Presentation is loading. Please wait.

Presentation is loading. Please wait.

Data Cleaning and Imputation Imputation done on economic variables (assets, income, consumption, financial transfers, health expenses), education, self-reported.

Similar presentations


Presentation on theme: "Data Cleaning and Imputation Imputation done on economic variables (assets, income, consumption, financial transfers, health expenses), education, self-reported."— Presentation transcript:

1 Data Cleaning and Imputation Imputation done on economic variables (assets, income, consumption, financial transfers, health expenses), education, self-reported health, being depressed last month, general limitations of activity, ADL, IADL, numeracy, self-reported reading skills, making ends meet, number of children and grandchildren, urban or rural, maybe body mass index, grip strength Outliers: look at distribution (1 th, 5 th, 50 th, 95 th, 99 th percentile, relation to the median)‏

2  Periodicity (monthly vs annual): consumption, pensions, pensions, occupational pensions)‏  Check range of values for categorical variables  Currency mix-up, relevant for euro countries (easier to detect for countries with high value of exchange rate, local currency/euro)‏  Consistency of info between ownership and brackets  Change in values between W1 and W2: should not be great for home, IRA's, life insurance, mortgage, pensions, consumption, long term insurance payment

3 Use info from related variables -> judgment call  Last year's income vs last payment this year  Health expenses and illnesses  Look at sequentially asked variables (4 questions for numeracy)‏  Body mass index (height and weight)‏  House acquisition and having received a bequest Assume that everything can go wrong, but be conservative about changes Notify MEA and the imputation group about any deviations from the standard questionnaire Wave 1?

4 Do the changes at a country levels for cases that are reasonably clear. Put missing if you think the value is definitely wrong but can't get the “correct” one. Preserve original values before changing Create a flag variable with remarks about each variable Please contact me via e-mail for questions/problems (cdimitri@unisa.it)‏cdimitri@unisa.it Discuss issues with country team leaders


Download ppt "Data Cleaning and Imputation Imputation done on economic variables (assets, income, consumption, financial transfers, health expenses), education, self-reported."

Similar presentations


Ads by Google