What are the essential steps of data preparation?
1.
Describe your data.
What are your sources? Link your sources to your research questions.
2.
Classify variables by analysis purpose.
Is your variable an outcome, a covariate, a confounder, or something else?
3.
Assess distributions.
What is the skew of the data? Will you need to make any transformations?
4.
Identify linking variables.
How will data be linked across forms? (Universal ID? Household codes?) How will indicators be linked to all processes? (Assessor ID? Cost tabulations?)
5.
Identify any composite indicators.
Socio-economic scales. Cognitive performance scales. This may require principal components analysis.
6.
What data needs to be transformed?
Do you need to convert to a z-score? Record to categorical outcomes?
What is a data analysis plan?
A data analysis plan is a systematic process describing how you will analyze your data based on some key determinants.
These determinants include:
Hypothesis, study design, variables collected, variables not collected, statistical distributions in your data, quality of your data, statistical expertise, meaningful conclusions for policy or action.
Why do you need a data analysis plan?
An analysis plan allows you to:
-
Confirm what you know about your data
-
Take an objective and systematic approach to reduce bias
-
Develop analytic, justified decisions that match your study design and data
Who should draft your data analysis plan?
-
Your principle investigator(s)
-
Your database administrator and statistician
-
Policymakers and decision makers
Need your videos translated? Hover here!
1. Enable Closed Captions on the video (CC)
2. Click on the Settings gear
3. Select Subtitles/CC
4. Click on Auto-translate and select your language!