sadc course in statistics the swaziland area survey choice ginindza
TRANSCRIPT
SADC Course in Statistics
The Swaziland area survey
Choice Ginindza
Overview
• Annual crop-cutting survey
• Concentrate on the measurement of areas– under each crop
• Forms for collection
• What happens to the data– after they have been collected
Each enumerator has a copy of this
guide
Holding layout sketch plan
Field 1 Maize
Field 2 60% maize,
40% pumpkins
Field 4 maize
Field 3 beans
Symbols for the sketch plan
Field area measurements
1 2 2 0 1 0 5 2
0 1
125 50 27.5
0.85
113 100
0.25
112 60 139 40
Area in each field
01 Maize 100% 113 0.85 02 M 60% Pk 40% 112 139 03 Beans 100% 131 0.53 04 Maize 100% 113 0.98
Total crop area
0.151.83
0.610.53
Other information is also recorded
Plus household information
What happens next?
• 75 staff collect these area data• between November and January each year
• The completed forms are sent to me in HQ– there are about 2000 forms each year– it is one form per household, unless a household
has more than 12 fields
• The forms are checked for consistency – before being sent for data entry
• They are then sent for data entry– By the end of January
Then – the data are entered
• Special software is used for the data entry– It is called CSPRO– and is designed specially for data entry
• A headquarters' staff member– has programmed CSPRO– to make the data entry screen – the same as the data entry form
• There are 7 staff – in the Government data entry team– and I add 2 of my staff just for this work
• It takes about 2 weeks
Checking the data once entered
• The resulting computer files – are returned to me
• I first look at them in an editor– it is called OX-edit and is just like Notepad– that is on every computer
• I use the editor to look for “shifts” – in the data from CSPRO– one character missing can make the transfer – to Excel or a statistics package very odd
• That checking just takes an hour perhaps
The data now
• Then I open the data in SPSS– That is a commonly used statistics package
• They look the same in Excel – – as shown on the next slide
• There are now 3945 records, – each one is a crop within a household.
• Plus the household information
Crop area information
Household information
Checking in SPSS
• I check the data in SPSS– That typically takes one day
• I also have the paper records available– And check odd values against them
• I correct in SPSS (and Ox-Edit)
• Sometimes records are returned – to the data entry people– and there is a further round
• Typically we find problems with 10 to 30 records
Analysis
• Now I analyse the data using SPSS
• That takes perhaps 1 day
• The most important is maize, – so that is done first.
• I get results from each Enumeration area
• And then total them for each district
• I compare with the previous year – as another check
• I then export the summaries from SPSS to Excel
Results
• Here is the sort of table I get.
Reporting
• This information is then sent – to our Meteorological Service
• They have to be there by the end of February
• The met service estimates yields
• And these estimates are multiplied – by these areas
• To give the estimated production
Publishing
• The estimates are then in a monthly– agroclimatic bulletin
• Which is part of our NEW system– NEW = National Early Warning
• to provide the food-balance sheet– for the current year
• Later we produce our own report– once we have the yields as well