Skip to content

Commit

Permalink
text edits
Browse files Browse the repository at this point in the history
  • Loading branch information
@janelai22
janelai22 committed Jan 11, 2024
1 parent 8a0f042 commit 677d6ae
Show file tree
Hide file tree
Showing 7 changed files with 30 additions and 25 deletions.
2 changes: 1 addition & 1 deletion apps/templates/design/config_summary/summary.html
View file
Expand Up @@ -186,7 +186,7 @@ <h6> Input different possible tailoring variable values to see how the intervent
<div class="tab-pane fade show active" id="home" role="tabpanel"
aria-labelledby="home-tab">
<p class="mb-0">In “Configuration Summary,” we show you the summary of your
JITAI configuration.
pJITAI configuration.
Thompson Sampling adjusts the intervention probability based on the
values of the tailoring variables. Please input a value for each
tailoring variable.
Expand Down
7 changes: 3 additions & 4 deletions apps/templates/design/covariates/covariate_attributes.html
View file
Expand Up @@ -152,11 +152,10 @@ <h3>Model Settings / {{cov_name}} / Attributes</h3>
<div class="tab-content" id="myTabContent">
<div class="tab-pane fade show active" id="home" role="tabpanel"
aria-labelledby="home-tab">
<p>We use the attributes of the covariates to check if the API receives
<p>We use the minimum value/maximum value/type of the covariates to check if the API receives
inappropriate/missing input from your mobile app. Please note that if
your covariate is binary, we ask you to describe how the values of 0 and
1 translate to the states an individual is in (by filling out
the "meaning of the 'Covariate=0/1'"). See the HeartSteps Example tab for an example.
your covariate is binary, we ask you to describe what information is being captured by
the different values of the covariates. See the HeartSteps Example tab for an example.
</p>

<p>Please note that if your covariate needs to be transformed, the
Expand Down
10 changes: 5 additions & 5 deletions apps/templates/design/model/proximal_outcome_attribute.html
View file
Expand Up @@ -82,24 +82,24 @@ <h3>Model Settings / Proximal Outcome Attributes</h3>
<div class="tab-content" id="myTabContent">
<div class="tab-pane fade show active" id="home" role="tabpanel"
aria-labelledby="home-tab">
<p> We use the attributes of the proximal outcome to check if the API
<p> We use the minimum and maximum values of the proximal outcome to check if the API
receives inappropriate values from your mobile app. Please read
[tutorial] for more details.
</p>
<p>In this version of pJITAI, we assume the proximal outcome is continuous. In the future,
categorical proximal outcomes will be incorporated.
</p>
<p> If you're reading the HeartSteps demo, the HeartSteps numbers look
<p> Note 1. If you're reading the HeartSteps demo, the HeartSteps numbers look
weird because in HeartSteps the proximal outcome (the 30-minute step count
after each decision point) is log-transformed.
(See the HeartSteps Example tab for more explanation.)
</p>
<p> [You determine whether your proximal outcome needs to be transformed.
If you decide to transform your proximal outcome, please read the following.]
<p> Note 2. You determine whether your proximal outcome needs to be transformed.
If you decide to transform your proximal outcome, please read the following.
If your proximal outcome needs to be transformed, the
transformation needs to be implemented in the mobile application; the
algorithm will not do the transformation. Hence, if there is transformation, we are asking
for the minimum and maximum values of the proximal outcome after the transformation.
for the minimum and maximum values of the transformed proximal outcome.
(If your proximal outcome need not be transformed, we are asking for the minimum and maximum
values of the proximal outcome without any transformation.)
</p>
Expand Down
18 changes: 12 additions & 6 deletions apps/templates/design/projects/general_settings.html
View file
Expand Up @@ -97,11 +97,17 @@ <h3>General Settings / General Info</h3>
<div class="tab-pane fade show active" id="home" role="tabpanel"
aria-labelledby="home-tab">
<p> Intervention components are the elements of
your intervention. For example, to help increase a person’s physical
activity level, you may consider an activity suggestion message as an
intervention component or a prompt for the participant to set goals.
Please tell us the intervention component you are designing to personalize.
The proximal outcome is a short-term outcome that the intervention is intended to
your intervention.
You may have several intervention components in your study.
Please tell us the intervention component you are designing to personalize.
For example, to help increase a person’s physical
activity level, you may consider the delivery of an activity suggestion message as an
intervention component or a prompt for the participant to set goals. Suppose you want
to personalize the delivery of an activity suggestion message by personalizing whether
to send an activity suggestion message. Then, this activity-suggestion intervention component
has two intervention options: to send or not to send an activity suggestion.
</p>
<p> The proximal outcome is a short-term outcome that the intervention is intended to
impact.</p>
<p> Need examples of how to fill in these inputs? You can go to the HeartSteps demo
where we have filled out all the inputs for you. In addition, you can find the explanations
Expand All @@ -112,7 +118,7 @@ <h3>General Settings / General Info</h3>
aria-labelledby="profile-tab">
<p>In the HeartSteps example, the intervention components are: 1) activity
suggestions and 2) planning for the next day’s activity. Here, we are
demonstrating designing the “activity suggestion” JITAI component.</p>
demonstrating designing the “activity suggestion” component.</p>

<p>In HeartSteps, the step count in a 30-minute window is observed following
each decision point. In addition, step counts are right skewed. Thus, we
Expand Down
11 changes: 6 additions & 5 deletions apps/templates/design/projects/personalized_method.html
View file
Expand Up @@ -52,11 +52,7 @@ <h3>General Settings / Personalization Method</h3>
<div class="rightsidebluetextbox">
<p>This version of Thompson Sampling uses a Bayesian linear
regression to model the relationship between the proximal
outcome, states, and intervention options. This algorithm
uses the data from each individual to learn and update that
individual’s coefficients. (In other words, in this version,
to personalize the decision rule for a participant,
the algorithm does NOT use other participants' data.)
outcome, states, and intervention options.
This method determines which
intervention option to deliver based on the posterior
probability that the intervention option is most effective
Expand All @@ -68,6 +64,11 @@ <h3>General Settings / Personalization Method</h3>
<li>There are two intervention options.</li>
<li>The proximal outcome is continuous.</li>
</ul>
<p>In addition, in this version, the algorithm uses only data from
each individual to learn and update that
individual’s coefficients. (In other words, in this version,
to personalize the decision rule for a participant,
the algorithm does NOT use other participants' data.) </p>
<p>Coming soon: a version of Thompson Sampling for categorical proximal
outcomes.</p>

Expand Down
5 changes: 2 additions & 3 deletions apps/templates/design/projects/projects.html
View file
Expand Up @@ -23,7 +23,7 @@

<h3>Intervention Components <span data-toggle='tooltip'
data-bs-placement="right"
title='If you want to edit, duplicate, or delete an intervention component, you can click the three dots on the top right of the intervention component. You can find your finalized design by clicking "Finalized" in the left-hand menu.'> <i
title='If you want to edit, duplicate, or delete an intervention component, you can click the three dots in the top right corner of the intervention component. You can find your finalized design by clicking "Finalized" in the left-hand menu.'> <i
class=" align-text-top questionmark far fa-question-circle"></i></span>
</h3>
</div>
Expand Down Expand Up @@ -122,8 +122,7 @@ <h3 class="card-title"> {{proj.get("study_name")}}
</p>
<p> For example, in the HeartSteps study, the intervention components are:
1) activity suggestions and 2) planning for the next day’s activities.
In the HeartSteps demo, we demonstrate designing the activity suggestion
intervention component.
In the HeartSteps demo, we demonstrate designing the personalization of the activity suggestions.
</p>
<p> We suggest that if you are not familiar with our toolbox, you start by going through our
HeartSteps demo (which can be accessed by clicking "Add New Project" and then "Add
Expand Down
2 changes: 1 addition & 1 deletion apps/templates/includes/sidebar_project.html
View file
Expand Up @@ -231,7 +231,7 @@ <h4>Have more questions?</h4>

</ul>
<h4>Want to see these instruction cards again?</h4>
<p>Click "Instruction cards" on the left-hand menu.</p>
<p>Click "Instruction cards" in the left-hand menu.</p>
</div>
<div class="col-3">
<img src="{{ config.ASSETS_ROOT }}/images/pjitai/mobile_api.png"
Expand Down

0 comments on commit 677d6ae

Please sign in to comment.