthly, or once a semester). Subsequent regression analysis found some of the five item-specific confidence ratings (e.g., confidence in graphically presented data interpretation) were positively and statistically significantly related to graph literacy. However, there were no statistically significant differences on graph literacy based on frequency of PM data use.

Conclusions: Prior research has found a link between teacher confidence and using student data to make educational decisions. Teachers who used student data more frequently were more confident in some RTI and graph literacy processes. Subsequent analysis indicate there was statistically significant but weak relations among confidence and data literacy. This indicates frequent use of PM data may increase confidence but that increased confidence may not meaningfully improve actual graph literacy, which is problematic when teachers are confident but make incorrect decisions. Our findings support possible Dunning-Kruger effects of teachers being unaware of their own (in)ability.