Raked / RIM weighting

In a survey we design a sample with the objective to represent the population and provide us insights about the population which would help researchers to draw inferences and conclusive findings for business or public agencies.

When we sample for a survey assign equal probability for individuals to be selected who would represent the population. However due to availability of individuals OR due to sensitivity of the research subjects OR accessibility of samples sometimes our samples could get skewed and would have the right representation of the population.

In RIM weighting we consider the an iterative proportional fitting procedure estimates the individual weights. The first iteration computes weights to match the first dimension (weighting variable) totals, the second iteration matches the second dimension totals, and so on. These steps for all the dimensions are performed repeatedly unless convergence is achieved within an acceptable margin of error.

Login to BI » Weightings » New weight scheme » Raked weighting » Select survey » Select the questions » Select data filter » Add a name » Add the propotions » Save Click to download video

Step 1: Weightings » New weight scheme

Step 2: Raked weighting » Next

Step 3: Select survey

Step 4: Select questions » Next

Step 5 (Optional): Create the required filter » Next

Step 5: Give a name » Add the propotions » Save

Processing can take anywhere from 1 minutes to 60 minutes. You can see the status here:
You can click on the weighting scheme name to see the stats for that particular scheme:
Important information:

• Effeciency: At what percentage of data efficiency did the weights converge..
• Iterations: The number of iterations required
• Status: Success (The data converged successfully) | Max stuck (More iterations could not improve the efficiency) | max iter (Reched the maximum number of iterations before data could converge)