Eren Alvarez, Bryner Carbonell, Yesenia Garcia, Kim Mai, Pamala Somchith PARTICIPATION WORKSHEET

Eren Alvarez, Bryner Carbonell, Yesenia Garcia, Kim Mai, Pamala Somchith

PARTICIPATION WORKSHEET #4

SAMPLING AND EVALUATION DESIGN

Please identify which sampling method you will use: Cluster Sampling

Please justify your choice. Why is your choice the best for your proposal? How do you make sure that Experimental and Control/Comparison Group are comparable? What are the indicators or stratifying characteristics? Be very specific

Cluster Sampling is the best choice for our sample because in order for us to get our participants, it requires us to get people from doctors and hospitals who recommend them to our program. As we gain participants from them we would be able to ask which hospital or doctor they came from and categorize them in that way thus, those that are in the controlled group are with one doctor in a different community and the comparison is in a different location/provider. Thus, some characteristics that we will be looking for are low SES, location, who their doctor is, age, and race.

Please identify which evaluation research design you will use: Quasi-Experimental Design – Nonequivalent Comparison Group

Please provide for each of the 12 Threats to Internal Validity: (1) an example specific to your program, and (2) an approach on how you would mitigate the threat. Be very specific to your program

THREAT TO INTERNAL VALIDITY

EXAMPLE

MITIGATION

HISTORY

During implementation it was found that some participants would go on vacations which influenced them to stop on their diet and physical activity.

We would be able to mitigate it by letting them know in order for the program to work they must not leave and go out of town that might influence them to stop their diet and physical activity.

MATURATION

It was found that during the program some may have cravings and would have the strong sense of urgency to fulfill this craving and go off track in regards to their diet and physical activity

We were able to mitigate this by advising that every two weeks they are able to fulfill one of their cravings but the meal that they get would also be in moderation or would be aware of the portion size

THREAT TO INTERNAL VALIDITY

EXAMPLE

MITIGATION

TESTING

Participants may expect that they will be tested for their hgA1c and might try to prevent from having high numbers

Testing them to see if they know which type of food can cause Type 2 Diabetes, or how much they eat to-go, and reading food labels

Keeping the test for hgA1c in close proximity to not allow for any irregular test numbers

Keeping the test in close proximity

INSTRUMENTATION

Instrument may be out of date and may cause wrong readings

Updating the instruments and tests that we use, making sure that it is up to date and anything that is getting used is not old.

REGRESSION TO THE MEAN

Anthropometric measurement of body weight has extreme cases of participants weighing around 190lbs with the averaging weights of participants being around 150 lbs.

Keep the participants whose measurements are around 190 lbs in the program, but don’t include them in the post-test analysis.

SELECTION

We may recruit participants who have different doctors but go to the same hospital or know each other due to something that is out of our control.

To mitigate we would change areas and be sure that doctors and hospital are not in the same area and that they are different companies in terms of hospital

ATTRITION

Participants may lose interest in our program or may not have enough time to continue on with the program

We would be able to mitigate by providing incentives everytime they pass a marker

INTERACTION

History & selection. History: participants may have other health conditions that could impact their ability to fully participate in the program. Selection: Participants in the control and experimental group are selected in Long Beach causing participants from the control group to want to be part of the program or experimental group to share information to the control group which would affect results.

Cluster sampling identifies which participants meet the eligibility for the program through referral. Participants for the control group would be selected in Santa Ana and participants in the experimental group would be selected in Long Beach to ensure that there is no interaction between the two groups.

DIFFUSION

Participants from the experimental and control group may have a slight chance knowing each other and discuss further about the program

Make sure that the non-equivalent comparison group is located in a different county or state in order to keep them geographically separate; lowering the chances of participants knowing others from the other comparison group.

COMPENSATION

The administrators of the program may empathize with others who are at risk of type 2 diabetes, but do not fit the criteria and furthermore give them extra resources to help manage.

Thoroughly train administrators to only implement the resources of the program on the participants exclusively.

COMPENSATORY RIVALRY

The control group participants’ knowledge of risk factors for type 2 diabetes did not increase by 40% by the 3rd session, so the promotoras of the control group offer extra workshops to reach expectations.

Thoroughly train the promotoras on how to implement the planned workshops and document the activities conducted during each session to further monitor the program.

DEMORALIZATION

With the given curriculum of the program, the experiment group’s test results were not averaging the same as the control group, so the promotora feels as if there is no hope to increase knowledge and management skills for type 2 diabetes and quits.

Make sure the promotoras are trained well in the curriculum presented to the group and document what occurs in each session. If struggling to implement the curriculum to the participants, ensure that another administrator can elaborate more on the information being presented.