CMF Pilots PALM Pilots

Sunset at Sangam

Akhand Tiwari (blog co-author), Divya Varma, Anand Shukla and I had a fun time in Allahabad implementing a short survey on PALM Centros. Allahabad feels strange. A potpourri of “sahib hazuri” from the 1960s, urban-Indian culture of today and ancient rituals of puratan Bharat, Allahabad really confuses me – more so than other Indian cities. A motorcycle ride through the Civil Lines followed by a boat trip at the Sangam highlights the melting-pot nature of Allahabad.

Getting back to why we are writing this blog entry, traditional paper surveys are often a nightmare and have given social researchers and economists many sleepless nights. Some of the obvious and frequent problems encountered are – (i) logical inconsistencies in surveyor entries (ii) questions left blank at the surveyor’s whim and (iii) inaccurate work by data entry companies. Prof. Chris Blatmann (http://chrisblattman.blogspot.com/search?q=PenDragon ) does an excellent job at expressing general frustration with paper surveys. CMF has been contemplating a move to electronic surveys for a while. Contemplation turned to action when Doug Johnson and Dan Kopf noticed Prof. Blatmann’s Pendragon blog (http://chrisblattman.blogspot.com/search?q=PenDragon) and put us in touch with Bryan Plummer (http://pluminliberia.blogspot.com). Bryan works with Prof. Blatmann and is carrying out a large PDA based survey in Liberia. With some initial encouragement and excellent notes, Bryan helped CMF start its own PDA journey.

We implemented a short 3 page micro-finance take-up survey in the Allahabad area with 5 Palm Centros (http://www.palm.com/us/products/phones/centro/index.html) and Pendragon (http://www.pendragon-software.com/). Pendragon is fairly user friendly and the entire process of hardware purchase, software procurement, survey design and testing took less than two-and-a-half weeks to complete.

We had some initial apprehensions about training surveyors on PDAs. Local languages and scripts in India change roughly every 600-800kms and Pendrgaon surveys can be in the Roman script only. We were also worried if surveyors would be comfortable and careful handling PDAs. The fortunate thing is that most Indian-college students have some knowledge of English although the skill level is typically not very high.

Akhand, Divya and Anand trained the surveyors ensuring that they would be at ease going back and forth between interviewing in Hindi/Allahabadi and entering the data in English. Akhand describes his experience with the training and the actual survey here.

Anil enters data into a PALM

Akhand:

I wasn’t that excited when the news came in about using PDAs for the take-up survey here at Allahabad. Gradually my optimism increased and today I am very happy that we are using PDAs for the take up survey.

Training – Finding surveyors with the right attitude and aptitude is always difficult. This time it was more difficult. The surveyors were required to have knowledge of handling digital devices and be familiar with computer-type keyboard in addition to others. I remember asking some of the candidates to send their complete mailing address to me via text message so that I know they are familiar with mobiles. It’s funny, but I asked candidates to show their cell phone during the interview in order to see if they are using a high-tech cell. Thanks to the IT revolution in India, my apprehensions were relieved to a great extent when each surveyor was able to complete the practical exercise we asked them to do (Some tasks - switch power on/off, press home, from the screen that appears choose Form 5.1, select particular options … etc).

Using PDAs – Apprehensions with the usage of machinery are always rife. It was true for me too. I asked Ajay to stay in Allahabad longer than he wanted to so that during the actual survey if problems arise (PDA may hang, it may not take values, surveyors press different buttons etc), he can sort things out. Unluckily we could start only after Ajay was gone. Fortunately, the PDA (I guess Pendragon’s) interface is very interactive and simple. One can easily edit the form code too. There has been no case where PDA has hung or worked badly (right now we have completed around 1000 households).

Precautions – You need to be careful during typing, it is very much possible that you pressed 2 while you wanted to choose 1. This is something we can take care of during backchecks. Moreover, gradually surveyor’s typing skills increase after they have done 50 surveys on PDA, (so initially you need to be careful and not look at the numbers achieved at the end of day).

It feels great to be able to look at survey data every evening. The data is already clean since all skips and logical checks can be taken care of during the software design stage itself. The software can be made to force surveyors to enter only logically consistent values. Also, minor changes in the questionnaire can be implemented overnight. We are looking forward to the day when plain-vanilla paper surveys will be a thing of the past at CMF.