130 likes | 260 Vues
This study explores the implications of survey timing on summative unit evaluations, particularly the risks and opportunities in administering surveys before and after exams and grades. By analyzing the responses from various demographic groups across different time intervals, insights into feedback, assessment effectiveness, and overall course quality are gained. The study highlights the challenges of representative sampling and the significance of gender balance in responses. Utilizing a comprehensive data analysis approach, this research seeks to enhance understanding of survey methodologies in academic settings.
E N D
John Ogier Making summative unit surveys totally summative - risks or opportunities
Collings AHEEF 2007 I misheard David’s method!! What are the implications of a long survey window? Into and after exams? Post grades? Response demographics?? Immediate response Post exams Post grades General - gender, grades etc Online Survey Experiment Wanted a wide range of large “units” Divide each randomly into 4 groups But high female response rate AUSSE 2007 - so maintain gender balance Got small selection of units Group 1 - normal pre-exams 15 days Group 2 - day after exam 15 days Group 3 - day after grades 13 days Group 4 - open entire 7 weeks Day 1 - personalised invitation email Week before close - reminder email Survey Timing
Variables C_FdBck “I received helpful feedback on my progress” C_Asmt “The assessments in this course measured my learning effectively” C_OAll “Overall, this was a good quality course” T_OAll “Overall, the lecturer is an effective teacher” Gender F, M Group 1, 2, 3, 4 (Survey Group) NumGrade -1 to 9 (“A+” = 9 through to “E” = -1) PostGrades PG (responded post grades/mid-year test return), NA (prior) Week W1 (responded Week 1), Rem (post-reminder email), NA (between) Also knew: - Ethnicity, International Student status, Citizenship, Age, intended Qualification, Overall GPA, Response Day & Time and Time taken
What about grades? Figure 4: ANOVA – C_FdBck, C_Asmt, C_OAll and T_OAll by Group
Dommeyer, Baum, Hanna & Chapman, 2004 - online surveys “do not produce significantly different mean evaluation scores” from paper UC Paper and Online survey history support this Not just means - also Std Deviations But are the responses “representative of the whole group”?? (Nulty, 2008) The dangers of “means”
Gender - a response bias? No significant diff @ 95% CI in Likert means or across Groups by Gender
Across the main units analysed: Bivariate fits Correlations for “Overall” Qtns Course R2 = 0.047 Teaching R2 = 0.0045 A+ is 9 ………… E is -1 What about grades?
Survey timing now apparent! Group 2 - post exams Group 3 - post grades Broken into Groups
Oct 6 - online surveys - 40 units Group 2 runs until 15th Dec. Grades released 5th Dec Where to from here?