5 tips: Making effective data collection forms

During the early study design phase, it is important to invest in electronic form design and the general architecture of your data. Here are some tips to consider while working through the form building process:

1. Better variable names

As a form is designed, the variable names that are assigned to each question will end up formatting your data dictionary. Variable names should be consistent, concise, and descriptive.

Consistent: Variable names should have a structure that is followed throughout a study. When a statistician is looking at your data dictionary they should be able to assign each variable name to it's respective form, a general question that was asked, and when it was asked. This dramatically improves the statistical analysis workflow.
Concise: The variable needs to be short so that when you extract your data to statistical analysis software, there aren't problems. For example, SAS has a 32 character limit on variable names and SPSS has a 64 character limit.
Descriptive: Even though, all statisticians will have access to a data dictionary, variable names should describe the variable in a meaningful way.

2. Keep forms short

Depending on your study design, bias can be introduced by creating form that look too long to your subject. There are many ways to work around this. Your forms can be segmented into smaller sections or you can use skip logic (more on this later) to make the forms expand or contract based on other questions in a form.

3. Automated survey invitations

For effective data collection you want to make it very easy for your study subject to get data into your database. There are many modes of data collection including: medical record abstractions, automated SMS based surveys, automated email survey, phone interviews, and direct data entry into a data capture system. If you are sending your data collection forms via email, using automated survey invitations could dramatically improve subject retention and improve response rates. This logic is typically designed in collaboration with BRIC staff and takes into consideration when a study subject is invited to complete a survey based on a specific study protocol.

4. Use skip logic

Skip logic is a way to direct a study subject through your data collection form based on their responses to other questions. For example, you can create skip logic in a form to hide or show sections of a form. Creating good skip logic will make forms concise (see tip #2) and create a much more user friendly experience for your data collection team or your study subject.

5. Use validation rules for responses

A few decades ago most of a study's data collection was happening via paper forms. All of those studies are now transitioning to web based data capture systems. When a paper based study is translated to a web based data capture system you have access to a lot more tools to improve data quality and data integrity. Included amongst these tools is the option to create validation rules for data capture. If you have a numerical response option, always use a minimum and a maximum numerical limit for response. This will prevent typos and result in cleaner data.