Challenges and opportunities of tablet-based electronic performance data collection and feedback system for artificial insemination delivery in dairy cattle: experience from the land O’Lakes Venture37 PAID project in Ethiopia

Abstract

Land O’Lakes venture37 (v37) has been using Open Data Kit (ODK) to collect Artificial Insemination (AI) technicians’ field performance within the Public Private Partnership for Artificial Insemination Delivery (PAID) Ethiopia Project from Amhara, Oromia, Southern Nations Nationalities Peoples (SNNP), and Tigray regions. This study assesses the challenges and opportunities of the Tablet-based project field performance data collection and feedback system. Information gathered from AI technicians using a structured questionnaire was analyzed using the crosstabulation procedures of SPSS and index methods. Almost all (96.7%) the interviewees had never used Tablet for data collection before PAID project. After training, about 82% of the respondents were able to use the ODK tool to collect and send data to PAID. However, there were reports on the challenges of using ODK that include poor or interruption of internet connectivity (index = 0.272), lengthy (time taking) nature of the forms (index = 0.241), and difficulty of updating the system (index = 0.093). Almost all the respondents were willing to use ODK in the future, provided internet is available and reliable (index = 0.248), and proper training is given (index = 0.238) for using ODK provides confidence that the report is delivered (index = 0.259), and it is timesaving (index = 0.24).

Keywords:

• Artificial Insemination
• Cattle
• Dairy
• Ethiopia
• ODK
• PAID

PUBLIC INTEREST STATEMENT

Field data can be collected using paper and/or electronically using tools like ODK. Artificial insemination (AI) technicians in Amhara, Oromia and SNNP regions who were trained to use ODK to collect data about Artificial Insemination Delivery in Dairy Cattle by Land O’Lakes Venture37 PAID project in Ethiopia were asked to provide their view on the challenges and opportunities of the Tablet-based data collection and feedback system. The result revealed that training about Tablet and ODK has helped AI technicians to use the technology to collect data and send it to PAID project. However, some of the respondents reported that poor or interruption of internet connectivity, the lengthy (time taking) nature of the forms, and difficulty in updating the system were the challenges of the system. Provided all the challenges are solved, AI technicians were willing to use ODK for future data collection.

1. Introduction

Most data collection in developing countries especially in the agricultural research and development sector is paper-based. However, increasing demands for quality statistical data and the need for their accessibility in real-time have caused the development of new modern technologies, especially in the data collection process. One of the solutions for accelerating the speed, accuracy, and accessibility of data is the use of modern IT instruments in data collection (Fitzgerald and FitzGibbon, 2014; King et al., 2014).

Open Data Kit (ODK) is an open-source software for collecting, managing and using data in resource-constrained environments. ODK is an extensible, open-source suite of tools designed to facilitate tasks at every level of data collection campaigns. It allows offline data collection using mobile devices in remote areas. The collected data can be submitted to a server when internet connectivity is available. It allows communities to aggregate data with full control over the collected data and the servers where the data are stored (JeffreyCoker et al., 2010; Terre des Hommes, CartONG and UNHCR, 2017); https://en.wikipedia.org/wiki/ODK_software.

ODK helps millions of people collect data quickly, accurately, offline, and at scale. It is the standard for offline data collection and is trusted by organizations such as the World Health Organization, Red Cross, Carter Center, CGIAR, Google, and many more (https://opendatakit.org/). The ILRI African Dairy Genetic Gains (ADGG) together with the National Animal Genetic Improvement Institute (NAGII) has established the ODK data collection system for dairy producers in Ethiopia to collect, aggregate and send electronic data directly to the main ADGG/NAGII data Platform on which data from different farm types (smallholders, medium-scale, and large-scale) is collected and stored (Oyieng et al., 2020).

The Public-Private Partnership for Artificial Insemination Delivery (PAID) Project, implemented by the Land O’Lakes Venture 37, is a project to address genetic constraints to dairy productivity growth in Ethiopia by strengthening Artificial Insemination (AI) delivery through public–private partnerships (PPPs). To bring about the intended dairy genetic gain in specific geographic coverage and time frame through an efficient AI service delivery system, a reliable supply of quality semen and liquid nitrogen (LN2) are prerequisites. With the intention of contributing to the country’s dairy genetic gain endeavors through efficient AI service delivery, Land O’Lakes Venture 37 (LOL V37) has been applying ODK tool to collect field performance data from the project-supported AI technicians in Amhara, Oromia, Southern Nations Nationalities Peoples (SNNP; currently, SNNP include Sidama and South-west Ethiopia regions), and Tigray regions. Efficient AI delivery and related performance data recording are critical for successful crossbreeding in dairy cattle and the sustainability of a profitable dairy business.

Understanding the effectiveness of the Tablet-based Data collection and feedback system and related challenges and opportunities as well as establishing a reliable data collection system that can be applied at all levels is required. The objective of this study, therefore, was to assess the effectiveness, challenges, and opportunities of Tablet-based Data Collection and Feedback System in the AI Delivery system in Amhara, SNNP, and Oromia regions of Ethiopia.

2. Materials and methods

A structured questionnaire was developed and employed to collect data on the use of the ODK tool to collect and report data on Artificial Insemination (AI) services provided by PAID project supported AI technicians in 2021. The data was collected from AI technicians trained on the use of ODK to collect AI service delivery-related field data in Amhara, SNNP, and Oromia regions and send to the PAID project. Of the total 621 project-supported AI technicians, 151 (50 from Amhara, 50 from SNNP, and 51 from Oromia) were randomly selected and interviewed through calls over the phone. The content of the interview included information about prior knowledge of the respondents on the use of IT tools; the current use of the ODK tool; the support provided to the AI technicians by regional coordinators and PAID staff; and the opportunities and challenges of ODK tool. Prior to administering the survey tool, interviewees were briefed about the objective of the assessment.

The data was encoded into SPSS software and analyzed using descriptive statistics (crosstabs) procedures (Jinn, 2011). The Crosstabs procedure includes the Mantel-Haenszel test of trend among its chi-square test statistics. This test was calculated as:

Chi-Square (MH) = (W-1)*r²

Where: W = the sum of the case weights and r² is the squared Pearson correlation between the row and column variables.

Those parameters with multiple responses and ratings were analyzed using index calculation using the formula provided by Kosgey (2004):

Index = Σ of [(5* the number of interviewees ranked first) + (4*the number of intervieweesranked second) + … + (1* the number of interviewees ranked 5^th)] given for particularvariables divided by Σ of [(5*sum of interviewees ranked first) + (4*sum of interviewees ranked second) + … + (1* sum of interviewees ranked 5^th)] for all the variables considered.

3. Results and discussion

3.1 General information of the respondents

While all the 151 sample respondent AI technicians were from the public sector, most (84.1%) of them were men (Table 1). There was not any association (p = 0.056) observed between the regions where the AI technicians are attached and their sex. The association analysis between the sex of the respondent AI technicians and data sending using ODK tool reveals a significant association (p = 0.000) where 87.4% of men respondents were able to use ODK, while only 54.2% of women AI technicians used ODK for field performance data collection and sending (Table 2). All the interviewees received either refreshments or full-fledged training in AI techniques by the PAID project staff.

Table 1. Descriptive statistics on the general characteristics of the respondents from Amhara, SNNP, and Oromia regions

Parameter	Amhara		SNNP		Oromia		Overall
	N	%	N	%	N	%	N	%
Sex of the respondent
Male	37	74.0	45	90.0	45	88.2	127	84.1
Female	13	26.0	5	10.0	6	11.8	24	15.9
Total	50	33.1	50	33.1	51	33.8	151	100
Pearson’s Chi-Square (P value)	6.713 (0.056)
Refreshment/full-fledged training on AI technique by PAID?
Yes	50	100	50	100	51	100	151	100
No	0	0.0	0	0.0	0	0.0	0	0.0
Total	50	33.1	50	33.1	51	33.8	151	100

Table 2. The association of sex of the respondent with the use of ODK tool to send data in Amhara, SNNP, and Oromia regions

Parameter	Sex of a respondent				Total
	Male		Female
	N	%	N	%	N	%
Have you ever collected and sent field performance data to PAID using ODK tool?
yes	111	87.40	13	54.20	124	82.10
No	16	12.60	11	45.80	27	17.90
Total	127	84.10	24	15.90	151	100.0
Pearson’s Chi-Square (P value)	15.18 (0.000)

3.2 Prior experiences of AI technicians on the use of Smart Phone and Tablet

As depicted in Table 3, most (61.6%) of the AI technicians had prior experience in using a smartphone. There was a significant (p = 0.00) association between region and prior experience in the use of a smartphone that 84.3% of the sample respondents from Oromia and 60% from SNNP regions reportedly had pre-PAID project experience in using a smartphone. About 60% of the respondent in the Amhara region, on the other hand, did not have prior experience. Overall, while 80.1% of the respondent AI technicians from all the regions had no experience of using a Tablet for any purpose, 96.7% of them reported to have never used Tablet for such field performance data collection and communication purposes prior to PAID Project.

Table 3. Prior experiences of respondent AI technicians from Amhara, SNNP, and Oromia regions in smartphone and Tablet use before PAID project

Parameters	Amhara		SNNP		Oromia		Overall
	N	%	N	%	N	%	N	%
Prior experience of using Smartphone
Yes	20	40.0	30	60.0	43	84.3	93	61.6
No	30	60.0	20	40.0	8	15.7	58	38.4
Total	50	33.1	50	33.1	51	33.8	151	100
Pearson’s Chi-Square (P value)	21.04(0.000)
Prior experience of using Tablet
Yes	20	40.0	3	6.0	7	13.7	30	19.9
No	30	60.0	47	94.0	44	86.3	121	80.1
Total	50	33.1	50	33.1	51	33.8	151	100
Pearson’s Chi-Square (P value)	19.98(0.000)
Tablet use for data collection purposes
Yes	0	0.0	3	6.0	2	3.9	5	3.3
No	50	100	47	94.0	49	96.1	146	96.7
Total	50	33.1	50	33.1	51	33.8	151	100
Pearson’s Chi-Square (P value)	2.90(0.234)

As revealed by the chi-square test, no apparent association was observed between prior experience in the use of smartphones (p = 0.113) and Tablet (p = 0.28) for field performance data collection and sending using the ODK tool (Annex Table 2 and Annex Table 3).

3.3 Information about the Tablet

All the respondents have received a Tablet from the PAID project of which 68.9% were without SIM card slot, with the difference (31.1%) having SIM card slot (Table 4). Overall, most (68.7%) of the Tablets provided to AI technicians are still functional. Among the Tablets that are not functional (31.3%), most of them are broken (53.2%), handed to others (19.1%), and have Software issues (14.9%). Only a few numbers of the respondents (7.3%) did not have the Table at hand.

Table 4. Information on the possession of Tablet from PAID project of respondent AI technicians in Amhara, SNNP, and Oromia regions

Parameter	Amhara		SNNP		Oromia		Total
	N	%	N	%	N	%	N	%
Have you received Tablet from the project?
Yes	50	100	50	100	51	100	151	100
No	0	0.0	0	0.0	0	0.0	0	0.0
Total	50	33.1	50	33.1	51	33.8	151	100
What type of Tablet did you receive?
Tablet with SIM card slot	20	40.0	18	36.0	9	17.6	47	31.1
Tablet without SIM card slot	30	60.0	32	64.0	42	82.4	104	68.9
Total	50	33.1	50	33.1	51	33.8	151	100
Is the Tablet still functional?
Yes	29	58.0	33	68.8	39	79.6	101	68.7
No	21	42.0	15	31.3	10	20.4	46	31.3
Total	50	34.0	48	32.7	49	33.3	147	100
If the Tablet is not functional, why is that?
Broken	8	38.1	12	80.0	5	45.5	25	53.2
Handed to others	9	42.9	0	0.0	0	0.0	9	19.1
Software problem	4	19.0	2	13.3	1	9.1	7	14.9
Battery problem	0	0.0	1	6.7	0	0.0	1	2.1
Other	0	0.0	0	0.0	5	45.5	5	10.6
Total	21	44.7	15	31.9	11	23.4	47	100
Do you still have the Tablet in your hand/at your disposal?
Yes	47	94.0	48	96.0	45	88.2	140	92.7
No	3	6.0	2	4.0	6	11.8	11	7.3
Total	50	33.1	50	33.1	51	33.8	151	100
If the Tablet is not at your hand, where is it?
Lost	1	33.3	2	100	4	66.6	7	63.6
Handed over to the district Bureau	0	0.0	0	0.0	2	33.3	2	18.2
Transferred to others	2	66.7	0	0.0	0	0.0	2	18.2
Total	3	27.3	2	18.2	6	54.5	11	100

3.4 Training on ODK by PAID

All the respondents in the three regions reported to have taken the ODK training by PAID project staff at least once (55.6%) and at most four times (7.9%) (Table 5). There was a significant association (p = 0.00) between the regions and training frequency where most (96.0%) AI technicians from the Amhara region reportedly took the training only once while a quite a few numbers of AI technicians in Oromia (41.2) and SNNPR (38.0) regions took the training twice. This might be attributed to the remoteness of the Amhara region from Addis Ababa, where the PAID Ethiopia project coordination team is based. Most of the respondent AI technicians rated the training given by the PAID project team as Satisfied (46.4%) and Agreed (66.9%) that training material and manuals prepared on the ODK system are helpful to understand the system and facilitate the process of data capturing and sending. The training material prepared was rated as important (54.7%).

Table 5. Training on Open Data Kit (ODK) by PAID project and level of satisfaction by respondent AI technicians in Amhara, SNNP, and Oromia regions

Parameters	Amhara		SNNP		Oromia		Overall
	N	%	N	%	N	%	N	%
Training on ODK as a data collection tool?
Yes	50	100	50	100	51	100	151	100
No	0	0	0	0	0	0	0	0
Total	50	33.1	50	33.1	51	33.8	151	100
Training frequency on ODK
Once	48	96.0	11	22.0	25	49.0	84	55.6
Twice	0	0.0	19	38.0	21	41.2	40	26.5
Three times	1	2.0	10	20.0	4	7.8	15	9.9
Four times and above	1	2.0	10	20.0	1	2.0	12	7.9
Total	50	33.1	50	33.1	51	33.8	151	100
Pearson’s Chi-Square (P value)	67.27(0.000)
Level of satisfaction on ODK training
Very satisfied	21	42.0	23	46.0	11	21.6	55	36.4
Satisfied	24	48.0	22	44.0	24	47.1	70	46.4
Neutral	1	2.0	5	10.0	9	17.6	15	9.9
Unsatisfied	4	8.0	0	0.0	7	13.7	11	7.3
Total	50	33.1	50	33.1	51	33.8	151	100
Pearson’s Chi-Square (P value)	17.68(0.007)
Agreement on the importance of the training material and manuals
Strongly agree	18	36.0	8	16.0	5	9.8	31	20.5
Agree	29	58.0	35	70.0	37	72.5	101	66.9
Neutral/no change	2	4.0	6	12.0	1	2.0	9	6.0
Disagree	1	2.0	1	2.0	8	15.7	10	6.6
Total	50	33.1	50	33.1	51	33.8	151	100
Pearson’s Chi-Square (P value)	24.40(0.000)
Level of importance of the training materials and manuals prepared on ODK system
Very important	16	32.7	19	38.0	4	7.8	39	26.0
Important	30	61.2	31	62.0	21	41.2	82	54.7
Neutral	1	2.0	0	0.0	26	51.0	27	18.0
Unimportant	2	4.1	0	0.0	0	0.0	2	1.3
Total	49	32.7	50	33.3	51	34.0	150	100
Pearson’s Chi-Square (P value)	67.52(0.000)

3.5 Use of ODK for data collection and sending

As summarized in Table 6, following a training, 82.1% of the respondents reported to have managed to send their respective field performance data using the ODK tool to the PAID project team at least once a month (54.8%), and four times a month (15.3%), once in a while (13.7), twice a month (10.5%), and every day (5.6%). The remaining about 18% of the respondents did not use ODK either to collect or send data to PAID due to internet connectivity issues (Index = 0.509) and they do not understand the system (Index = 0.205) (Table 6). This implies that data collectors need to get the required training until they internalize the system as a few numbers of the respondents (19%) rated the system as difficult. About 38% of all the respondents rated using ODK as Easy (38.4%) followed by Very easy (21.9%).

Table 6. Use of Open Data Kit (ODK) tool for data collection and sending by respondent AI technicians from Amhara, SNNP, and Oromia regions

Parameters	Amhara		SNNP		Oromia		Overall
	N	%	N	%	N	%	N	%
Used ODK tool to send data
Yes	39	78.0	43	86.0	42	82.4	124	82.1
No	11	22.0	7	14.0	9	17.6	27	17.9
Total	50	33.1	50	33.1	51	33.8	151	100
Pearson’s Chi-Square (P value)	1.09(0.579)
Frequency of data sending using ODK
Everyday	1	2.6	2	4.7	4	9.5	7	5.6
Every Week	7	17.9	7	16.3	5	11.9	19	15.3
Every two weeks	1	2.6	4	9.3	8	19.0	13	10.5
Every month	30	76.9	17	39.5	21	50.0	68	54.8
Once in a while	0	0.0	13	30.2	4	9.5	17	13.7
Total	39	31.5	43	34.7	42	33.9	124	100
Pearson’s Chi-Square (P value)	27.63(0.001)
Opinion about using ODK tool
Very easy	20	40.0	10	20.0	3	5.9	33	21.9
Easy	21	42.0	21	42.0	16	31.4	58	38.4
Neutral	2	4.0	12	24.0	17	33.3	31	20.5
Difficult	6	12.0	7	14.0	12	23.5	25	16.6
Very Difficult	1	2.0	0	0.0	3	5.9	4	2.6
Total	50	33.1	50	33.1	51	33.8	151	100
Pearson’s Chi-Square (P value)	31.32(0.000)
Current use of ODK tool for data sending
No	50	100	50	100	51	100	151	100
Yes	0	0.0	0	0.0	0	0.0	0	0.0
Total	50	33.1	50	33.1	51	33.8	151	100

In addition, AI technicians need to get the required support and follow-up in the ODK system from regional cluster coordinators and PAID project staff. There were quite a few AI technicians who reported to have not received the required support from the regional cluster coordinators (30%) and PAID project staff (18%) (Table 7 and 8). There was a significant association between the regions and the support from the regional coordinators (p = 0.01), and PAID project staff (p = 0.00). More than 55% of AI technicians from Oromia regions stated that they are either neutral or disagree with the support from the PAID project staff. However, though there were skill and knowledge differences among regional coordinators and budget and logistic problems, information gathered from both regional coordinators and PAID staff revealed that the required support was given to AI technicians.

Table 7. The priority (index) reasons of respondent AI technicians in Amhara, SNNP, and Oromia regions for not sending data using ODK at all (N = 27), and currently (N = 151)

Reasons for not using ODK	Amhara	SNNP	Oromia	Overall
Reasons for not using ODK for data sending at all (N=27)
Internet problem	0.329	0.522	0.778	0.509
Couldn’t understand the system	0.414	0.087	0.000	0.205
I am not interested	0.000	0.217	0.111	0.093
I have changed my location (district)	0.071	0.087	0.000	0.056
The tablet is Brocken	0.129	0.087	0.000	0.081
Incentive is not attractive	0.057	0.000	0.111	0.056
Reasons for not using ODK for data sending currently (N=151)
Lack of internet access	0.199	0.291	0.361	0.281
The project is not receiving ODK data	0.273	0.283	0.096	0.214
Difficulty updating the list in the system	0.128	0.091	0.318	0.184
Old version of ODK	0.129	0.122	0.130	0.128
The project is not paying incentive for ODK data	0.074	0.008	0.034	0.042
I am not interested to send data through ODK	0.041	0.028	0.023	0.031
My tablet is not functional	0.125	0.130	0.016	0.088
Due to unstable ODK forms	0.032	0.010	0.002	0.015
Tablet stolen	0.000	0.020	0.012	0.010
I stopped working for the project, change of place	0.000	0.016	0.000	0.005
Communication gap b/n the project and me	0.000	0.004	0.008	0.004

Table 8. Opinions of respondent AI technicians from Amhara, SNNP, and Oromia regions about the support on ODK from different actors

Parameters	Amhara		SNNP		Oromia		Overall
	N	%	N	%	N	%	N	%
Support and follow up from regional Cluster coordinators on the ODK system
Strongly agree	3	6.0	3	6.0	5	9.8	11	7.3
Agree	26	52.0	23	46.0	24	47.1	73	48.3
Neutral	0	0.0	13	26.0	10	19.6	23	15.2
Disagree	17	34.0	11	22.0	11	21.6	39	25.8
Strongly disagree	4	8.0	0	0.0	1	2.0	5	3.3
Total	50	33.1	50	33.1	51	33.8	151	100
Pearson’s Chi-Square (P value)	20.12(0.010)
Support and follow up from PAID project staff in the ODK system
Strongly agree	17	34.0	15	30.0	4	7.8	36	23.8
Agree	26	52.0	21	42.0	19	37.3	66	43.7
Neutral	1	2.0	12	24.0	11	21.6	24	15.9
Disagree	6	12.0	2	4.0	14	27.5	22	14.6
Strongly disagree	0	0.0	0	0.0	3	5.9	3	2.0
Total	50	33.1	50	33.1	51	33.8	151	100
Pearson’s Chi-Square (P value)	34.66(0.000)
Use of the regional Telegram platform to raise issues related to ODK system
Yes	44	88.0	28	56.0	19	37.3	91	60.3
No	6	12.0	22	44.0	32	62.7	60	39.7
Total	50	33.1	50	33.1	51	33.8	151	100
Pearson’s Chi-Square (P value)	27.72(0.000)
Expected amount of incentive payment to ODK data
Yes	44	88.0	30	61.2	27	64.3	101	67.8
No	6	12.0	19	38.8	15	35.7	40	26.8
Total	50	35.5	49	34.8	42	29.8	141	100
Pearson’s Chi-Square (P value)	27.46(0.000)

Despite all the possession and functionality of the tablets, training on ODK, and easiness of the system, none (100%) of the respondents are currently using the ODK tool for data sending to PAID project. Overall, the three priority reasons (sum index = 0.679) of the respondent AI technicians for not using the ODK tool for data collection and sending are lack of internet access (index = 0.281), the project is not receiving ODK data (index = 0.214), and difficulty of updating the list in the system (index = 0.184) (Table 7).

Information obtained from the project revealed that repeated efforts were undertaken, and a significant budget was allocated to use the ODK system as a sole source of performance data collection and sending. However, the volume of data received through the system was not justifiable to use as the sole source of the data collection system. On the other hand, the system was administered from Nairobi, Kenya and this has been a bottleneck to fix system-related challenges at the right time.

3.6 WiFi for data sharing

About 71% of the respondents received WiFi Router from PAID project of which 46.7% took themselves while the rest (53.3%) was taken by the respondents friends (Table 9). Most of the respondents (92.0%) who took the WiFi themselves reportedly have the router in their hands. However, only 40.0% of them have used WiFi to send data to PAID project four times (52.5%), three times (17.5%), once (15%), and twice (15%). There was a positive association (p = 0.00) between the regions and router possession that most from the Amhara region (58.0%) did not get a router from the project.

Table 9. Possession and use of WiFi Router by respondent AI technicians from Amhara, SNNP, and Oromia regions for data collection and sending

Parameters	Amhara		SNNP		Oromia		Overall
	N	%	N	%	N	%	N	%
Did you receive a WiFi Router from the project
Yes	21	42.0	43	86.0	43	84.3	107	70.9
No	29	58.0	7	14.0	8	15.7	44	29.1
Total	50	33.1	50	33.1	51	33.8	151	100
Pearson's Chi-Square (P value)	30.19(0.000)
Who received the WiFi router?
My self	18	85.7	9	20.9	23	53.5	50	46.7
My Colleagues	3	14.3	34	79.1	20	46.5	57	53.3
Total	21	19.6	43	40.2	43	40.2	107	100
Pearson's Chi-Square (P value)	25.11(0.000)
Do you still have the WiFi router in your hands/at your disposal?
Yes	17	94.4	8	88.9	21	91.3	46	92.0
No	1	5.6	1	11.1	2	8.7	4	8.0
Total	18	36.0	9	18.0	23	46.0	50	100
Pearson's Chi-Square (P value)	0.28(0.87)
If not, where is it?
Handed to district Bureau	1	100	0	0.0	2	100	3	75.0
Lost	0	0.0	1	100	0	0.0	1	25.0
Total	1	25.0	1	25.0	2	50.0	4	100
Pearson's Chi-Square (P value)	4.0(0.135)
Have you ever used WiFi Router for data sending
Yes	14	29.2	6	66.7	20	46.5	40	40.0
No	34	70.8	3	33.3	23	53.5	60	60.0
Total	48	48.0	9	9.0	43	43.0	100	100
Pearson's Chi-Square (P value)	15.76(0.003)
How many times did you send data using WiFi Router?
Once	3	21.4	0	0.0	3	15.0	6	15.0
Twice	1	7.1	0	0.0	5	25.0	6	15.0
Three times	1	7.1	1	16.7	5	25.0	7	17.5
Four and more times	9	64.3	5	83.3	7	35.0	21	52.5
Total	14	35.0	6	15.0	20	50.0	40	100
Pearson's Chi-Square (P value)	32.78(0.000)
Have you ever used a shared WiFi router for data sending?
Yes	0	0.0	2	6.1	7	35.0	9	11.1
No	28	100	31	93.9	13	65.0	72	88.9
Total	28	34.6	33	40.7	20	24.7	81	100
Pearson's Chi-Square (P value)	63.55(0.000)
How many times did send data using a shared router?
Once	0	0	0	0.0	3	42.9	3	33.3
Twice	0	0	2	100	1	14.3	3	33.3
Three times	0	0	0	0.0	2	28.6	2	22.2
Four and more times	0	0	0	0.0	1	14.3	1	11.1
Total	0	0.0	2	22.2	7	77.8	9	100
Pearson's Chi-Square (P value)	34.64(0.000)
How enough is the monthly WiFi top-up provided by the project to send data and update the list?
Enough	19	42.2	6	66.7	5	35.7	30	44.1
Not Enough	26	57.8	3	33.3	9	64.3	38	55.9
Total	45	66.2	9	13.2	14	20.6	68	100
Pearson's Chi-Square (P value)	63.63(0.000)

3.7 Challenges of using ODK tool for data collection and sending

The three most important challenges (sum index of 0.606) of using the ODK tool to collect and send data to PAID project as reported by the respondents include lack of internet service (index = 0.272), the forms in the system being too long to complete (index = 0.241), and the difficulty of updating an older version of ODK (index = 0.093), among the list of challenges reported (Table 10).

Table 10. Priority (index) challenges of using ODK tool to collect and send data by respondent AI technicians in Amhara, SNNP, and Oromia regions

Challenges to collect and send data using ODK tool	Amhara	SNNP	Oromia	Overall
Lack of Internet Service	0.227	0.288	0.303	0.272
Forms are too long/too many questions	0.269	0.216	0.239	0.241
Difficulty in updating older versions of ODK	0.147	0.071	0.062	0.093
Lack of ear tag for proper animal identification	0.080	0.117	0.043	0.081
Farmers have no mobile phone/Error log problem	0.083	0.087	0.069	0.080
Delayed incentive payment	0.023	0.079	0.074	0.059
Difficulty in updating the list of farmers and animals	0.052	0.021	0.081	0.051
Lack of support from the regional coordinators	0.053	0.027	0.027	0.036
Lack of skill and knowledge in tablet system	0.035	0.017	0.048	0.033
Language problem (problem understanding English)	0.007	0.007	0.053	0.022
Frequent changes in the system/forms	0.019	0.048	0.000	0.023
Poor awareness of farmers	0.000	0.012	0.000	0.004
Lack of support from the project staff	0.004	0.007	0.001	0.004
Incorrect/unexpected payment (incentive)	0.000	0.003	0.000	0.001

Despite the list of challenges reported, almost all the respondents are willing (96.7%) to use ODK as a data collection and sending tool if the challenges are solved, and choose ODK (88.1%) as a future method of data collection (Table 11). Therefore, working on these three challenges can improve the problem by more than 60%.

Table 11. Willingness and choice of data collection method by respondent AI technicians in Amhara, SNNP, and Oromia regions

Parameters	Amhara		SNNP		Oromia		Overall
	N	%	N	%	N	%	N	%
Willingness to use the tablet-based data capturing tool if problems/challenges are solved
Yes	50	100	49	98.0	47	92.2	146	96.7
No	0	0.0	1	2.0	4	7.8	5	3.3
Total	50	33.1	50	33.1	51	33.8	151	100
More convenient method of data collection
Tablet Based	47	94.0	42	84.0	44	86.3	133	88.1
Paper forms	3	6.0	8	16.0	7	13.7	18	11.9
Total	50	33.1	50	33.1	51	33.8	151	100
Pearson’s Chi-Square (P value)	2.62(0.27)

The reasons forwarded for the choice of ODK system by the respondents include confidence on the timely delivery of the report (index = 0.259), time efficient (index = 0.240), and ease and convenience of use (index = 0.133) (Table 12). Whereas only 3% of the respondents reported the paper-based data collection methods to be easier and more convenient (index = 0.310), time-saving (index = 0.279), and energy-saving (index = 0.159). There are earlier research results that reported IT-based data collection methods to be cost-effective, timely and accurate compared with that of paper-based one (Fitzgerald and FitzGibbon, 2014; Erin Satterlee et al., 2015).

Table 12. Priority (index) reasons of respondent AI technicians in Amhara, SNNP, and Oromia regions for the choice of data collection and sending methods

Reason for choosing data collection methods	Amhara	SNNP	Oromia	Overall
Tablet-based data collection
Have confidence that my report is delivered	0.279	0.245	0.252	0.259
Time saving	0.264	0.181	0.270	0.240
Easy and Convenient	0.121	0.199	0.082	0.133
Saves Money	0.107	0.083	0.094	0.095
Saves energy	0.061	0.081	0.133	0.092
Incentive payment is fast	0.003	0.110	0.121	0.076
It is new/modern technology	0.096	0.091	0.039	0.075
It is reliable	0.069	0.011	0.009	0.031
Paper-based data collection
Easy and Convenient	0.244	0.355	0.291	0.310
Time saving	0.289	0.264	0.291	0.279
Save energy	0.178	0.100	0.214	0.159
It is Reliable	0.222	0.136	0.097	0.136
Incentive is better	0.067	0.145	0.107	0.116

3.8 Solutions to re-start the Tablet-based data collection system

To restart the use of ODK as a data collection and sending tool, a list of activities is generated and prioritized by the respondent AI technicians (Table 13). Among the solutions forwarded, availing internet system/service to all AI technicians (index = 0.248), providing proper training to AI technicians about ODK (index = 0.238), and changing those tablets without SIM card slots to those with SIM card slots (index = 0.203) were the three very important ones accounting about 70% of all the activities forwarded. With respect to internet service, however, the project has provided WiFi router for about 70% of the respondents. Moreover, there was no significant association (p = 0.52) between receiving Tablet with SIM card slot and data sent to PAID project (Annex Table 1). It is pointed out that, however, improvements in telecommunications infrastructure may provide a more enabling environment for the use of ICTs in research in Ethiopia (Fitzgerald and FitzGibbon, 2014).

Table 13. Order (index) of potential solutions to re-start Tablet-based data collection system by respondent AI technicians in Amhara, SNNP, and Oromia regions

Activities to be performed	Amhara	SNNP	Oromia	Overall
Avail internet system to all AI technicians	0.236	0.225	0.283	0.248
Training of AI technicians	0.269	0.196	0.249	0.238
Change those tablets without SIM slots	0.281	0.184	0.145	0.203
Train regional coordinators to get the required support	0.029	0.111	0.032	0.058
Test the system for its effectiveness	0.059	0.092	0.009	0.053
Enough quantity of printed ear tags should be provided	0.005	0.105	0.044	0.052
Revise incentive rate for tablet based data	0.044	0.023	0.075	0.047
Reducing the prolonged and number of the questioner	0.000	0.028	0.092	0.040
Use local Language	0.022	0.016	0.057	0.032
Farmer and animal registration should be handled by other stakeholders	0.055	0.007	0.012	0.024
Facilitate and arrange motor cycle for the purpose	0.000	0.013	0.001	0.005

4. Conclusions and Recommendations

All the AI technicians interviewed were public; therefore, including those private AI technicians could have added value to the data collection process and adoption of the ODK technology in the future. The most important challenges of using ODK were internet problem, and understanding the system. Therefore, availing internet by changing those tablets without Sim card slots with those with SIM card slots and proper training, and supporting AI technicians are important to use ODK as future data collection means. All the respondents are willing to use ODK provided the challenges are solved to which most of them rated the use of ODK as easy. This is because ODK provides confidence for data delivery, saves time and it is easy and convenient. Finally, to collect accurate and transfer and share data, the ODK data collection system should be applied by providing proper training and availing internet access to data collectors.

Acknowledgments

The authors would like to express their great appreciation to the PAID project for the financial support.

Disclosure statement

The authors declare that they have no known conflicts of interest concerning this article’s research authorship and publication.

Additional information

Notes on contributors

Mengistie Taye

Zelalem Yilma Public Private Partnership for Artificial Insemination Delivery (PAID) was a seven years project implemented by the government of Ethiopia and Land O’Lakes Venture37. The objectives of the project were to improve the efficiency of AI technicians to increase demand among farmers and to enhance the input delivery system. The project was implemented in more than 200 districts in five regional states (Amhara, Oromia, SNNP, Sidama, and Tigray).