In Luwero

Olivia and I left Kampala for the trip to the Save the Children sub-distict office in Luwero on Monday. It is about 60 kms north of Kampala, and I am excited to get out of the city and into the project areas of Save. They had some pretty bad rains here and the roads out of the city are clogged. Not far north of Kampala a large and impressive bridge is being built which is at right angles to the main north-south highway. I have no idea how long it has been constructed, but there was no noticeable evidence of activity going on. The bypass that went below and around the bridge, was filled with muddy water, some of which was pouring into the adjacent shops lining the bypass. A low car had attempted to cross through the flooding water and had stalled, effectively stopping the heavy traffic of vehicles, over-crowded wildly decorated buses and myriad trucks. Fortunately for us, our direction was fairly clear since most traffic mid-morning is back into Kampala.



After but a couple of hours we arrived at the municipality of Luwero where Save's office is. The area that we had just passed through was hilly and filled with trees and lush shrubbery. Banana trees abounded. I have found that the diet here is high in starches and rather uninteresting for my palate. I tried cassava for the first time, and potatoes, both sweet and Irish (our regular American Idaho variety), rice abounded. Chicken was omnipresent, but I found that is was rather tough and sinewy. We were greeted warmly as we entered the compound. I was met first by the administrator, David. There was a moment of confusion since we both thought we were repeating the others' name rather than providing our own (mine is David as well, in case I have not identified myself prior...).



Also, a quick note, I always have a predilection for both drivers and the lonely compound gate guards. At one point, after having always greeted the gate keepers upon entering each day, I went and sat in the tiny guard booth for a chat. These booths, as I have seen all over the world, are so small an afford essentially a place out of the rain. I noted after I left the booth after concluding a delightful comparison of cell phones, that I was cast a few glances. Perhaps, one does not fraternize in the guard booth. It had 2 windows and an open door, but I am somewhat egalitarian nature so I did not inquire. The picture of the guard follows this text.



We were graciously conducted to a large meeting room and provided with refreshments continually. Luc was busy in the Kampala office but would join us later today. The purpose of this training was for the Health Extension workers (HEW) who would actually use the PDA form for the Family Planning survey. this frankly the most exciting aspect of my visit, since these are the persons in direct contact with the beneficiaries. this is where the mobile processor connects with the people in need.



I was really surprised at how quickly and eagerly the HEW's took to the PDA's. I never noticed any sleepy eyes during the next three days and I was plied with questions. The HEW's explored the PDA's beyond my introductory training evidenced by needing to re-configure in preparation for my upcoming training in our Vietnam CO.

To go backwards in this time line, at the conclusion of the PDA training with the M&E, IT persons at the Kamapala office, Luc and I knew that we needed a form to be used by the HEW's in Luwero. The LQAS Family Planning (FP) form that Luc had derived from the Flexible Fund FP master form was one of the most articulate and well constructed form that I have seen introduced during on of my sessions. While it was strait forward, it needed a series of scripts behind the scenes of the PDA form for control of data. for example, a interviewer can literally record with pencil any numerical value for a persons age in years; in this case the LQAS FP survey is inclusive only of women from age 15 to 49. This might seem a low possibility happening, but actual field survey circumstances are less that optimal. Crying children distract a Moms attention, cooking chores, husbands, hot, humid, precipitation. So, inadvertent values are records for the simplest questions, and unless found by the interviewer, the person who then receives the form to hand-transcribe the survey data into a computer perhaps miles and weeks later has no idea of the correct value.



So, in our form, over the weekend, I placed an implicit value range allowed in the age field and made it it a required field, i.e."you must provide a value and it must be correct. This increases the data accuracy and frankly incorrect values are difficult to surmise post-interview and are either discarded or altered. The LQAS-type survey contains mostly dichotomous questions, Yes or No, etc. Also the LQAS statistical model provides a high degree of confidence with only a sample count of 19 (really). Therefore, if only one survey is inaccurate of the total of 19, the degree of confidence diminishes quickly.

I realized that Olivia, who I observed had both the mental regimen and eagerness to build these mobile databases, did not yet have the knowledge to build this type of form. However, I invited her to give it a try and then we will compare ofrms for our Monday training. I later told her that this put her in an almost impossible position, but I wanted her to give it a try independently of my instruction. I worked most of the weekend, only going out once to walk a mile or so, in order to the have a viable test form ready. In addition the LWAS FP form incorporates a skip questioning pattern. Th is simply a built in logic. If the respondent answers Yes, then the interviewer asks the succeeding question. However, if the respondent answers No, then the interviewr goes to question 42 in the 3rd section. Using a paper form in the field, this seems fairly straight forward once again, but there can be many distraction during the interview. also, paper forms, although much larger than a PDA screen, are often crowded with details and instructions and can be disorienting. On a PDA-designed form, a simple scrit automatically moves to the correct question based on the response. The scripts are simple in construct and actually make sense:

exitscreen:
if answer == Y then
goto [FP303A]
else (which is equivalent to a "No")
goto [FP301B]
endif

The [FP303] is actually a question code within the form, as well as [FP301B]. It's pretty simple. So, by monday we had a form to train with. Of course Olivia was stopped at something she had not encountered in the training, and quickly resolved it. So, with 7 PDA's passed out ot the HEW's we began our training!