Diffusion of Improved Lentil Varieties in Ethiopia

Diffusion of Improved Lentil Varieties in Ethiopia

A Comparison of Adoption Estimates from Expert Panel, Community Focus Group

Discussions and Sample Household Surveys

Chilot Yirga

Yigezu Atinaf Yigezu

Aden Aw-Hassan

Research Report 113

የኢትዮጵያ የግብርና ምርምር ኢንስቲትዩት Ethiopian Institute of Agricultural Research

Diffusion of Improved Lentil Varieties in Ethiopia

A Comparison of Adoption Estimates from Expert Panel, Community Focus Group

Discussions and Sample Household Surveys

©EIAR, 2016

http://www.eiar.gov.et

Tel +251-11-6462633

Fax +251-11-6461294

P.O. Box: 2003

Addis Ababa

ISSN: 9789994466290

Copy editor: Abebe Kirub

Contents

Summary 1

1. Introduction and Background 3

2. Trends in Lentil Production and Research 6

3. Methodology 9

4. Results and Discussion 14 4.2.1 Expert estimates of adoption and diffusion of lentil varieties 16

4.2.2 Community estimates of adoption and diffusion of lentil varieties 18

4.2.3 Adoption and diffusion of lentil varieties based on household survey 22

5. Conclusions and Lessons Learned 35

6. References 36

7. Appendices 37

1

Summary

The study presents adoption and diffusion of improved lentil varieties

drawing from several complementary data collection approaches including

desk reviews, expert panel interviews, community and household surveys

in major lentil producing areas of Amhara and Oromia regions of Ethiopia.

Descriptive analyses were largely used to summarize and present survey

results.

Evidence from the desk reviews revealed that the lentil improvement

program in Ethiopia has been working hand in hand with International

Center for Agricultural Research in Dry Areas (ICARDA) crop

improvement program that ensured a continuous flow of genetic resources

vital for the release of 12 improved lentil varieties. The number of

improved lentil varieties released during 2001-2010 period is higher

compared to the 1990s and 1980s further suggesting the synergy

achieved by the partnership. Despite the release of a fairly good number

of improved varieties, however, use of improved lentil varieties by

smallholder farmers is still below expectations. At a national level, the

expert panel estimates of the area weighted and simple (non-weighted)

estimates stands at 10.8% and 10.2% respectively, suggesting improved

varieties of lentil have not yet aggressively demonstrated in the major

lentil growing areas of the country. Estimates from the expert panel

further indicated that un-weighted estimates of the area share of

improved lentil varieties are higher in Oromia with 27.9% compared to

Amhara (17.8%). Similarly the community survey indicated that use of

improved lentil varieties in Ethiopia is low with considerable variability

across regions, zones and districts. Holder and area weighted adoption

estimates measured as proportion of households using improved lentil

varieties and areas share of improved varieties stands at 7.1% and

13.4%, respectively.

Adoption estimates based on the household survey generally follow a

similar pattern to the expert panel and community survey based

2

estimates. The household surveys indicated that the simple (un-weighted)

adoption rates measured as proportion of households using improved lentil

varieties and area share of improved lentil varieties is fairly low across the

study locations estimated at 12.5% and 15.2%, respectively. At a regional

level, the share of lentil area under improved varieties is much better in

Oromia with 38.6% compared to Amhara region estimated at 2.2%

reflecting the influence of on-farm demonstrations and pre-scaling up

activities conducted in the former. Further disaggregation of the data by

zone revealed that adoption rates are highest in East Shewa and West

Shewa Zones of Oromia Region and North Shewa Zone of Amhara all of

which are closer to Addis Ababa and have been the main targets of

outreach programs of the lentil improvement program. District level

adoption figures also reveal that adoption of improved lentil varieties varied

considerably within zones signifying adoption estimates at zonal level hide

interesting results. The weighted adoption estimates from the household

survey follow a similar trend to the simple adoption estimates. At a

national level, holder and area weighted adoption rates are estimated at

12.0% and 15.6%, respectively, with an absolute difference of 0.5 and 0.4

percentage points, suggesting at a national level both weighted and un-

weighted estimates are comparable. Similarly, at regional level the holder

and area weighted adoption estimates are comparable with non-weighted

adoption estimates with absolute difference of one percentage point or less

further providing evidence of the comparability of both estimates.

A comparison of the adoption estimates from the three data source reveal,

adoption rates are highest from the sample household survey. Irrespective

of the data source, weighted adoption rates are lower than the simple (un-

weighted) adoption rates indicating failure to use proper weights over

estimates adoption rates. The difference between weighted and un-

weighted adoption estimates, however, appears to be small from the

household survey than from community survey.

3

1. Introduction and Background

Food legumes constitute 11.5% of the grain crop area and

9.6% of the total production are the second most important

crops after cereals of immense economic importance both at

household and national level (CSA 2011). With little

intervention, the food legume market has developed to a 90

million USD industry (IFPRI, 2010). Food legumes offer

significant potential for Ethiopia to expand its foreign market

presence while increasing smallholder income. Expanding

food legume production provides multiple benefits. First, food

legumes are the major source of protein, especially for the

majority of the farming community who cannot afford to

purchase animal products. Second, food legumes improve soil

fertility through fixing atmospheric nitrogen thereby

contribute to cost savings. Third, in view of the growing

domestic and international demand, food legumes could be an

alternative cash sources for smallholder farmers. Realizing the

immense potential of food legumes to supply high quality

products for both the domestic and export market, increase

farm income and contribute to food security, a number of

initiatives have been underway aimed at increasing the

competitiveness of smallholder farmers. Prominent among the

initiatives include: the generation, adaptation and promotion

of improved food legume production technologies involving

high yielding varieties, recommended fertilizer rates and crop

protection practices; market liberalization, promotion of pulse

export trade and financing incentives aimed at enhancing the

4

competitiveness of pulse exporters (IFPRI, 2010). However,

despite efforts to increase on-farm productivity of food

legumes and improved incentives to exporters, continued

contribution of the sector to the economic development of the

country is threatened by low productivity, poor quality of

production and inconsistent exports.

Attempts to generate improved food legume technologies

adapt on-farmer’s field and demonstrate for producers to

encourage uptake of the technologies by producers dates back

to the early 1970s. The Ethiopian Institute of Agricultural

Research through its outreach program and the agricultural

extension wing of the Agriculture Bureaus of the Regional

States were involved in demonstrations and scaling up of

proven food legume technologies to farmers. Information on

the adoption and diffusion of improved food legume

technologies by smallholder farmers, however, are scarcely

available. To date, very few lentil technology adoption studies

were conducted in Ethiopia.

Owing to the dearth of information on the adoption and

diffusion of improved lentil varieties in Ethiopia and the need

for generating reliable and nationally representative adoption

estimates, the Ethiopian Institute of Agricultural Research

(EIAR) in collaboration with the International Research

Center for the Dry Areas (ICARDA) carried out an assessment

to track varietal change and assess the impact of crop genetic

improvement research in sub-Saharan Africa (SSA) under the

auspices Diffusion and Impact of Improved Varieties in Africa

5

(DIIVA) project. The DIIVA project aims at laying the

groundwork for tracking the successes and failures of crop

improvement efforts and for understanding the impact of those

investments on poverty, nutrition, and food security. In

Ethiopia, the project collected data on varietal releases,

strength of NARS and expert and community perception on

adoption rates of barley, faba bean, chickpea and lentil

varieties (Walker et.al, 2014). This report documents the

adoption and diffusion of improved varieties of lentil based on

broadly representative data collected through expert opinion

surveys, focus group discussions involving community

representatives and questionnaire based household surveys in

Ethiopia.

The report is organized into five sections. The next section

presents the trends of lentil production as well as the research

and extension efforts in Ethiopia. Section three discusses the

study locations, the sampling methods, data collected and

analysis. Results of the study are discussed in section four.

The last section, section five summarizes the main findings.

6

2. Trends in Lentil Production and

Research

In Ethiopia, lentil is one of the 12 important food legume

mainly cultivated by smallholder farmers in cooler Central

and Northern highlands. Lentil ranks 6th

in terms of number of

growers; accounting for, 6.2% of total pulse growers, area,

i.e., 5.7% of total area under pulse crops and production for

4.1% of total pulse production (Figure 1) (CSA, 1997/98 -

2010/11). Over the last 12 years lentil production and area

under lentil increased at the rate of 16.5% and 6.9%,

respectively during the last 12 years from 1997/98 to

2010/2011 (Figures 1 and 2). The annual growth rate,

however, is lower than the other highland food legume crops

in the country. Productivity also grew at an annual rate of

9.6%, higher than most of the pulses. `

7

Figure 1: Area under major highland pulses, Ethiopia, 1997/98 -2010/11

Source: CSA (different issue from 1998-2011)

Lentil production is concentrated in the Amhara and Oromia

regions with 59% and 31% producers, respectively, engaged

in production (Figure 5). The two regions account for 92% of

the total area occupied by lentil (CSA, 2011). Lentil is largely

grown on black vertisols which often suffer from inadequate

drainage during the main rainy season (June-August).

Consequently, lentil is planted late in the season (September-

October) on residual moisture. On well drained black soils,

lentil and chick pea could be grown as a second crop after

barley.

8

Figure 2: Production (0000' ton) of Major Highland Pulses, Ethiopia, 1997/98 - 2010/11

Source: CSA (different issue from 1998-2011)

Lentil provides important economic advantages to the small-

scale farm households in providing food, feed, cash income

and foreign currency earnings. Besides being rich in protein,

its ability to use atmospheric nitrogen through biological

nitrogen fixation (BNF) is economically appealing and

environmentally friendly. In spite of its importance, lentil

productivity has remained very low mainly due to the use of

low yielding local cultivars, biotic and abiotic constraints, and

poor management practices.

9

3. Methodology

Data for this study was collected in 2010 as part of the

Diffusion and Impact of Improved Varieties in Africa

(DIIVA). Several techniques including desk reviews,

elicitation of expert and community opinions as well as a

structured questionnaire were used to collect relevant data that

would allow address the objective of tracking the diffusion of

improved varieties in Ethiopia. The key data collected

includes crop varietal release, expert and community

perception of varietal knowledge and adoption, plot

characteristics; input use and production from sample

households (Walker et.al, 2014; Chilot et.al, 2015). The sub-

sections below describe the approaches employed for

collecting, summarizing and presenting the data.

3.1 Desk review

Annul national variety registry books published by the

Ministry of Agriculture of are the main source of the varietal

release data. Official published data were compared then

compared with personal interviews with senior chickpea

breeders and unpublished documents.

3.2. Expert opinion

Cool season grain legume breeds stationed at various research

centers were the core team members of the expert panel.

10

Elicitation of expert opinion on the adoption of improved

lentil varieties involved several steps. First, a consultation

workshop involving experts (mainly senior researchers and

research coordinators of chick pea, lentil, barley and faba

beans) was held in Addis Ababa, Ethiopia from 29-31 July,

2010 aimed at formulating expert elicitation procedures and

gain a general understanding of the diffusion pattern of

improved varieties of barley, faba bean, chickpea and lentils.

In the workshop, breeders, research managers and economists

from the Ethiopian Institute for Agricultural Research (EIAR)

headquarters and three of the major research stations

(Debreziet, Holetta and Kulumsa) assembled together with

economists from ICARDA. The panel of experts classified the

major growing agro-ecological zones of the four ICARDA

mandated crops into high, medium and low potential areas.

Second, the national research coordinators for the individual

crops, researchers involved in the respective commodities held

discussions and reviewed available evidence that would help

them estimate the adoption/diffusion levels for each crop

variety. Third, the panel of experts for each commodity

generated estimates of the area under local and improved

varieties of each crop by agro-ecological zone. Forth, these

figures were further disaggregated by variety. Finally, the

estimated adoption rates were checked by the respective crop

research coordinators and further disaggregated by

administrative zones for ease of understanding and practical

usefulness.

11

3.3. Community survey

The kebele also referred to as peasant association was the

primary sampling unit for the community survey. In each

study kebele one focus group discussion (FGD) was carried

out. The objective of the FGD is to obtain as much useful

information as possible in relatively short time period through

group interactions. The supervisor using the community

questionnaire facilitated the community level focus group

discussions. The major items included in the community level

focus group discussions include: geo-physical and agro-

climatic characteristics, and socio-economic features of the

communities; new varieties grown and their attributes, as well

as agricultural production; and prices of agricultural inputs

and outputs, and of household assets.

The information from the community survey provided useful

insight into the farming systems of the areas. The community

survey was conducted parallel to the household survey in all

the 125 sample kebeles in Amhara and Oromia Regions.

Extension personnel in respective sample kebels selected 6-8

community leaders who have lived for a long time in the

respective communities and believed to have extensive

knowledge of the communities. The supervisors facilitated the

community focus group discussions using a semi-structured

questionnaire

12

3.4 Household survey

The household survey was conducted following the

completion of the desk review and the expert opinion

elicitation. The household survey involved the collection of

cross-section data from representative sample households in

the main chick pea and lentil growing districts of the

highlands of Amhara and Oromia regions of Ethiopia during

2011. A multistage sampling procedure was used for

purposive selection of cereal-legume growing highlands

followed by a random selection of kebeles also known as

peasant associations (PAs) within districts and finally sample

households in the selected kebele. First, major cool season

grain legume growing districts were identified based on area

and production statistics of the Central Statistical Agency

(CSA, 2010). Second, given the study objectives and

employing power calculations, 33 districts from 9

administrative zones were selected (Table 1). Third, form each

district with the assistance of the respective district agriculture

officers, three kebeles were randomly selected from a list of

chickpea and lentil growing kebeles in the respective districts.

Finally, given the selected kebeles, 12 households were

randomly selected from a list of households solicited from the

district agricultural development offices.

13

Table 1: Distribution of Sample Households for the Lentil and Chickpea Adoption Study,

Ethiopia

Region Zone No. of

districts

No of

kebeles

Number of Households

Male Women total

Amhara North Wello 4 12 137 7 144

South Wello 7 21 238 12 250

North Shewa 7 21 229 23 252

South Gonder 3 9 101 7 108

North Gonder 3 9 105 3 108

Oromia North Shewa 3 9 104 4 108

East Shewa 2 6 68 4 72

West Shewa 2 6 68 4 72

Southwest Shewa 2 6 67 5 72

Total 33 99 1117 69 1186

Necessary data were collected from face to face interviews

using a structured household survey questionnaire. The

collected data from the household survey include improved

variety knowledge and adoption, plot characteristics (size,

distance from residence, severity of soil degradation, fertility

level, and slope); input use and production. Major socio-

economic variables measured include demographic structure

of households, farm size, livestock owned, and access to

credit, extension and improved inputs.

14

4. Results and Discussion

4.1 Varieties from Lentil Improvement Program

The development, release and promotion of improved lentil

varieties have been the main focus of the lentil improvement

program. Equally important of the lentil research program has

been the development of complementary production packages

that would enable producers to benefit from the cultivation of

improved lentil varieties. To date, the Debreziet Agricultural

Research Center of the Ethiopian Institute of Agricultural

Research (EIAR) coordinates and leads research in lentil

while the Regional Agricultural Research Institutes (RARIS)

of Amhara and Oromia, partake in the development and

release of improved lentil varieties and other complementary

innovations. The International Center of Agricultural Research

(ICARDA) provided either genetic materials used in the

breeding program or elite materials ready for release. Most of

the research work focused on identifying high yielding and

disease resistant varieties suitable for the diverse agro-

ecologies of the country. The strong partnership between the

lentil improvement program and ICARDA are instrumental in

release of 12 improved lentil varieties (Figure 3). Most of

these varieties have been promoted for use among smallholder

farmers in the major lentil production areas of the country. Of

the improved varieties released so far, four varieties has been

sourced from ICARDA. Ada’a was the first improved lentil

variety released in 1995 while Teshale and Alme Tena were

released in 2004. As depicted in Figure 3, there has generally

15

been an increasing trend in the total number of varieties

released. In particular, the number of chickpea varieties

released is remarkably high during 2001-2010 periods. List of

lentil varieties released for which records are available with

cultivar name, genetic background, year of release, origin and

selected characteristics of the cultivars, such as date of

maturity, plant height, grain yield and other environmental

requirements are provided in Appendix 1. Improved lentil

varieties, on the average, provide 2.5 ton/ha grain on farmers'

field.

Figure 3: Released improved lentil varieties by decade, Ethiopia

16

4.2 Adoption of Improved Lentil Varieties

4.2.1 Expert estimates of adoption and diffusion of

lentil varieties

Simple (un-weighted) expert panel estimates of the degree of

use of improved lentil varieties by smallholder farmers

measured as share of lentil area under improved varieties is

provided in Table 2. The expert estimates indicates that

adoption rates are higher in Oromia with 14.8% compared to

Amhara (9.4%) with considerable variations within zones.

Nationally, the area weighted expert estimates of the area

under improved lentil varieties is moderate, with about 11%

revealing experts degree of belief that improved lentil

varieties are not yet widely disseminated in the major lentil

growing areas of the country. And yet, according to the panel

of experts, use of improved lentil varieties by smallholder

farmers are relatively high in the central highlands (East

Shewa, West Shewa, North West Shewa and North Shewa of

the Amhara Region) where much of the lentil technology

transfer works have been intensively conducted.

17

Table 2: Un-weighted expert estimates of the adoption level of improved lentil varieties, Ethiopia

Region

Zone

Lentil area Area

under

improved

varieties

(ha)

Un-weighted

expert

adoption

estimates

(% area)

ha % of total

in the

region

% of total

in the

country

Amhara

North Shewa 17528 40.9 22.9 3506 20.0

East Gojam 1102 2.6 1.4 44 4.0

South Wello 11896 27.8 15.5 357 3.0

North Gonder 4430 10.3 5.8 44 1.0

South Gonder 2606 6.1 3.4 26 1.0

North Wello 5300 12.4 6.9 53 1.0

Total Amhara 42862 94.4 55.9 4030 9.4

Oromia

East Shewa 9983 39.9 13.0 2995 30.0

South West Shewa 3501 14.0 4.6 280 8.0

West Shewa 2429 9.7 3.2 170 7.0

North Shewa 7840 31.3 10.2 235 3.0

Arsi 1289 5.1 1.7 26 2.0

Total Oromia 25043 94.3 32.7 3706 14.8

National 71668 93.5 7829 10.9

Table 3 depicts varietal level adoption estimates of the

diffusion of improved lentil varieties by the panel of

experts. According to the expert pane, only 4 out of the

12 released improved lentil varieties are currently under

cultivation by smallholder farmers in Ethiopia. The

expert panel also indicted that about 73% of the

improved lentil area is covered by a single improved

variety, Alemaya, highlighting the risks posed in

relaying on a single variety in lentil production.

18

Table 3: Expert estimates of adoption rates of

improved varieties of lentils in Ethiopia, 2010

Variety Area

(ha)

% area

Alamaya 5765 8.0

Teshale 777 1.1

Alem Tena 109 0.2

Ada’a 903 1.3

Other Improved 275 0.4

All Improved 7829 10.9

4.2.2 Community estimates of adoption and

diffusion of lentil varieties

As noted in the methodology, community level adoption

estimates were solicited from 75 communities in 25

districts. Results of the community level adoption measured

both in terms of percent lentil area under improved varieties

and proportion of households cultivating improved lentil

varieties are summarized in Table 4. Out of the 25 districts

for which community level adoption estimates were

collected, non-zero adoption rates are reported in six

districts namely, Minajar Shenkora, Fogera, Wegera, Aleltu,

Ada’a and Gimbichu. The data further indicated that,

adoption of improved lentil varieties though limited to few

districts, both the rate and intensity of use of improved lentil

varieties is relatively higher in Oromia than in Amhara

Region. None of the communities in North Wello and South

Wello Zones of the Amhara Region reported any experience

in growing improved lentil varieties. The community survey

further elucidated that the use of improved lentil varieties is

concentrated in three districts namely Ada’a, Gimbichu and

Minajar Shenkora all of which are close to the Debreziet

19

Agricultural Research Center suggesting the center’s

outreach programs have been the main instruments for

triggering adoption of improved lentil varieties in these

districts.

Table 4: District level community estimates of non-weighted adoption rates of improved lentil

varieties, Ethiopia, 2011

Region Zone District Adoption Rate (%)

area HH

Amhara North Wollo Gubalafto 0.0 0.0

Habru 0.0 0.0

Meket 0.0 0.0

Dawint 0.0 0.0

South Wollo Kelala 0.0 0.0

Wegidi 0.0 0.0

Legehida 0.0 0.0

Tenta 0.0 0.0

Tehuledere 0.0 0.0

Werebabu 0.0 0.0

Desezuria 0.0 0.0

North Shewa Moretina Jiru 0.0 0.0

Siadebrina Wayu 0.0 0.0

Ensaro 0.0 0.0

Mojona Wedera 0.0 0.0

Menze Keye 0.0 0.0

Hageremariam 0.0 0.0

Minjar Shenkora 20.0 15.1

South Gonder Fogera 8.0 1.0

North Gonder Wegera 2.9 0.6

Oromiya North Shewa Aleltu 1.7 6.7

East Shewa Ada’a 100.0 54.7

Ginbichu 53.0 60.0

West Shewa Ejere 0.0 0.0

South West Shewa Elu 0.0 0.0

Table 5 presents the simple (un-weighted) community

estimates of adoption rates aggregated at zonal, regional

and national levels. Zonal level community estimated

adoption rates follow a similar pattern as that of expert

20

panel estimates. Adoption is relatively higher in the

central highlands (East Shewa, West Shewa, North

West Shewa and North Shewa of the Amhara Region)

than the rest of the country. The large difference in the

use of improved lentil varieties reported by

communities between Oromia and Amhara region is

also evident in the community survey. Nationally, the

community based estimates indicated about 5.5% of the

households adopted improved varieties on about 7.4%

of the lentil area. Low awareness about improved lentil

varieties and lack of access to seed were the two main

reasons emphasized by communities of the low/non-

adoption of improved varieties.

Table 5: Community estimates of non-weighted adoption rates of

improved varieties of lentils by zone and region,

Ethiopia

Region Zone HH (%) Area (%)

Amhara 0.8 1.5

North Gonder 0.6 2.9

North Shewa 2.2 2.9

North Wello 0.0 0.0

South Gonder 1.0 8.0

South Wollo 0.0 0.0

Oromia

24.3 31.0

East Shewa 57.3 76.5

North Shewa 6.7 1.7

South west Shewa 0.0 0.0

West Shewa 0.0 0.0

National All Zones 5.5 7.4

Varietal level adoption rates from the community

survey are presented in Table 6. In tandem with the

21

expert panel estimates, the community survey reveal

that only 5 of the 12 released improved lentil varieties

are currently cultivated by smallholder farmers in

Ethiopia. Even then, production is dominated by one

improved variety, Alemaya, apparently grown by about

20% of households occupying about a quarter of the

area under improved varieties (Table 6).

Table 6: Community estimates of adoption rates (un-weighted) of lentils disaggregated by variety

in Ethiopia, 2010

Variety Communities

(No.)

HH (%) Area

(%)

Alemaya 26 20.4 25.1

Ada’a 2 0.0 0.0

Gudo 1 0.0 0.0

Improved, but not matched with official name 2 2.0 1

Table 7 compares un-weighted (simple) adoption

estimates of the proportion of households and area share

of improved varieties with holder and area weighted

adoption estimates at regional and national levels based

on the community survey. On the average, weighted

adoption estimates exhibit a similar pattern to un-

weighted adoption estimates. At national level, holder

and area weighted adoption rates are estimated to bet

7.1% and 13.4%, respectively, with an absolute

difference of 1.6 and 6.0 percentage points, providing

evidence that the two estimates differ markedly.

Similarly, at regional level the holder and area weighted

adoption estimates are comparable the Amhara Region

22

but differ by a significant margin for Oromia. Weighted

adoption estimates for lentil at a zone level, however,

were not estimated due to data limitations.

Table 7: Weighted and Un-weighted adoption estimates of the level of use of improved

lentil varieties based on the community survey, Ethiopia, 2011

Region Un-weighted

adoption

Weighted

adoption

Absolute

difference

% HH % Area % HH %

Area

%

HH

%

area

Amhara 0.8 1.5 0.8 2.0 0.0 -0.5

Oromia 24.8 31.0 19.1 32.9 5.7 -2.0

National 5.5 7.4 7.1 13.4 -1.6 -6.0

4.2.3 Adoption and diffusion of lentil varieties

based on household survey

An adopter in the household survey is defined as one who

used improved lentil varieties on at least one of his/her plots

during the study year. Un-weighted (simple) adoption rates

measured as proportion of households growing improved

lentil varieties and the share of lentil area under improved

varieties are presented in Table 8. The data indicate that

adoption of improved lentil varieties varied considerably

within a zone and across zones. Of the 32 districts included

in the study, improved lentil varieties are cultivated in 7

districts (22%). Meaningful level of use of improved lentil

varieties, however, are limited to Minajar Shenkhora,

Moretina Jiru and Ensaro Districts in North Shewa Zone of

Amhara region, Ada’a and Ginbichu districts of East Shewa

Zone of the Oromia Region with over 10% of the lentil area

under improved varieties. All the districts with significant

adoption rates are traditional area where the EIAR have

23

been conducting lentil technology transfer activities. Further

scrutiny of the district level adoption figures indicate that

two districts, Ada’a and Gimbichu in the East Shewa, both

in the Oromia Region, has the highest adoption rates with

about 90% of the lentil area under improved varieties.

Similarly in the Amhara Region, only Minajar Shenkhora

registered above the national average of 15.2% further

suggesting research center based extension efforts have been

the main triggers of improved varietal adoption and

diffusion in Ethiopia.

24

Table 8: Simple (non-weighted) adoption rates of lentils varieties from the household survey

disaggregated by district, Ethiopia, 2011

Region Zone District Sample size (n)

Adoption rate

% of HH % of area

Amhara

North Wollo

Guba Lafto 7 0.0 0.0

Habru 2 0.0 0.0

Meket 27 0.0 0.0

Dawint 35 0.0 0.0

South Wollo

Kelala 17 0.0 0.0

Wegidi 1 0.0 0.0

Legehida 35 0.0 0.0

Tenta 22 0.0 0.0

Tehuledere 11 0.0 0.0

Werebabu 4 0.0 0.0

Dese Zuria 13 0.0 0.0

North Shewa

Moretina Jiru 35 14.3 4.8

Siadebrinawayu 25 4.0 1.1

Ensaro 21 14.3 12.1

Mojona Wedera 30 0.0 0.0

Menze Keye 24 0.0 0.0

Hageremariam 31 0.0 0.0

Minjar Shenkora 15 40.0 38.5

South Gonder

Farta 9 0.0 0.0

Fogera 9 0.0 0.0

Misrak Este 4 0.0 0.0

North Gonder Gonder Zuria 10 0.0 0.0

Wegera 23 0.0 0.0

Oromiya

North Shewa

Debrelibanos 15 0.0 0.0

Hidabu Abote 10 0.0 0.0

Aleltu 31 9.7 4.3

East Shewa Ada'a 27 88.9 90.3

Gimbichu 35 100.0 87.7

West Shewa Dendi 17 0.0 0.0

Ejere 24 0.0 0.0

South West Shewa

Tole 21 0.0 0.0

Elu 28 0.0 0.0

Adoption estimates at a zone level are provided on Table

9. Of the 9 major lentil growing zones included in the

25

household survey cultivation of improved lentil varieties

is limited to three zones, namely, East Shewa and West

Shewa Zones of Oromia Region and North Shewa Zone

of Amhara Region all of which are closer to Addis Ababa

and have been the main targets of the outreach programs

of the lentil and chick pea improvement programs.

Among the three zones, adoption rates are highest in East

Shewa where the lentil improvement program is based.

Adoption of improved lentil varieties in the other major

lentil growing zones such West Shewa, North and South

Wello, North and South Gonder, however are

practically nil suggesting research center based

extension efforts have been the main drivers of adoption

of improved lentil varieties (Tables 9). At a regional

level, the proportion of households using improved lentil

varieties and area under improved lentil varieties is much

better in Oromia with 29.8% and 34.4%, respectively,

compared to Amhara region estimated at 3.7% and 2.5%,

respectively, reflecting the influence of on-farm

demonstrations and pre-scaling up activities conducted

by the lentil research coordinating center in its proximate

areas. Further un-weighted aggregation of the household

survey data revealed that the proportion of households

using improved lentil varieties and area under improved

lentil varieties at a national level is fairly low estimated at

12.5% and 15.2%, respectively.

26

Table 9: Un-weighted adoption rates of lentils varieties from household survey by

administrative zone and region, Ethiopia, 2011

Region Zone Sample size (n)

Adoption rate

% of HH

% Area

Amhara

All Sample Zones 410 3.7 2.5

North Gonder 33 0.0 0.0

North Shewa 181 8.3 5.2

North Wollo 71 0.0 0.0

South Gonder 22 0.0 0.0

South Wollo 103 0.0 0.0

Oromiya

All Sample Zones 208 29.8 34.4

East Shewa 62 95.2 88.6

North Shewa 56 5.4 3.5

South West Shewa 49 0.0 0.0

West Shewa 41 0.0 0.0

Whole Sample 618 12.5 15.2

Varietal level adoption rates are given in Table 10. The

household survey data indicated that only two of the 12

released improved varieties are currently grown during

the study year indicating the vulnerability of lentil

production posed by farmers' dependence on limited

lentil varieties. The data also indicated that almost all of

the lentil area covered by improved variety is covered by

a single variety Alemaya. .

Table 2: Un-weighted varietal level adoption of

lentil varieties based on the household

survey, Ethiopia, 2011

Variety Simple Adoption Rate

% HH % Area

Alemaya 11.42 14.43

Ada’a 0.63 0.78

Local 87.95 84.79

27

As noted in the methodology section, statistics on lentil area

and number of holders cultivating lentil are available at

zonal level. Hence, weighted adoption estimates are

calculated at regional and national level. Table 11 compares

simple adoption estimates of the proportion of households

and share of area under improved varieties with weighted

adoption estimates at regional and national levels. On the

whole, weighted adoption estimates from the household

survey follow a similar trend to the simple adoption estimates.

At a national level, holder and area weighted adoption rates

are estimated at 12.0% and 15.6%, respectively, with an

absolute difference of 0.5 and 0.4 percentage points,

suggesting at a national level both weighted and un-weighted

estimates are comparable. Similarly, at regional level the

holder and area weighted adoption estimates are comparable

with non-weighted adoption estimates with absolute

difference of 1 percentage point or less further providing

evidence of the comparability of both estimates. Weighted

adoption estimates for lentil at a zone level, however, were

not estimated due to data limitations.

Table 31: Weighted and Un-weighted adoption estimates of the degree

of use of improved lentil varieties based on the household

survey, Ethiopia, 2011

Region Un-weighted

adoption

Weighted

adoption

Absolute

difference

%

HH

%

Area

% HH %

Area

%

HH

%

Area

Amhara 3.7 2.5 2.6 2.2 1.1 0.3

Oromia 29.8 34.4 30.1 38.6 -0.3 -4.2

National 12.5 15.2 12.0 15.6 0.5 -0.4

28

4.4 Adoption Estimates from Expert

Panel, Community FGD and

Household Survey

One of the objectives of the study is to explore whether

expert panel and community survey would provide a

reliable adoption estimates compared to the more

expensive and time demanding household surveys.

Consequently, as outlined in the methodology section of

the report, data were collected from expert panels,

community and household sample surveys. Unlike the

expert panel survey, the household and community

surveys provided adoption estimates at the lowest

administrative unit of a kebele in terms of proportion of

area share of improved lentil varieties as well as

proportion of holders growing improved varieties. The

panel of experts, however, provided adoption estimates

at a zonal level in terms of area share of improved lentil

varieties out of total cultivated lentil area.

Consequently, meaningful pair-wise comparison for un-

weighted adoption estimates for community and

household surveys could be made at district, zone,

regional and national levels. Pair-wise comparisons

involving expert panel estimates, however, are possible

at a regional level and only for the proportion of area

under improved varieties. Furthermore, comparison of

29

weighted and un-weighted adoption estimates from the

three estimates is made at a regional and national level.

4.3.1 Household and community level adoption

estimates

Table 12 presents un-weighted adoption estimates from

the household and community surveys in terms of

proportion of households using improved lentil varieties

and area share of improved lentil varieties. Zonal level

adoption rate estimates of household and community

survey based estimates show a similar pattern.

Nonetheless, on the average, weighted and un-weighted

adoption estimates from the household survey are

higher than community estimates by about 5 percentage

points suggesting community surveys are likely to

underestimate the use of improved lentil varieties by

smallholder farmers. Evident from the comparison is

that the divergence of adoption estimates from the two

data sources is higher for the Oromia Region where

relatively higher adoption is reported compared to the

Amhara Region where adoption rates are generally low.

30

Table 12: Comparison of un-weighted and weighted adoption estimates (% area under improved chickpea) from household and community surveys, Ethiopia

Region Zone Un-weighted adoption rate

(% of HH)

Un-weighted adoption rate

(% of Area)

HH survey

(A)

Community

estimate (B)

Deviation (A -

B)

HH survey

(A)

Community survey

(B)

Deviation (A -

B)

Amhara

North Gonder 0.0 0.6 -0.60 0.0 1 -1.0

North Shewa 8.3 2.2 6.09 5.2 20 -14.8

North Wollo 0.0 0 0.00 0.0 1 -1.0

South Gonder 0.0 1 -1.00 0.0 1 -1.0

South Wollo 0.0 0 0.00 0.0 3 -3.0

Oromia

East Shewa 95.2 57.3 37.86 88.6 30 58.6

North Shewa 5.4 6.7 -1.34 3.5 3 0.5

Southwest Shewa 0.0 0 0.00 0.0 8 -8.0

West Shewa 0.0 0 0.00 0.0 7 -7.0

Whole sample 12.5 7.1 5.40 15.2 10.2 5.0

31

4.3.2 Household and expert panel weighted

adoption estimates

The household sample survey was conducted in the two

major chickpea growing areas of Amhara and Oromia

regions whereas the expert panel provided estimates for

a much wider area. Table 13 presents area weighted

adoption rates from the expert panel and sample

household survey in terms of percent area share under

improved chickpea varieties. Comparison of the

regional level adoption estimates by considering only

the zones included in the national survey shows that the

expert panel estimates are higher for Amhara region by

about 7 percentage points and lower for Oromia Region

by about 24 percentage points. At national level, the

adoption estimates from the two estimates are pretty

close differing by about 5 percentage points (Table 13).

The fact that the estimates have converged better as we

aggregated from the regional level to the national level

is consistent with the theoretical expectation as upward

and downward deviations normally offset each other

leading to better estimates of the mean at a higher level.

32

Table 4: Comparison of area-weighted adoption estimates of improved lentil varieties from

expert panel with household sample survey estimates at Regional and

National level

Region Expert panel

(A1)

Household survey

(B1)

Deviation

(A1 - B1)

Amhara 9.4 2.2 7.2

Oromiya 14.8 38.6 -23.8

Total 10.8 15.6 -4.8

4.3.3 Weighted adoption estimates between

expert panel and community FGD

The area weighted adoption estimates at regional and

national level measured as percentage of area under

improved lentil varieties from the expert panel and

community survey are provided in Table 14. Area weighted

adoption estimates are higher in Amhara region where the

absolute level of adoption is small and higher for the

Oromia region where the absolute level of adoption is

relatively higher. At a national level the adoption levels

from the two estimates differ by about 3 percentage points

in favor of community survey.

33

Table 14: Comparison of area-weighted adoption rates (% area) of improved lentil

varieties from expert panel and community surveys at regional and

national levels

Region Expert panel

(A1)

Community survey

(B1)

Deviation

(A1-B1)

Amhara 9.4 2.0 7.4

Oromiya 14.8 32.9 -18.1

Total 10.8 13.4 -2.6

4.4.4 Weighted adoption estimates between

household and community levels

Table 15 compares area weighted adoption rates from

the household and community sample surveys in terms

of percent area share under improved lentil varieties.

Estimates from household survey are lower than the

community estimates in the Amhara region by about 7

percentage points but higher in Oromia region by about

24 percentage points. At a national level the two

estimates differed by 5 percentage points with 15.6%

and 10.8% adoption rates for the household and

community survey based estimates, respectively.

Table 5: Comparison area-weighted adoption rates of improved lentil varieties from

household survey with community estimates at regional and national levels

Region Household survey

(A1)

Community survey

(B1)

Deviation

(A1-B1)

Amhara 2.2 9.4 -7.2

Oromiya 38.6 14.8 23.8

Total 15.6 10.8 4.8

34

Table 16 summarizes the un-weighted and weighted

adoption rate estimates from the expert panel, community

and household surveys. Generally, the weighted and un-

weighted adoption estimates follow a similar pattern. At a

national level, on the average area weighted adoption

estimates are higher than un-weighted estimates

regardless of the data sources. The divergence between

weighted and un-weighted adoption estimates is not

unexpected as the later takes into account lentil area

difference between the regions. Of the three data sources,

on the average, adoption estimates from household

surveys are higher than expert panel and community

based survey estimates.

Table 6: Summary of adoption rates of improved lentil varieties at

national level from expert panel, community FGD and

household survey

Data Source Estimation Type

Simple Weighted

%

HH

%

Area

%

HH

%

Area

Expert Panel n.a 10.2 n.a 10.8

Community FGD 5.5 7.4 7.1 13.4

HH Survey 12.5 15.2 12.0 15.6

Note: n.a= not available

35

5. Conclusions and Lessons

Learned

The study documents and compares the adoption and

diffusion of improved lentil varieties derived from three

complementary data collection approaches namely,

expert panel, community and household sample surveys

conducted in major lentil producing areas of Amhara and

Oromia regions of Ethiopia. Of the three data source,

adoption rates are highest from the sample household

survey. Irrespective of the data source, weighted adoption

rates are on the average higher than the simple (un-

weighted) adoption rates indicating failure to use proper

weights under estimates adoption rates. The difference

between weighted and un-weighted adoption estimates,

however, appears to be small from the household survey

than from community FGDs.

On the whole, at the national level, the expert panel,

community and household survey based estimates of the

area share of improved varieties correspond fairly well

with 10.8%, 13.4 and 15.6%, respectively; suggesting

expert panel and community survey could be used to

generate the desired information quickly and cheaply.

All the three estimates witnessed higher levels of

varietal use by smallholder farmers in the central

highlands where agricultural extension efforts by the

research system has been intensively conducted

36

suggesting research based extension efforts have been

the main trigger for improved varietal adoption.

6. References

Central Statistical Agency (CSA). 2011. Report on Area and production of

major Crops. Agricultural sample survey 2010 / 2011 (2003 e.c.),

statistical bulletin, volume Ii. Addis Ababa, Ethiopia.

Chilot Yirga, Yigezu A., and Aden Aw-Hassan. (2015). Tracking Adoption

and Diffusion of Improved Chickpea Varieties: Comparison of

Approaches. Research Report 107. EIAR, Addis Ababa, Ethiopia.

International Food Policy Institute (IFPRI). 2010. Pulses Value Chain

Potential in Ethiopia: Constraints and opportunities for enhancing

exports.

Walker T, A Alene, J Ndjeunga., R Labarta, Y Yigezu, A Diagne, R Andrade,

R Muthoni Andriatsitohaina,, H De Groote, K Mausch,, C Yirga, F

Simtowe, E Katungi, W Jogo, M Jaleta, and S Pandey. 2014. Measuring

the Effectiveness of Crop Improvement Research in Sub-Saharan Africa

from the Perspectives of Varietal Output, Adoption, and Change: 20

Crops, 30 Countries, and 1150 Cultivars in Farmers’ Fields. Report of

the Standing Panel on Impact Assessment (SPIA), CGIAR Independent

Science and Partnership Council (ISPC) Secretariat: Rome, Italy.

37

7. Appendices

Appendix 1: Improved lentil varieties released by the national agricultural research system

Variety Year of

release

100-seed

wt. (g

Planting date Seed rate

(kg/ha)

Adoption area Yield

(t/ha) Altitude(m)

Rainfall(mm)

EL 142 1980 2.0-4.0 Mid July 50-60 1650-2000 600 400-600

R 182 2.6 Late June to Early July 60.70 1800-2400 500-1200 1.7-2.5

R-186 1980 2.0-3.0 Mid June to Early July 65-80 1800-2400 500-1200 1.8-2.5

Chalew(NEL358) 1984 2.0-3.0 Mid July 50-65 1850-2450 500-1200 1.9-2.6

Checol(NEL-2704 1984 3.0-4.0 Mid July 50-65 1600-2200 500-1200 1.5-2.2

Gudo(FLIP84-78L) 1995 5.0-6.0 Mid July 80-120 1850-2450 500-1100 1.8-2.5

Ada(FLIP-86-14L) 1995 1.0-2.0 Mid July 80-120 1850-2450 500-1100 1.9-2.6

Alemaya(FLIP89-63L) 1997/98 2.5-3.2 Early July 75-80 1600-2600 500-1200 2.0-3.0

Assano(FLIP(88—46) 2003 4.3 Aug. 1- 15 65 1800-2600 750-1000 3.2

Alem Tena(FLIP96-49L) 2004 2.9-3.9 Early to late July 85-90 1600-2000 400-600 1.7-2.3

Teshale(FLIP96-46L) 2004 3.1-4.3 Early to late July 85-100 1800-2400 400-800 1.8-3.7

Derso(Alemaya FLIP-88-411-02-AK-14) 2010 3.2-3.7 Early to Mid July 85-100 1600-2400 400-800 2.3-3.0

Source: crop variety register (different issue No. from 1998-2010)