quantitative sotl research methods krista trinder, college of medicine brad wuetherick, gmcte...
TRANSCRIPT
Quantitative SOTL Research
MethodsKrista Trinder, College of Medicine
Brad Wuetherick, GMCTEOctober 28, 2010
Session outcomesO Identify components of good survey
questionsO Choose an appropriate research
methodO Willingly and enthusiastically
participate in more workshops in thisO Meet and work with other dedicated
and scholarly teachers
Quantitative ToolsO StatisticsO NumbersO Measurement
O Surveys/Questionnaires are the most common tool.
Survey SelectionO Good to use pre-existing instruments
that have been found to be reliable and valid.
O May have to create your own if no suitable tools exist.
Creating a SurveyO Decide what you want to measureO Generate a large number of itemsO Decide on question formatO Have experts review your itemsO Consider including validation itemsO Give items to a development sample
(some suggest 300)O Evaluate itemsO Adjust scale length
Examples of Survey Questions
O This course was enjoyable and provided me with information that will benefit my future career.O A. Strongly DisagreeO B. DisagreeO C. NeutralO D. AgreeO E. Strongly Agree
Examples of Survey Questions
O I dislike courses that do not include small group work.
O Better: I like courses that include small group work.
Good Survey QuestionsO Keep items short and concise.O Aim for a Grade 6 reading levelO Avoid double negativesO Avoid double-barreled items (only
ask one thing per question)O Avoid ambiguous pronoun references
and phrases with unintended meanings
Quantitative Approaches
O Experimental
O Comparative and Associational
O Descriptive
ExperimentalO Randomized Experimental Approach
O CausalityO Gold StandardO Randomly assign participants to
experimental and control groups
Quasi-ExperimentalO Similar to an experimental approach,
but without random assignment.
O Poor designs: O One group posttest onlyO One group pretest-posttestO Posttest only with non-equivalent
groupsO Better design:
O Pretest-posttest non-equivalent comparison group
Comparative and Associational ApproachesO Examine effect of an attribute
independent variable on the dependent variable.
O No random assignmentO No control over independent variables
(IVs)
O Comparative: ordered IVs with few levels (men, women).
O Associational: continuous IVs with more levels (age, grades)
Descriptive ApproachO Purpose is to summarize data
O One variable; no comparisons or relationships investigated
O Report percentages, means, frequencies etc.
Discussion of Approaches
O What approach might you use?
Parametric vs. Non-Parametric
O Usually use parametric analyses, which compare means (i.e., ANOVAs, t-tests).
O Sometimes it is more appropriate to conduct non-parametric analyses, which compare distributions (i.e., Chi-square). O Very small sampleO Different data collection methods
ReliabilityO Shows similar performance at
different times.O Sources of reliability:
O Test-retestO Internal Consistency (i.e., Cronbach’s
alpha – should be at least .70)O Rater Consistency (i.e., inter-rater
reliability)
Content ValidityO Extent to which the content of the
instrument measures what you want it to.
O Define your concept.O See how it is represented in the literatureO Develop test items assessing this
concept.O Expert judges review items.O See handout
Criterion ValidityO How well your instrument compares
to an outside criterion that is known to measure the same concept.
O E.g., How well does self-reported voting correlate with actual voting shown through records?
Construct ValidityO The extent to which a measure
correlates with other measures that should be related.
O Expect correlations with measures that measure the same thing (convergent validity).
O Expect low correlations with measures that should not be related (discriminant validity)
Reliability and ValidityO A measure can be reliable without
being valid, but cannot be valid without being reliable.
O See cartoon
Mixed MethodsO Do not have to use strictly
quantitative or qualitative tools!
O Research projects often benefit from including both quantitative and qualitative methods.
Choosing your methodO See handoutsO Work independentlyO Exchange thoughts and ideas: consultO More individual workO Leave at your own discretion
Thank you!See you on November 25th!