Is that still the same? Has that changed? On the accuracy of measuring change with dependent interviewing