outcomes evaluation a good evaluation is . useful to its audience practical to implement conducted...
DESCRIPTION
6 Steps of Evaluation Establish an evaluation plan from beginning of program Obtain buy-in from administrators Allow enough staff time to make evaluation a priorityTRANSCRIPT
Outcomes Evaluation
• A good evaluation is ….– Useful to its audience– practical to implement– conducted ethically– technically accurate.
Evaluation
• Outcomes/impact evaluation–measurement of changes in attitude,
knowledge, health status, behavior, nutrition status
6 Steps of Evaluation
• Establish an evaluation plan from beginning of program
• Obtain buy-in from administrators• Allow enough staff time to make
evaluation a priority
6 Steps of Evaluation
• Obtain permission & encourage participation from participants
• Be flexible & creative
6 Steps of Evaluation
• Use a strong research design & measures which generate data you need to support your program’s goal
Validity
• Internal validity– extent to which an observed effect can
be attributed to a planned intervention
Validity
• External validity– extent to which an observed impact can
be generalized to other settings & populations with similar characteristics
Threats to Validity
• History• Measurement• Selection
History
• External event (H)• Internal programmatic or internal
participant events (I)• Treatment effects (X)
Measurement
• Methods used to collect data• Instruments need to be– reliable – valid
Selection
• Define eligibility to participate in a program– criteria– is there a difference between those who
stay in the program & those that drop
Selection
• Of those eligible, who accept & who refuse– 3222/5000 = 64%– attended/eligible
Selection
• Drop out rate– 1600/3222 = 50%– drop out/initially say yes
Selection
• Lost to follow-up– 300/1622 = 18%– can’t find/completed program
Selection
• Identify contextual or structural variables to decrease selection bias
Regression Effects
• If score high on pretest, little room for improvement
• Will show program has poor impact• Is pretest score a threat to validity?
Synergistic Effects
• All work together to lower internal validity
Evaluation Designs
• Design to increase internal validity
Evaluation Designs
• Design is selected based on– objectives of the program– purpose of the evaluation– availability of eval resources– type of health & behavior problem,
setting & audience
Evaluation Designs
• Notation– R - random assignment– E - intervention group– C - true control group– C - comparison group– X - treatment
Evaluation Designs
• Notation– N - number of subjects– O - observation to collect data– T - time
One Group
• One group & one time– posttest only
• E OXO• Non-experimental• No random assignment
One Group
• No control/comparison group
One Group
• What are some of the main weaknesses of this design for increasing internal validity?
One Group
• When could this design be used & be appropriate?
Nonequivalent Comparison
• E OXO• C OXO• Comparison group– any group not formed by random
assignment
Nonequivalent Comparison
• What threats to internal validity are lessened?
Baseline DataVariable
Age (yr)Cog scoreBelief score
Exp45.771.81.5
Control44.170.91.5
SigNS*NSNS
Self-examNoneOneTwoThree
33%20%19%27%
23%23%14%40%
Sig0.01
*not significant
Time Series
• E OOO X OOO• Pattern of outcome variable• Know stability of outcome measure• Collect outcome variable
unobtrusively
Time Series
• Multiple data points– increases the power of the design
• Equal intervals
Time Series
• Must still try to control of history, selection & measurement
Time Series
• With control or comparison group, much stronger
• Better control over history threat to validity
Time Series with Comparison Group
• E OOO X OOO• C OOO X OOO
True Experimental
• R E OXO• R C O O• Establish at baseline 2 groups not
sig different• Best control of threats to validity
True Experimental
• Advantages?• Disadvantages?
Post-then-pre
• Appropriate for assessing behavior change
• Participants have limited knowledge at beginning of program
Post-then-pre
• Example of a typical pre-test question– Do you include one food rich in vitamin
C in your diet daily?
Post-then-pre
• Implement• See handout, Table 2• After program give only a posttest
Post-then-pre
• Q1 - Asks about behavior because of program
• Q2 - What the behavior had been before the program (i.e.. The pretest question)
Post-then-pre
• U of NE handout data analysis problem 6
• U of NE ETHT report
Success Stories
• Testimonials• Qualitative info• Audience testing
Case Study
• Story of an individual• Can be biased• Cannot be generalized
More Services Available
• Change in the environment
Evaluation Plans
• Read Moving to the Future example• PERT & Gantt charts
Evaluation Plans
• Must be in place– objectives– specifications of the intervention &
program methods
Evaluation Plans
• Must be in place–measurement & data collection
procedures– description of methods
Evaluation Plans
• Main reason– when– from whom– how– by whom
Evaluation Plans
• Worksheet 1 & 2 • Worthen & Sanders, 1987• Fill out form as a team
END Evaluation
• Questions?