1: What are your hypotheses? Clearly define your causes and effects.

Also indicate if an effect

is the cause of another effect, too. Discuss why you think these hypotheses can be valid.

Hypothesis I: If the physicians are too busy throughout their workday, then they won’t be
able to input quality data into the EMRs.

Cause and effect, I: Cause: Physicians being too busy. Effect: Not being able to create high
quality information to enter into the EMRs.

This effect can cause another effect because this poor quality of information won’t help any
patients with potential risks to diabetes.

This hypothesis can be valid since the physicians can repeat this same study multiple times.
They then can observe if the information they entered is falsifiable.

Hypothesis II: If the patients with potential diabetes gave incorrect information, then the
quality of the information entered in the EMRs by the physicians will be poor.

Cause and effect: Patients gave incorrect information. Cause: Physicians entering poor quality
of information into EMRs.

This effect will also affect patients with potential diabetes to not be treated or incorrectly
treated. If a patient gave information to a physician in which would give them indication that
they have diabetes but in reality, they don’t then the that patient would be treated improperly.

This hypothesis can be valid since it can be testable. They can analyze If the data given by
patients were incorrect then they could determine that the data will have poor quality. This can
also be falsifiable since a different observation can be imagined by another researcher.

2: Discuss your variables in the scope of interest and variables outside of the scope of
interest. For each hypothesis, identify the following:

Independent variables:

Hypothesis I: Physicians helping with other patients not in the study. Physicians spending time
with other research. Physicians forgetting to enter data into EMR.

Hypothesis II: The type of incorrect of information given by the patient. Patients making up
information
Dependent variables:

Hypothesis I: Number of hours the physicians focus on entering data. Whether the physician is
burnt out from the workload.

Hypothesis II: How much of the incorrect information was entered into the EMRs.

Are there any constants? If yes, name them:

Hypothesis I: The same patient in the testing. The location of the 36 clinics. The same questions
being asked to collect the data.

Hypothesis II: The same patient in the testing. The location of the 36 clinics. The same
questions being asked to collect the data.

Are there any confounders, controls, or background variables? If yes, name them:

Hypothesis I: Confounders: The physicians do not have an effective way or improved process in
entering the data. Controls: Different physicians in the 36 clinics are having access to the EMR
instead of a few physicians. Background variables: The age, sex, ethnicity of the patients.

Hypothesis II: Confounders: The physicians do not have an effective way or improved process
in entering the data. Controls: Patient’s information Is taken the same way and collected in the
same room. Background variables: the age, sex, ethnicity of the patients.

3: How do you ensure the internal validity of the study? Address all biases to internal validity
and explain how these biases can be avoided.

You can ensure the internal validity by making sure you rule out any other explanations that can
affect your results from the study you are testing. Your treatment and response variable need
to change together. Your treatments precede any changes in the response variables and make
sure there are no confounding factors that can explain the results of the study.

Some threats that can occur in your study are: Participants, Instruments, Artificial Behavior, and
Settings.

To avoid the threats in participants, you can make sure you the selection of subjects are
random but make sure to have a large enough sample size to it doesn’t lead to the Simpson
Paradox. You also want to make sure the study doesn’t last several years since some of the
subjects could improve in health.

To avoid threats in instruments, you can make sure the tools used in the study and testing are
the same. That goes for the questions being asked to the patients and the data being entered
by the same physician. If you were to change an instrument, you must make sure you change it
as a whole since it could produce different outcomes.

To avoid threats in artificial behavior, you can make sure the patients aren’t aware of the study
so that they can’t lie about the information given to the physicians. Also, the physicians must
not look for answers to support their research question. For this, you could have one physician
collect the information from the patient without them knowing why they are collecting it.

To avoid threats to the settings, you can make sure no unfortunate event or mistaken events
occurs during the collection of data. You can also make sure that the patients sign an
agreement that they must not dropout of the study until the collection of data has ended.