د/ايمان زغلول قاسم

استاذ تكنولوجيا التعليم المساعد بكلية التربية بالزلفي

book B30

researchers do not succeed in finding convincing evidence regarding the effectiveness of
various feedback types. We doubt that there is one best way of providing feedback, given the
interaction between student characteristics, task characteristics and feedback characteristics.
This is in line with the findings of Hattie and Timperley (2007) and Shute (2008), who
conclude that the literature provides inconsistent results with respect to different methods for
providing feedback on students‘ learning outcomes. Also, in many studies that investigate the
effects of feedback on students‘ learning outcomes, the time students spend reading feedback
is not taken into consideration. The results of this study, however, suggest that time spent
reading feedback varies widely depending on the different ways of delivering feedback as
well as between students within one feedback condition. Therefore, it is recommended that
future research take into account time spent reading feedback. Unfortunately, the time
students spent reading the feedback was not available at the item level within this experiment.
If this data were to become available, it would be possible to investigate the relationship
between item difficulty, the ability of the student and time spent reading feedback. This type
of analysis could lead to new insights into the effects of different feedback types and feedback
timing on learning, especially between students with varying ability levels. These insights
could be a starting point for combining assessments for learning with computerised adaptive
testing.
A limitation of this study was that the assessment for learning and the summative
assessment were not constructed from a calibrated item pool. Unfortunately, after
administering the summative assessment, we had to reduce the test length from 30 to 11 items
due to the multidimensionality of the assessment and poor quality of some of the items. Also,
the summative assessment appeared to be more difficult than the assessment for learning. It
might, therefore, be possible that there was an effect as a result of the feedback condition, but
the summative assessment was not sensitive enough to measure this effect.
In future research, it is recommended that longer assessments of previously calibrated
items be used in order to develop assessments that are more reliable. Application of the
Spearman Brown prophesy formula predicts that Cronbach‘s alpha will be above .80 for a
comparable 30-item summative assessment. Besides, it is recommended to use parallel forms
of assessments to compare the results for both the assessment for learning and the summative
assessment. In this way, the effects of different feedback conditions can be measured with
more precision than was the case in this study.
Previous research has shown that the effects of various methods for providing
feedback differ concerning varying levels of learning outcomes. This study did not distinguish
between items that measured a specific level of learning outcomes because of the limited
amount of items used in the assessments. Making a distinction between different levels of
learning outcomes could lead to more insight into the conditions under which feedback is
effective.
In future research, it is recommended that larger groups be used in order to increase
the statistical power, and therefore the chance of finding significant effects of different
feedback conditions. Also, future research should point out if students benefit from computerbased
assessments for learning in the long run. Since in this study the summative assessment
was administered immediately after the assessment for learning, only short-term learning
effects could be measured in this experiment.

الوقت من ذهب

اذكر الله


المصحف الالكتروني