Environ Dev Sustain (2013) 15:887–902

DOI 10.1007/s10668-012-9417-x

Pesticide handling practices by vegetable farmer

in Kenya

Ibrahim Macharia • Dagmar Mithöfer • Hermann Waibel

Received: 6 August 2012 / Accepted: 7 November 2012 / Published online: 24 November 2012
 Springer Science+Business Media Dordrecht 2012

Abstract Pesticide handling practices have a strong bearing on the exposure of toxic
effects to target and nontarget organism. A clear understanding of determinants of pesticide
handling practices is a precondition in the design and implementation of policy inter-
vention. To accomplish this, a household survey of 425 respondents was conducted in
2008. Majority of the farmers (85 %) had inappropriately handled pesticides, mainly
through, unsafe storage (23 %), unsafe disposal of leftover in either sprays solutions, or
rinsate and empty pesticide containers (40 %), failure to wear the required minimum
protective gear (68 %), or overdosed pesticides (27 %). However, majority of those
farmers were aware of the risks of pesticide use, with over 81 % expressing the view that
pesticides have harmful effects on human health, livestock, beneficial arthropods, and on
water. Econometric models showed that pesticide handling practices were significantly
influenced by variation in record keeping, main source advice on pesticide use, toxicity of
pesticide, and geographical location. Pesticide risk perception and negative impacts
experiences had no association with handling practices. The study recommends policy-
makers to design effective, participatory, and location targeted outreach programmes,
which deal specifically on promotion of record keeping and reduction in use of harmful
pesticides.

Keywords Pesticides  Agriculture  Kenya  Handling practices  Risks perception 

Farmers

I. Macharia (&)
Department of Agribusiness Management and Trade, Kenyatta University,
P.O. Box 43844, Nairobi 00100, Kenya
e-mail: mach4ibra@yahoo.com

D. Mithöfer
Rhine Waal University of Applied Science, Landwehr 4, 47533 Kleve, Germany
e-mail: dagmar.mithoefer@hochschule-rhein-waal.de

H. Waibel
Faculty of Economics and Business Administration, Leibniz University of Hanover,
Königsworther Platz 1, 30167 Hannover, Germany
e-mail: waibel@ifgb.uni-hannover.de

123
888 I. Macharia et al.

1 Introduction

Farmers use a wide range of pesticides to prevent crop losses from pest attacks. The ideal
situation for pesticides is that it reaches the target organism and, having achieved its
intended effect, decomposes rapidly into harmless compounds. However, this is not always
the case as most of the pesticides are accompanied by negative side effects. These side
effects include impairment of human and animal health, surface and ground water con-
tamination, pest resistance and resurgences, reduction in natural enemy populations,
damage to fisheries, fauna, and flora.
In Kenya, some research findings have indicated existence of pesticide-related neg-
ative effects, for example, pesticides threatening Lake Naivasha local hippopotamus
populations (IUCN 2005) and farmers’ health impairment (Ohayo-Mitoko et al. 2000;
Okello 2005; Asfaw 2008). Many of these impacts are a direct result of the inappropriate
handling of pesticides, often due to deviation from recommended application and han-
dling procedures.
Unsafe handling of pesticides usually due to negligence, lack of information, or lack of
training can pose a serious health risk for farmers who are the major pesticide users and are
regularly exposed to pesticides (Reeves and Schafer 2003). Several dimensions of unsafe
practices in the handling of pesticides include the following: Farmers may apply higher
than recommended dosage, store pesticides unsafely, dispose pesticides left over and
containers unsafely, or fail to wear the required personal protective equipment (PPE).
Extremely unsafe practices include mixing pesticide with bare hands, splashing pesticides
onto crops using brushes or twigs and tongue testing to assess concentration strength of the
chemical (Dinham 2003). This consequently increases the chances of pesticide side effects
on the farmer and the environment as a whole.
Safe handling of pesticides is considered a pivotal aspect in the reduction of health and
environmental hazards of pesticides (Keifer 2000). A study conducted by Mancini et al.
(2005) demonstrated that handling pesticides unsafely during spraying enhanced health
risks of farmers.
Production of vegetables in Kenya is mainly carried out by small-scale farmers tar-
geting local market (Mithöfer et al. 2008; HCDA 2010). The produce is marketed
through the informal sector, that is, open markets and kiosks, and currently, only 5 % are
sold through supermarkets (Tschirley et al. 2004). Kenya also exports green beans, peas,
and Asian vegetables to the European Union (EU) and the Middle East market (Harris
et al. 2001). Smallholders are estimated to account for 27 % of exported vegetables
(Jaffee 2003).
Insect pests and diseases generally limit production of vegetables. The majority of
smallholder vegetable farmers rely heavily on spraying pesticides to reduce the damage
caused by pests and diseases. Currently, relatively little is known about pesticide han-
dling practices in Kenya. A clear understanding of how vegetable farmers are handling
pesticides and the factors that influence those practices were deemed necessary for the
design and implementation of any policy intervention. Thus, the main objective of this
study was to identify the determinants for pesticide handling practices and develop
recommendations that can reduce the health and environmental hazards associated with
those practices. The study offers an opportunity for policy makers to understand the
underlying factors related to pesticide handling such that they can target capacity-
building efforts in those areas.

123
Pesticide handling practices 889

Reinforcing variables:
Previous experiences of pesticides Pesticides risk perception on:
related adverse event, observing Human health, livestock, water, birds,
others' actions and experiences, and beneficial arthropods
perceived impact of other farmers'
chemical use

Enabling variables:
Training, experience, main source of
pesticide use information, market, Safe pesticide handling practices:
social values, use of timely and • Personal protective equipment
reliable sources of information • Safe storage
• Safe disposal of empty containers
and rinsate
Producer characteristics variables: • Correct dose
Basic demographics, knowledge

Farm structure variables:

Farm size, crops, farm income

Fig. 1 Factors associated with risk perception and pesticide handling practices. Source: Own presentation

2 Materials and methods

2.1 Conceptual framework

The conceptual framework for the analysis of factors associated with farmers’ pesticide
risk perception and the determinants of their pesticide handling practices is presented in
Fig. 1. This framework is a combination of two existing analytical tools: the farm structure
theory developed for agricultural studies (Tucker and Napier 2001) and the psychometric
paradigm framework used in risk perception research (Slovic 2000). These tools incor-
porated farm-specific factors such as farm size and crops; individual features such as age,
gender, knowledge, training; and psychometric factors such as risk perception and who are
trusted sources for providing pesticide risk information. Within the psychometric para-
digm, people make quantitative judgments about the current and desired riskiness of
diverse hazards and the desired level of regulation of each. It is thus hypothesized that
farmer risk perception is influenced by socioeconomic and demographic factors, and in
turn drives farmer’s decisions on how to handle pesticides among the other factors.

2.2 Surveys and data

This paper uses cross-sectional data collected through a survey carried out in seven major
vegetable producing districts of Central province (Kiambu, Kirinyaga, Murang’a, Nyandarua
and Nyeri North) and Eastern province (Makueni and Meru Central) of Kenya in 2008.
The sample was a random subsample1 of 425 farmers, sampled by probability proportional
to size (PPS) from two previous surveys conducted in 2005 by the International Centre of

1
Follow up survey but due to resource constraints the sample was reduced.

123
890 I. Macharia et al.

Fig. 2 Study sites. Source: Own presentation based on GIS mapping of potential vegetable production areas

Insect Physiology and Ecology (ICIPE), that is, the Diamondback moth biological control
impact assessment survey (‘DBM’ with 295 farmers) and the Global Good Agricultural
Practices (‘GLOBALGAP’ with 544 farmers) assessment survey. GLOBALGAP (formerly
known as EUREPGAP) is a private sector body that sets voluntary standards for the certification
of agricultural products around the globe. The aim is to establish one standard for good agri-
cultural practice (GAP) by translating consumer requirements into agricultural production.
In both surveys, a multistage sampling procedure was employed to select districts, sub-
locations, and farmers. First, districts were purposely sampled according to intensity of
vegetable production and agroecological zone. Figure 2 shows the study sites.
Lists of farmers that were compiled by extension workers at the sub-location level,
served as a sampling frame from which the farmers were randomly sampled using the PPS
procedure. Table 1 displays the distribution of farmers in the sampled districts.
Structured questionnaires were used to collect information on pesticide use and practices,
risk perceptions, knowledge and experiences of pesticide negative impacts (health effects,
intoxication of livestock, mortality of beneficial arthropods and birds). For all the pesticide-
related impacts listed above, only the effects observed during the spraying operation or within
24 h after spraying were considered. In addition, the number of empty pesticide container
lying around the farm was also assessed and counted to further confirm and ascertain farmers’

123
Pesticide handling practices 891

Table 1 Regional distribution of survey respondents

Province District Main vegetable crops Previous No. of
surveys farmers
Domestic Export (2005) sampled
(2008)

Central Kiambu Cabbages, kales, spinach 48 27

Kirinyaga Peas, tomatoes French beans 155 74
Muranga Tomatoes, kales French beans 51 24
Nyandarua Cabbages, potatoes 119 52
Nyeri North Peas, cabbage, onions, carrots French beans 277 116
Eastern Makueni Cabbages, kales Asian vegetablesa 49 22
Meru Central Peas, tomatoes, cabbage, French beans 140 110
onions

Source: Own presentation

a
Brinjals, karella, dudhi, okra, turia, valore and aubergine

pesticide containers disposal methods. To test farmers’ recognition of the natural enemies and
soil biota known to them, a self-made arthropods zoo (clear plastic jar containing the most
common—natural enemies: ladybird beetles, praying mantis, spiders, dragonflies, fire ants
and soil biota: earthworms, millipedes, and crickets) was shown to them. Respondents were
also asked to show samples of pesticide containers or labels they had used to facilitate the
accurate recording of the names of pesticide products applied.
Ten pictograms used on pesticide labels in Kenya for instructions on appropriate pesticide
use were also shown to farmers in order to assess farmers’ understanding of them and their link
with actual pesticide handling practices. Interviewers were instructed to record all responses in
the exact words of the farmer, which were later compared with the correct instruction and scored
1 if it was right or zero when it was wrong. To measure farmer risk perception of the harmful
effects of pesticides on human health and the environment, five questions were asked. ‘In your
opinion, what is the effect of pesticide used in vegetable production on (1) human health (2)
livestock (3) beneficial arthropods (bees, natural enemies, soil biota), (4) water, frogs and fish
(5) birds?’ Farmers could choose one from the following answers: (1) beneficial (2) somewhat
beneficial (3) no effect (4) somewhat harmful (5) harmful, and (6) do not know.

2.3 Model

To model determinants of the degree of pesticide risk perceptions and pesticide handling
practices, we used a count data model (Poisson regression). Poisson regression model is com-
monly applied in the analysis of count data (Wooldridge 2006). The count of risk perceptions or
the counts of handling practices yi were assumed Poisson distributed and can be expressed as:
 
probðYi ¼ ki Þ ¼ eki kyi
i yi ! ð1Þ

where yi = 0, 1, 2….
The expected parameter ki is assumed both the mean and the variance of the count data
(yi). This property is called equi-dispersion. The distribution is extended to obtain a
regression model by allowing each observation yi to have a different value of k. The
popular formulation is an exponential relationship between the expectation rate and a set of
regressors, expressed as:

123
892 I. Macharia et al.

E½yi jxi  ¼ var½yi jxi  ¼ ki ¼ expðb0 xi þ li Þ ð2Þ

where xi is the vector of regressors, b is the vector of estimated parameters, which captures
the effect of the regressors on the dependent variables (i.e., pesticide risk perceptions and
pesticide handling practices), and l is the error term.
We used maximum likelihood estimation procedure for the estimation of the parameters b
and l in Stata Version 10 statistical package. For the empirical model, we began by specifying
the risk perception equation as Eq. 3 below and handling practices as the fourth equation.
To simplify counts of the farmer risk perception on human health, livestock, water,
birds, and beneficial arthropods, responses with the following: (1) beneficial (2) somewhat
beneficial, and (3) no effect were grouped together and coded as 0 and were referred to as
pesticide are ‘harmless’ whereas responses with (4) somewhat harmful and (5) harmful
were grouped and coded as 1 and referred as ‘harmful’.
Theoretically, farmers’ pesticide risk perception (PERCEPTION) can be shaped by a
variety of independent factors, including the potential health implications, formal educa-
tion, and experience2 (Warburton et al. 1995). The perception equation considered these
factors along with the gender, main source of pesticide use information, being the primary
pesticides applicator, target markets, toxicity levels of pesticide handled, and location.
Variable definitions and summary statistics are shown in Table 2.
0 1
IMPACT; EDUCATION; EXPERIENCE; EXPERIENCESQ; GENDER;
B C
B GLOBALGAP; APPLICATOR; RECORD; EXTENS; FARMER; RADIO; C
B C
PERCEPTION ¼ f B C
B TRADER; EXPORT; LOCAL; NPEST; PWHOIab; PWHOII; PWHOIII; C
@ A
PWHOU; District Dummies
ð3Þ
The pesticide handling practices (HANDLING) was constructed as a count of each of
the responses for overdose, unsafe storage of pesticides, unsafe disposal of pesticides
rinsates and containers, and failure to wear of the minimum PPE (long-sleeved shirt, long
trousers or overalls, gloves, and gumboot). Each was coded as 1 and referred to as inap-
propriate handling practices and 0 otherwise. The equation was then specified as a function
of farmer’s pesticide risk perception, gender, level of education, experience in agricultural
production, farm size, main source of pesticide use information, target markets, toxicity
levels of pesticide handled, being the primary pesticide applicator, and location.
0 1
PERCEPTION; IMPACT; EDUCATION; EXPERIENCE; EXPERIENCESQ;
B C
B GENDER; GLOBALGAP; APPLICATOR; RECORD; FARMSIZE; C
B C
HANDLING ¼ f B C
B EXTENS; FARMER; RADIO; TRADER; EXPORT; LOCAL; NPEST; C
@ A
PWHOIab; PWHOII; PWHOIII; PWHOU; District Dummies
ð4Þ
All variables were cross-checked for problem of multicollinearity, through the simple
correlation matrix between all the variables and the highest correlation coefficient was

2
Experience squared (EXPERIENCESQ) was added in the equation as there exist a concave relationship
between experience and returns to experience (Mincer 1974). Applied in our case it means that the more
experienced the farmers are, they may be expected to perceive pesticide to be risky and handle them safely.
However, this expectation may decline after a certain point due to depreciation effects of human capital.

123
Pesticide handling practices 893

Table 2 Definition and summary statistics of variables used in empirical estimations

Variable Definition Meana

Dependent variables
PERCEPTION Farmer perceive pesticides to be harmful to human health, 4.00 (1.23)
livestock, water and fish, beneficial arthropods and birds
(count)
HANDLING Farmer overdose pesticides, unsafely stored pesticides, unsafely 1.57 (0.98)
disposed of pesticides containers and failed to wear the
minimum PPE (count)
Reinforcing independent variables
IMPACT Farmer experienced/witnessed a pesticide associated human 2.16 (1.72)
health impairment, livestock poisoning, mortality of beneficial
arthropods and birds (count)
Farmer characteristics independent variables
EDUCATION None (%) 1.6
Primary school (%) 47.8
Secondary school (%) 39.5
College (%) 10.8
GENDER 1, if farmer is a male; 0, otherwise 0.70 (0.45)
Enabling independent variables
EXPERIENCE Farming experience in agriculture production (years) 20.85 (12.03)
EXPERIENCESQ Farming experience in agriculture production (years squared) 579.30 (634.73)
GLOBALGAP 1, if farmer has ever been GLOBALGAP certified 0.21 (0.40)
(proxy for pesticide use training); 0, otherwise
APPLICATOR 1, if the farmer is the primary pesticides applicator; 0, otherwise 0.86 (0.35)
EXTENS 1, if extension officers are the main source of advice on pesticide 0.23 (0.42)
use; 0, otherwise
FELLOW 1, if other farmers are the main source of advice on the pesticide 0.05 (0.21)
use; 0, otherwise
LABEL 1, if label is the main source of information on the pesticide use; 0.20 (0.40)
0, otherwise (base in estimation)
RADIO 1, if radio is the main source of information on the pesticide use; 0.29 (0.45)
0, otherwise
TRADER 1, if pesticide traders including agro-vets are the main source of 0.23 (0.22)
advice on the pesticide use; 0, otherwise
BOTH 1, if farmer producing for domestic as well as export market; 0, 0.46 (0.50)
otherwise (base in estimation)
EXPORT 1, if farmer produce exclusively for the export market; 0, 0.04 (0.19)
otherwise
LOCAL 1, if farmer produce exclusively for the local market; 0, 0.50 (0.50)
otherwise
Farm management and other independent variables
RECORD 1, if the farmer keep records of the pesticide use activities; 0, 0.31 (0.46)
otherwise
FARMSIZE Total farm size (hectares) 1.01 (0.96)
NPEST Number of pesticide products farmer handled (count) 4.06 (2.40)
PWHOIab Amount of pesticides applied and classified as WHO Ia or Ib 32.96 (227.77)
(extremely or highly hazardous) (g)

123
894 I. Macharia et al.

Table 2 continued

Variable Definition Meana

PWHOII Amount of pesticides applied and classified as WHO II 654.20 (1,244.38)

(moderately hazardous) (g)
PWHOIII Amount of pesticides applied and classified as WHO III 177.96 (495.60)
(slightly hazardous) (g)
PWHOU Amount of pesticides applied and classified as WHO U 301.44 (772.83)
(unlikely to present any acute hazard use) (g)
KIAMBU 1, if the farmer is located in Kiambu; 0, otherwise 0.06 (0.24)
KIRINYAGA 1, if the farmer is located in Kirinyaga; 0, otherwise 0.18 (0.02)
MAKUENI 1, if the farmer is located in Makueni; 0, otherwise 0.05 (0.22)
MERU 1, if the farmer is located in Meru Central; 0, otherwise 0.26 (0.44)
CENTRAL
MURANGA 1, if the farmer is located in Muranga; 0, otherwise (base in 0.06 (0.23)
estimation)
NYANDARUA 1, if the farmer is located in Nyandarua; 0, otherwise 0.12 (0.33)
NYERI NORTH 1, if the farmer is located in Nyeri North; 0, otherwise 0.27 (0.45)

Source: Own survey

a
Figures in parenthesis are standard deviations

0.38. Likewise, for endogeneity, none of the independent variables was suspected to be
explained within the equation in which it appeared. To check the robustness of the model,
we also fitted a negative binomial regression model, which is preferred when there is over-
dispersion (Long and Freese 2003). The likelihood ratio test and the statistical evidence did
not indicate over-dispersion.3 In addition, to check the robustness of all the models, other
restricted models were estimated in which subsequently insignificant variables were
dropped. The statistical quality of the models and the direction of the signs did not change,
and the coefficients deviated only marginally.

3 Results and discussion

3.1 Descriptive statistics

3.1.1 Farmer and vegetable production characteristics

The average age was calculated to be 46 years with nearly 23 % of the sample being under
the age of 35 and 29 % older than 50 years. Most farmers were literate, and only 2 % of
the farmers had not attended any formal school. Almost, half (47 %) of the farmers had
received formal primary education, whereas about 40 % had received secondary education,
and about 11 % had earned at least a diploma. The female farmers were approximately
30 %, and the average years of experience in agricultural activities by all the farmers were
21 years. Vegetable plot sizes varied between 0.02 ha to 10.1 ha with a median of 0.30 ha
per farmer. Farms varied in size from 0.02-18 ha and comprised five general cropping

3
Over-dispersion normally occurs when the mean \ variance.

123
Pesticide handling practices 895

Table 3 Pictograms presented to farmers and level of understanding i.e. knowledge and their practices (%)
Pictogram Meaning Know Practice p valuea

Yes No

Keep in a safe place out of reach of children Yes 47 13 0.41

No 29 10

Protect your feet/wear gumboots boots Yes 80 12 0.35

No 8 0

Wear protective clothing/apron Yes 59 24 0.34

No 10 6

Wear gloves Yes 32 61 0.56

No 2 5

Harmful to farm animalsb Yes 3 49 0.38

No 1 42

Harmful to aquatic animals like fishc Yes 33 18 0.25

No 28 22

Wash hand after use Yes 80 0 –

No 20 0

Cover your face/use a face shield Yes 32 50 0.43

No 8 10

Source: Own survey

a
From Fisher’s exact test
b
Practice: do not feed livestock with freshly sprayed vegetable residue
c
Practice: do not dispose of leftover spray solutions, rinsates or empty pesticide containers into pond, rivers
or dams

systems, that is, cereals, legumes, fruit crops, fodder crops for dairy, and intensive vege-
table plots, with a median of 0.80 ha.

3.1.2 Pesticides use

Farmers used 66 pesticide products comprising of 44 active ingredients. These included

insecticides (52 %), fungicides (42 %) herbicides (5 %) and 1 % acaricides. The com-
monly used products included dimethoate (WHO II), used by 48 % of farmers, lambda
cyhalothrin (WHO II—27 %), cymoxanil (WHO II—22 %) cypermethrin (WHO II—
22 %), cyfluthrin (WHO Ib—20 %), mancozeb (WHO U—18 %), and deltamethrin (WHO
II—14 %). The total amount of pesticides used was estimated at 570 metric tonnes for the
whole year (Macharia et al. (2009) estimation method). Forty-one percent of the volumes
belong to the group of carbamates, 19 % to pyrethroids, 16 % to organophosphates, 13 %
to acetamides, and 5 % to inorganics.

123
896 I. Macharia et al.

According to World Health Organization (WHO 2009) risk classification, 7 % of the

pesticide pesticides commonly used are extremely hazardous (WHO Ia and WHO Ib e.g.,
Methomyl (Lannate 90SP, Agrinate 90SP, Methomex 90) and Dichlorvos (Phosvit)) and
36 % moderately hazardous (WHO II e.g., dimethoate, cyhalothrin, cypermethrin).
Approximately 61 % are indicated by Pesticides Action Network North America (PAN) to
be bad actor chemicals.4 Comparing the products to the Environmental Impact Quotient5
(EIQ) following Mazlan and Mumford (2005) classification showed that 11 % were rated
as low (EIQ = 0–20), 61 % as medium (EIQ = 21–40) and 27 % as high (EIQ C 41).

3.1.3 Knowledge and handling practices

Table 3 summarizes the farmers’ knowledge of the safety measures pictograms nor-
mally found on pesticide labels on the Kenyan market and how they responded to them
(practiced). Though the majority (63 %) of farmers stated that they read and understand
pesticide labels, a sizeable percentage (65 %) did not know the correct meaning of all
the main and simple pictograms used in pesticide labeling with only 4 farmers adhering
to all.
A sizeable proportion (23 %) stored the pesticides unsafely in places such as in the
kitchen, bedrooms, and farm store together with farm produce and other equipment without
any safety precaution. Approximately 32 % reported wearing the required minimum
protective gear. By cross-observation a clear under use of PPE, particularly the use of
gloves was revealed. Only 1 out of 7 farmers was seen in gloves during spraying,
strengthening the validity of farmers’ responses. Low use of PPE has also reported among
farmers in other countries like in Ethiopia and the United States (Carpenter et al. 2002;
Mekonnen and Agonafir 2002). In most cases, use of PPE was very low despite the
availability of PPE and farmer awareness of the potential impact of pesticides on their
health. The farmers disposed empty pesticide containers within the farm by burying or
throwing into the latrine (56 %), disposal pit (28 %), dumping by the field (13 %), or
washed and reused (2 %). By cross-checking, an average of 2 pesticide containers was
observed lying either in or near the vegetable field, water ponds, and near homestead of 33,
2, and 13 % of the sample farmers, respectively. Hurtig et al. (2003) Ntow et al. (2006) and
Recena et al. (2006) reported similar unsafe disposal methods of empty pesticide con-
tainers. The majority of the farmers also indicated that they continuously sprayed the
leftover pesticide solutions on the same crop the same day (60 %), 12 % emptied the
leftover spray solutions nearby wells or ponds. Equipments used to apply the pesticides
were washed with a water hose near the homestead (44 %) or in the field using water from
ponds, streams or from the wells with 17 % releasing the rinsates into pond or stream.
The majority of the farmers (91 %) stated that they followed the label instructions when
determining the application rates. However, the comparison between the farmers’ appli-
cation rates to the existing recommended rates, (i.e., the application rates indicated
on pesticide labels, which were also cross-referenced with those conventionally put in

4
Chemicals that are highly acutely toxic, cholinesterase inhibitor, known/probable carcinogen, known
groundwater pollutant or known reproductive or developmental toxicant.
5
The EIQ calculation developed by Kovach et al. (1992) uses active ingredients of pesticides and applies a
rating system in ten categories to identify a single value of the environmental impact rating. The ten
categories include: (1) action mode of pesticides, (2) acute toxicity to birds, (3) fish, (4) bees, (5) acute
dermal toxicity, (6) long term health effects, (7) residue half-life in soil and (8) plant surface, (9) toxicity to
beneficial organisms, and (10) groundwater and runoff potential.

123
Pesticide handling practices 897

Table 4 Farmers experience and perception of harmful effects of pesticides (%), 2008
Category Impact Perception

Experienced Witnessed from neighbors Harm-less Harmful Don’t know

Human health 35 50 5 93 2
Livestock 12 30 8 85 7
Beneficial arthropods 54 57 25 70 4
Water, frogs and fisha – – 10 80 10
Birds 6 9 14 74 12
Average 27 37 12 81 7
Source: Own survey
a
It was difficult for farmers to associated pesticide use with fish mortality or notice contamination of water
by pesticides apart from the smell, which is subjective

company catalogs), showed that only three farmers actually sprayed the recommended
rates6 for all the pesticide sprayed. Even allowing an error of 25 % for the total number of
pesticides sprayed did not change the number of farmer. Approximately 27 % of the
farmers had overdosed pesticides with an average overuse rate of 0.42 kg/application.
However, only 10 % overdosed pesticides on the export crops as compared to 17 % on the
domestic crops. Overdose results in financial losses because of waste of pesticides, and
decreased yields due to phytotoxicity (Asogwa and Dongo 2009). However, the biggest
risk of overdose is the increased likelihood for the development of resistance against
pesticides, which can have devastating large-scale effects on crop production (Meijden
1998). It also increases the chances of pest resurgence due to destruction of natural control
organisms (Meijden 1998).
Many farmers recognized the major natural enemies’ found in the vegetable crops with
over 90 % identifying 3 out of the 4 insects that they were asked to identify (ladybird
beetles 90 %, spiders 99 %, fire ants 99 %), while only 66 % clearly identified dragonflies.
However, they were uncertain of their role.

3.1.4 Experience and perception of pesticide negative impacts

Over 35 % of the farmers reported at least one of a variety of acute illness symptoms of
pesticide poisoning within 24 h after spraying pesticides, with half indicating that they
witnessed a fellow farmer intoxicated by pesticides (Table 4). The most common symp-
toms reported were sneezing, headache, stomach pains, dizziness, burning skin/rash, eye
irritation, shortness of breath, backache, vomiting, blurred vision, and coughing in the
order of most frequently reported. These symptoms are associated with pesticide acute
poisoning (Extension Toxicology Network 2004). In addition, a sizable number had also
observed/witnessed livestock poisoning, mortality of beneficial arthropods, and birds
during or 24 h after spraying pesticides, and all attributed to the pesticide sprayed. Many
(81 %) farmers also regarded pesticides as harmful for their health and environment.

6
The amount the farmer exceeded from recommended dose indicated on the label of the pesticide container
for each of the individual pesticides was first calculated. If the farmer used more than the recommended
dose, over-dose was coded as 1 and if it was less than the recommended dose under-dose likewise coded 1.
This calculation was performed for all the pesticide products used in each application, then summed across
each farmer.

123
898 I. Macharia et al.

Table 5 Estimation results of farmer risk perception and pesticide handling model, 2008
Risk perception Pesticide handling
a
Coefficient Coefficienta

Unrestricted Restricted Unrestricted Restricted

IMPACT 0.01 (0.01) 0.03 (0.02)

PERCEPTION 0.01 (0.03)
EDUCATION 0.00 (0.02) -0.05 (0.04)
EXPERIENCE 0.01 (0.00) 0.01 (0.01)
EXPERIENCESQ -0.00 (0.00) -0.00 (0.00)
GENDER 0.06 (0.03)* 0.06 (0.03)* 0.05 (0.07)
GLOBALGAP 0.13 (0.04)*** 0.13 (0.04)*** -0.04 (0.09)
APPLICATOR -0.09 (0.04)** -0.09 (0.04)*** -0.04 (0.08)
RECORD 0.02 (0.03) -0.17 (0.07)** -0.18 (0.07)**
FARMSIZE -0.00 (0.03)
EXTENS 0.06 (0.04) 0.03 (0.11)
FARMER 0.14 (0.06)** 0.13 (0.05)*** 0.07 (0.14)
RADIO 0.04 (0.04) 0.05 (0.09)
TRADER -0.06 (0.05) 0.15 (0.09)* 0.09 (0.06)*
EXPORT -0.21 (0.09)** -0.26 (0.08)*** -0.11 (0.15)
LOCAL 0.08 (0.03)** 0.08 (0.03)*** -0.01 (0.07)
NPEST 0.02 (0.01)*** 0.02 (0.01)*** 0.02 (0.01)* 0.02 (0.01)*
PWHOIab 0.00 (0.00) 0.00 (0.00)
PWHOII -0.00 (0.00) 0.00 (0.00)***,c 0.00 (0.00)***
PWHOIII 0.00 (0.00) -0.00 (0.00)
PWHOU -0.00 (0.00)**,b -0.00 (0.00)** -0.00 (0.00)
KIAMBU -0.41 (0.11)*** -0.36 (0.08)*** -0.32 (0.22)
KIRINYAGA -0.21 (0.09)** -0.19 (0.04)** 0.40 (0.13)*** 0.34 (0.06)***
MAKUENI 0.05 (0.09) 0.32 (0.16)** 0.22 (0.11)**
MERU CENTRAL -0.09 (0.09) -0.29 (0.15)*** -0.34 (0.09)***
NYANDARUA 0.01 (0.09) 0.10 (0.15)
NYERI NORTH -0.02 (0.08) -0.00 (0.14)
Constant 1.23 (0.11)*** 1.31 (0.05)*** 0.22 (0.23) 0.36 (0.06)***
Observations 411 416 411 415
Log likelihood -729.92 -743.51 -567.46 -577.99
Wald v2 142.34*** 110.37*** 108.12*** 80.55***

Source: Own survey

a
Figures in parenthesis are robust standard errors, statistical significant at the 0.01 (***), 0.05 (**), 0.1 (*)
level of probability
b
z value = -1.89 and c z value = 2.61

3.2 Models estimation results

The results of the two models are presented in Table 5. Starting with the risk perception
model (Column 2, unrestricted), the results indicate that the probability of risk perception
significantly increases with male farmers, GLOBALGAP certification, fellow farmers as

123
Pesticide handling practices 899

the main sources of advice on pesticide use, production of vegetable geared exclusively for
domestic market as well as the number of pesticides handled. Farmers as the primary
pesticides applicator, growing vegetables targeting export markets, handling pesticide in
WHO U, and being located in the districts of Kirinyaga and Kiambu reduces the proba-
bility of pesticide risks perception. Neither experience/witness of pesticide-related negative
impact7 nor education had any significant effect on risk perceptions. The restricted model
reestimated by dropping insignificant variables, shows the estimates of the coefficients, and
their directions were robust.
The second equation for the pesticide handling model in Table 5 (column 3, unre-
stricted) clearly shows that the probability of inappropriate handling of pesticides is lower
with record keeping and being in the district of Meru Central. In general, record keeping of
pesticide products handled, their application dosage, application techniques, and produc-
tion activities enables a farmer to increase profits through better pesticide use planning.
With records, a farmer can also see how well she/he is managing production operations and
can identify the strengths and weaknesses in those activities. The positive significant
coefficient on pesticide traders as the main sources of advice on pesticide use, numbers of
pesticides handled, and handling of pesticides in WHO II, suggests a probability of
inappropriate handling of pesticides association. In the Philippines, increased pesticide
misuse was found to be strongly associated with visits by chemical company representa-
tives or by agricultural technicians (Tjornhom et al. 1997). Furthermore, pesticides dealers
particularly the companies have an incentive to push pesticides use by advertising and
promotion, and this creates a bias in favor of their use (Tisdell et al. 1984).
Analysis further shows that farmer located in the districts of Kirinyaga and Makueni
have higher probability of inappropriate handling of pesticides. Contrary to theoretical
expectations, pesticide risk perceptions and previous experience/witness of a negative
pesticide impact had no significant influence on pesticide handling practices.8 Kishi et al.
(1995) reported that farmers take pesticides poisoning symptoms as normal effects so they
get used to them. Similarly, the study in Côte d’Ivoire by Ajayi (2000) also showed that
pesticide applicators tended to accept a certain level of illness as an expected and normal
part of farming. This could be the reason why farmers did not handle pesticide safely even
after experiencing a negative impact. In addition, most farmers do not keep records of their
pesticides related losses, as they do not appreciate its importance. Furthermore, the lack of
diagnosis attributed to pesticide exposure make farmers also to ignore the dangers of
pesticide use as the long-term effect is not easy to prove (Pimentel and Greiner 1997).
Similar findings on lack of association between handling practices and risk perceptions
were reported in studies that showed that knowledge of the pesticide negative effects was
not directly reflected on the use of PPE (Martinez et al. 2004) or did not influence or
change farmers crop production practices (Ecobichon 2001). Tucker and Napier (2001)
also found that although some Midwestern US farmers were aware of potential negative
effects of pesticides use, they still relied heavily on chemical control.
Though the coefficients of farmer as the primary applicator, farm size, and GLOB-
ALGAP certification are insignificance, they had the expected negative sign. Close check

7
In an alternative model specification, with set of dummies for impacts on health, livestock, beneficial
arthropods, and birds yielded no statistical significance.
8
Even after controlling for specific impact and risk perceptions, i.e. on health, livestock, beneficial
arthropods, and birds no statistical significance was found.

123
900 I. Macharia et al.

on GLOBALGAP certification showed that many of the farmers (69 %)9 who were earlier
certified did not maintain their certification at the time of survey. Probability of these
farmers not following the recommended practices as required might offer a partial
explanation of this apparently perverse result.
When the model was reestimated (restricted) by dropping insignificant variables, the
estimates of the coefficients were again robust.

4 Conclusions and recommendations

The study shows that most farmers are aware of the pesticide side effects; however, they do
not handle them appropriately. The PPE gear in most cases was inadequate, and there exist
no regulations that require the use of protective gear during pesticide handling.
The regression results showed that record keeping, play a significant role in the
reduction of inappropriate pesticides handling practices, highlighting the need for
encouraging farmers on record keeping of their vegetable production activities. Handling
pesticides in WHO II and receiving advice on pesticides use from pesticide traders sig-
nificantly increase inappropriate pesticides handling practices.
It is remarkable that pesticide risk perceptions and previous experience/witness of a
negative pesticide impact have no direct influence on farmer’s pesticides handling prac-
tices. Hence, the learning effect of experience is very little. Such a trend is very worrying
because not only are the health of farmers affected, but the whole household also suffer.
Furthermore, the effect on entire society is likely to be considerable since water sources
and the entire environment are affected.
The results further suggest widespread inappropriate handling of pesticides in Kirinyaga
and Makueni districts. The district of Meru appears to be less prone to these practices,
perhaps due to record keeping.
The study ultimately recommends policymakers to design effective, more participatory
and targeted outreach programmes, which deal specifically on promotion of record keeping
and reduction on use of pesticides particularly in WHO II.
Farmer should be enlightened of the broader long-term negative effects of pesticides as
most farmers seems not to learn after experiencing the short-term pesticide negative
impact. Similarly, the results also point to specific district of Kirinyaga and Makueni
experiencing higher prevalence of inappropriate handling of pesticides, targeting these
areas may have the most measurable effects on reduction of inappropriate handling of
pesticides.

Acknowledgments This research was supported by DAAD and an ICIPE-GTZ/BMZ funded project. The
author expresses appreciation to farmers as well as staff members of the Ministry of Agriculture and other
individuals who provided valuable assistance.

References

Ajayi, O. C. (2000). Pesticide use practices, productivity and farmers’ health: The case of cotton-rice
systems in Cóte d’lvoire, West Africa. Pesticide Policy Project Special Issue No. 3. Institute of Hor-
ticultural Economics, Hannover, Germany.

9
It has been argued that smallholder vegetable farmers can achieve GLOBALGAP certification, but
continuous maintenance is a problem due to the high costs of compliance.

123
Pesticide handling practices 901

Asfaw, S. (2008). Global agrifood supply chain, EU food safety standards, and African small-scale pro-
ducers: The case of high value horticultural exports from Kenya. PhD. Dissertation. Hannover Uni-
versity, Germany.
Asogwa, E. U., & Dongo, L. N. (2009). Problems associated with pesticide usage and application in
Nigerian cocoa production: A review. African Journal of Agricultural Research,4, 675–683.
Carpenter, W. S., Lee, B. C., Gunderson, P. D., & Stueland, D. T. (2002). Assessment of personal protective
equipment use among Midwestern farmers. American Journal of Industrial Medicine,42, 236–247.
Dinham, B. (2003). Growing vegetables in developing countries for local urban populations and export
markets: Problems confronting small-scale producers. Pest Management Science,59, 575–582.
Ecobichon, D. J. (2001). Pesticide use in developing countries. Toxicology,160, 27–33.
Extension Toxicology Network. (2004). Cholinesterase inhibition. Toxicology information brief. Cornell
University. www.pmep.cce.cornell.edu/profiles/extoxnet/TIB/cholinesterase/html Accessed 12 Sept
2008.
Harris, C., Hegarty, P. V., Kherallah, M. X., Mukindia, C. A., Ngige, J. A., Sterns, P. A., Tatter, J. (2001).
The impacts of standards on the food sector of Kenya. Report organized by the Institute of Food and
Agricultural Standards at Michigan State University.
HCDA. (2010). Horticultural crops development authority. Horticultural Policy: CFR Policy Paper, HCDA,
Kenya.
Hurtig, A. K., Sebastian, M. S., Soto, A., Shingre, A., Zambrano, D., & Guerrero, W. (2003). Pesticide use
among farmers in the Amazon Basin of Ecuador. Archives of Environmental Health,58, 223–228.
IUCN. (2005). International union for the conservation of nature and natural resources: In FAO forestry
country profile. http://www.fao.org/forestry/foris/ Accessed 1 Nov 2007.
Jaffee, S. (2003). From challenge to opportunity: The transformation of the Kenyan fresh vegetable trade in
the context of emerging food-safety and other standards. Agricultural and Rural Development Working
Paper 2, The World Bank, Washington, DC.
Keifer, M. C. (2000). Effectiveness of interventions in reducing pesticide overexposure and poisonings.
American Journal of Preventive Medicine,18, 80–89.
Kishi, M., Hirschon, N., Djajadisastra, M., Satterlee, L. N., Strowman, S., & Dilts, R. (1995). Relationship
of pesticide spraying to signs and symptoms in Indonesian farmers. Scandinavian Journal of Work
Environmental Health,21, 124–133.
Kovach, J., Petzoldt, C., Degni, J., & Tette, J. (1992). A method to measure the environmental impact of
pesticides. New York’s Food and Life Science Bulletin,139, 1–8.
Long, J. S., & Freese, J. (2003). Regression models for categorical dependent variables using Stata (2nd
ed.). College Station, TX: Stata Press.
Macharia, I., Mithöfer, D., & Waibel, H. (2009). Potential environmental impacts of pesticide use in the
vegetable sub-sector in Kenya. African Journal of Horticultural Science,2, 138–151.
Mancini, F., Bruggen, A. H. C., Jiggins, J. L. S., Ambatipudi, A. C., & Murphy, H. (2005). Acute pesticide
poisoning among female and male cotton growers in India. International Journal of Occupational and
Environmental Health,11, 221–232.
Martinez, R., Gratton, T. B., Coggin, C., Rene, A., & Waller, W. (2004). A study of pesticides applicators in
Tarrant County. Texas Journal of Environmental Health,66, 34–37.
Mazlan, N., & Mumford, J. (2005). Insecticide use in cabbage pest management in the Cameroon Highlands.
Malaysia Crop Protection,24, 31–39.
Meijden, G. (1998). Pesticide application techniques in West Africa. A study by the agricultural engineering
branch of FAO through the FAO Regional Office for Africa, p. 17.
Mekonnen, Y., & Agonafir, T. (2002). Pesticide sprayer’s knowledge, attitude, and practice of pesticide use
on agriculture farms of Ethiopia. Occupational Medicine,52, 311–315.
Mincer, J. (1974). Schooling, experience and earnings. New York: Columbia University Press.
Mithöfer, D., Nang’ole, E., & Asfaw, S. (2008). Smallholder access to the export market: The case of
vegetables in Kenya. Outlook on Agriculture,37, 203–211.
Ntow, W. J., Gijzen, H. J., Kelderman, P., & Drechsel, P. (2006). Farmer perceptions and pesticide use
practices in vegetable production in Ghana. Pest Management Science,62, 356–365.
Ohayo-Mitoko, G. J., Kromhout, H., Simwa, J. M., Boleij, J. S., & Heederik, D. (2000). Self reported
symptoms and inhibition of acetyl-cholinesterase activity among Kenyan agricultural workers.
Occupational and Environmental Medicine,57, 195–200.
Okello, J. J. (2005). Compliance with international food-safety standards: The case of green bean pro-
duction in Kenyan family farms. PhD dissertation. Michigan State University, Michigan.
Pimentel, D., & Greiner, A. (1997). Environmental and socio-economic costs of pesticide use. In D.
Pimentel (Ed.), Techniques for reducing pesticide use: Economic and environmental benefits
(pp. 51–78). Chichester, UK: Wiley.

123
902 I. Macharia et al.

Recena, M. C. P., Caldas, E. D., Pires, D. X., & Pontes, E. R. J. C. (2006). Pesticides exposure in Culturama,
Brazil: Knowledge, attitudes, and practices. Environmental Research,102, 230–236.
Reeves, M., & Schafer, K. S. (2003). Greater risks, fewer rights: US farm workers and pesticides. Inter-
national Journal of Occupational and Environmental Health,9, 30–39.
Slovic, P. (2000). The perception of risk. London Sterling, VA UK: Earth scans Publications.
Tisdell, C. A., Auld, B., & Menz, K. M. (1984). On assessing the biological control of weeds. Protection
Ecology,6, 169–179.
Tjornhom, J. D., Norton, G. W., Heong, K. L., Talekar, N. S., & Gapud, V. P. (1997). Determinants of
pesticide misuse in Philippine onion production. Philippines Entomology,11, 139–149.
Tschirley, D. L., Ayieko, M., Mathenge, M., Weber, M. T. (2004). Where do consumers in Nairobi purchase
their food and why does this matter? The Need for investment to improve Kenya’s ‘‘traditional’’ food
market, Tegemeo Institute for Agricultural Policy and Development, Egerton University, Nairobi.
http://www.tegemeo.org/.
Tucker, M., & Napier, T. L. (2001). Determinants of perceived agricultural chemical risk in three water-
sheds in the Midwestern United States. Journal of Rural Studies,17, 219–233.
Warburton, H., Palis, F. G., & Pingali, P. L. (1995). Farmer perceptions, knowledge, and pesticide use
practices. In P. L. Pingali & P. A. Roger (Eds.), Impact of pesticides on farmer health and the rice
environment (pp. 59–95). Boston: Kluwer.
WHO. (2009). The World Health Organization (WHO) recommended classification of pesticides by hazard
and guidelines to classification. http://www.inchem.org/documents/pds/pdsother/class_2009.pdf.
Accessed 6 Nov 2012.
Wooldridge, J. M. (2006). Econometrics: A modern approach (3rd ed.). Cincinnati, OH: Southwestern
Press.

123