Report from first NUFU visit to University of Dar es Salaam, March 18-27 2003

The main objectives of my travel was to meet with the chief Tanzanian collaborators, make more detailed plans for the project and finalize the project budget, and in addition to have preparatory talks with Statistics Tanzania. All these objectives were met, in addition to being introduced to the local NUFU coordinater, ms Muamba. Especially the exchanges with the UiS project coordinater dr Rugaimukamu were fruitful, while more in depth discussions with other faculty members in the Dept of Statistics will have to be carried out during my next stay which is planned for the turn of September/October 2003.

Before the deadline on March 20, the project plan and budget was revised via the NUFU website. A detailed budget was worked out by dr Rugaimukamu and discussed with prof Katapa and myself. The main difficulty is to accommodate a national sampling strategy within a reduced project budget, but we believe this can be done successfully by interviewing less people in each household, and reducing the number of in-depth interviews somewhat. A lower US dollar exchange rate and falling prices on computer equipment was of course helpful. A local, but independent project account is being set up at the University of Dar es Salaam; a letter about this has been sent from the Department to the Dean. As soon as the account has been established, we shall rely on SEMUT to help us with the first transfer of funds.

A more detailed plan for activities during project year 1, 2003 was worked out:

The first version of the questionnaire had been worked out and discussed before my visit and further comments were discussed. A preliminary bibliography was created on the basis especially of the online library catalogue. A number of doctoral theses on regional infant mortality differentials written within the Demographic Training Unit were browsed and the most relevant parts copied.

At a productive meeting on Monday March 24 with the Director General Cletus P.B. Mkai of the National Bureau of Statistics Tanzania it was learned that anonymized microdata from 1988 is already available for researchers at the University of Dar es Salaam. The complete 20 % sample with the long form and 100 % with the short form can be analyzed. In a trial run we successfully analyzed the individual level SPSS data set from the long form. Our joint survey project is also to be based on a sample drawn from the 2002 census address database, which will be cleaned by July 2003 according to plans. Use of a substantial sample, maybe even 100% from the 2002 census, will further strengthen our project by giving more detail about household composition and a basis for evaluating the representativity of our survey sample.

1978 is on tapes of dubious quality, which have been sent to be read in the US. I offered to try in Norway if no luck there. Software from the Minnesota Population Center will be instrumental in creating constructed variables. Next time I am scheduled to visit the University of Dar es Salaam is in late September / early October 2003.

I thank the UiS coordinator for a fruitful stay!

Dar es Salaam / Tromsø, ultimo March 2003
Gunnar Thorvaldsen
UiN project coordinator