Measuring critical thinking in pre-registration midwifery students: A multi-method approach

Test the concurrent validity of three newly developed tools (student self-rating, preceptor rating, and reflective writing) that aim to measure critical thinking in midwifery practice.

A descriptive matched cohort design was used.

Australian research intensive university offering a three year Bachelor of Midwifery programme.

Fifty-five undergraduate midwifery students.

Students assessed their ability to apply critical thinking in midwifery practice using a 25-item tool and a 5-item subscale in Motivated Strategies for Learning Questionnaire. Clinical preceptors completed a 24-item tool assessing the students' application of critical thinking in practice. Reflective writing by students was assessed by midwifery academics using a 15-item tool. Internal reliability, and concurrent validity were assessed. Correlations, t-tests, multiple regression and confidence levels were calculated for the three scales and associations with student characteristics.

The three scales achieved good internal reliability with a Cronbach's alpha coefficient between 0.93 and 0.97. Matched total scores for the three critical thinking scales were moderately correlated; student/preceptor (r=0.36, p<0.01); student/reflective writing (r=0.38, p<0.01); preceptor/reflective writing (r=0.30, p<0.05). All critical thinking mean scores were higher for students with a previous degree, but only significant for reflective writing (t (53)=-2.35, p=0.023). Preceptor ratings were predictive of GPA (beta=0.50, p<0.001, CI=0.10 to 0.30). Students' self-rating scores were predictive of year level (beta=0.32, p<0.05, CI=0.00 to 0.03).

The student, preceptor, and reflective writing tools were found to be reliable and valid measures of critical thinking. The three tools can be used individually or in combination to provide students with various sources of feedback to improve their practice. The tools allow formative measurement of critical thinking over time. Further testing of the tools with larger, diverse samples is recommended.