A quantitative analysis of the characteristics of pesticide use among colombian agricultural workers

Abstract

The campesino agricultural communities in Colombia are vulnerable in many domains (economic, health) and they remain a crucial target population for medical and public health studies. Campesinos are at a high risk of occupational hazards, such as those derived from pesticide use and exposure. This study focused on individual factors that influence pesticide use and exposure. I collected 79 questionnaires with adult (male and female) rural agricultural campesinos in San Cristobal, Antioquia, Colombia. We examined campesinos' perceptions on pesticides' harm for human health and the environment, as well as beliefs, attitudes, perceived control and perceived confidence related to pesticide use. The findings differed between two campesino groups sampled: pesticide users and agroecological adherents. Pesticide users experienced lower perceptions of pesticide harm for human health and the environment, lower perceived control about stopping pesticide use (an increase in one unit in control decreases the logit of using pesticides by 74%), and lower confidence about stopping pesticide use than agroecological adherents (an increase in one unit in confidence decreases the logit of using pesticides by 64%). We discuss individual influences on occupational safety and health and recommend that future public health and educational interventions should improve safety training and confidence so campesinos can gain control of the process for implementing behavioral change related to pesticide use reduction.

Full text

A quantitative analysis of Colombian campesinos’
use of pesticides: perceived control and confidence
in this use
Un análisis cuantitativo del uso de Plaguicidas en los campesinos
colombianos: percepción del control y la confianza en este uso
Ysabel Polanco1; Juan C. Salazar2; Barbara Curbow3
1 PhD en Salud Pública, Universidad de Antioquia, Medellín, Colombia. Correo electrónico: ysabelpoldem@gmail.com.
2 PhD en Estadística. Universidad Nacional de Colombia, Medellín, Colombia. Correo electrónico: jcsalaza@unal.edu.co.
3 PhD Salud Pública. Department of Behavioral Sciences and Community Health, College of Public Health and Health Professions. University
of Florida. Correo electrónico: bcurbow@phhp.ufl.edu.
Recibido: 04 de junio de 2013. Aprobado: 30 de abril de 2014.
Polanco Y,, Salazar JC,Curbow B. A quantitative analysis of Colombian campesinos’ use of pesticides: perceived control and
condence in this use. Rev. Fac. Nac. Salud Pública 2014; 32(3): 373-382.
Abstract
Objective: this study aims to explore how campesino non
pesticide users differ in their knowledge, attitudes and beliefs,
and perceptions of control and condence surrounding
pesticide use compared to pesticide users. Methodology:
we collected 79 questionnaires with adult (male and female)
rural agricultural campesinos in San Cristobal, Antioquia,
Colombia. We examined the campesinos’ perceptions on
pesticides’ harm for human health and the environment, as
well as their beliefs, attitudes, perceived control, and perceived
condence related to pesticide use. Results: The ndings
differed between two campesino groups sampled: pesticide
users and agroecological adherents. Pesticide users showed
lower perceptions of pesticide harm for human health and the
environment, lower perceived control about stopping pesticide
use (an increase of one unit in control decreases the logit of
using pesticides by 74%), and lower condence about stopping
pesticide use than agroecological adherents (an increase of
one unit in condence decreases the logit of using pesticides
by 64%). Discussion: Several individual factors inuence
the occupational safety and health of campesinos, including
perceived control and condence. We recommend that future
public health and educational interventions should improve
safety training and condence so campesinos can gain control
of the process for implementing behavioral change related to
pesticide use reduction.
-----------Keywords: pesticides, campesinos, risk perception,
perceived control.
Resumen
Objetivo: Este estudio pretende explorar como los campesinos
no usuarios de plaguicidas dieren en su conocimiento, actitudes
y creencias y percepciones de control y conanza en torno al
uso de plaguicidas comparado con usuarios de plaguicidas.
Metodología: Colectamos 79 cuestionarios con hombres
y mujeres adultas que son campesinos agricultores en San
Cristóbal, Antioquia, Colombia. Examinamos las percepciones
de estos campesinos sobre el daño de los plaguicidas sobre
la salud humana y el medio ambiente. También se exploraron
sus creencias, actitudes, percepciones de control y conanza
relacionadas con el uso de los plaguicidas. Resultados:
Los hallazgos dieren entre los dos grupos de campesinos
muestreados: usuarios y no usuarios de plaguicidas. Los
usuarios de pesticidas experimentaron menores percepciones
del daño de los plaguicidas sobre la salud humana y el medio
ambiente, menor percepción del control acerca de parar el
uso de plaguicidas (el incremento de una unidad en control
disminuyó el logit de usar pesticidas en un 74%), y menor

374 Universidad de Antioquia
Rev. Fac. Nac. Salud Pública Vol. 32 N.º 3 septiembre-diciembre 2014
conanza acerca de parar el uso de pesticidas comparado
con los no usuarios de plaguicidas (el aumento de una unidad
en la conanza, disminuyó el logit de usar pesticidas en un
64%). Discusión: Varios factores individuales inuencian la
salud ocupacional y la salud de los campesinos incluyendo
la percepción del control y la conanza. Recomendamos que
futuras intervenciones educativas en salud pública aumenten el
entrenamiento sobre prácticas seguras y la conanza. Así los
campesinos pueden ganar control en el proceso de implementar
cambios en su comportamiento relacionado con la reducción
del uso de pesticidas.
-----------Palabras clave: plaguicidas, campesinos, percepción
del riesgo, control percibido
* The term pesticide refers to any synthetic chemical substance intended for preventing, destroying, repelling, or mitigating pests. Pests can be
insects, other animals, weeds, fungi, or microorganisms that cause damage to crops or animals. The term includes insecticides, herbicides, fun-
gicides, and other substances used to control pests. Many of these substances are known to have adverse effects on human and ecosystem health
because they are designed to kill or otherwise adversely affect living organisms. The term includes Organophosphate, Carbamates, Pyrethroid,
and Organochlorineinsecticides. The definition excludes biopesticides derived from natural materials such as animals, plants, bacteria, and cer-
tain minerals (e.g., canola oil, baking soda, garlic, peppers, herb extracts often used in agroecological practices). Biopesticides are divided into
three categories: microbial pesticides (bacterium, fungus, virus or protozoan), biochemical pesticides (non-toxic naturally occurring substances
such as insect sex pheromones and scented plant extracts), and plant-incorporatedprotectants [source http://www.epa.gov/pesticides/about/].
Introduction
Pesticides* have constituted an immediate solution to pest
problems in crops [1]. However, pesticides have produced
adverse effects on humans as well as the environment that
make their use unsustainable [2-4]. Pesticides are also
responsible for the emergence of increasingly resistant
pests/pathogens and the destruction of natural pest enemies
[1]. Biodiversity has been strongly altered by synthetic
pesticides. Pesticide residues in different ecosystems, soil,
water, air, animals, and plants are other deleterious results
of conventional agricultural practices [5].
Campesinos in many countries where conventional
agrochemical-based agriculture constitutes the
predominant form of production in the agricultural
sector directly suffer from all of the severe threats to
human health and well-being posed by exposure to
pesticides [6, 7]. Most of the pesticides that are banned
in rst-world countries are widely employed in Latin
America [8]. In this article, we describe and analyze
the individual factors of risk perceptions, knowledge,
beliefs, and attitudes associated with pesticide use by
campesinos employing a quantitative approach.
Many health alterations have been associated with
pesticide use [9-12]. A variety of human systems can
be affected, including the neurological, immunological,
respiratory, and reproductive systems [6, 13-15].
Several potential means of exposure exist [5, 16, 17].
The greatest risk and highest toxicity are linked to skin
and mucosa contact with pesticides during preparation,
mixing, and application [18].
Effects of exposure range from acute intoxication
to chronic conditions including developmental and
neurologic alterations [6, 10, 12, 16]. Exposure not
only affects the farmworker directly involved in
the manipulation of toxic substances but also other
household members such as children and pregnant
women [11, 14, 19]. Even though a growing concern
exists about pesticide exposure of farmworkers and their
families, relatively few studies have tried to test directly
the association of behavioral factors with pesticide
exposure in an agricultural population [20].
Several research studies have explored beliefs, attitudes,
knowledge, risk perceptions, perceived condence, and
perceived control surrounding pesticide use [21-25]. Some
of these studies have explored the perceptions of control,
perceptions of risk, and pesticide knowledge among Latino
farmworkers in the United States, nding limited knowledge
and low risk perceptions (they do not feel at risk when they
mix, prepare, or spray with pesticides). These ndings
suggest that more research is needed and that different public
health interventions are important [21, 23, 26, 27].
A study performed with adolescent Latino
farmworkers exploring knowledge and risk perception
about pesticides found that use of protective equipment
was decient. A large proportion of these farmworkers
(42.2%) reported the belief that they were never
exposed to pesticides in their work, and many (40.2%)
reported that there were no ways to protect themselves
from pesticide exposure. However, the large majority
(79.4%) acknowledged that pesticides can cause health
problems, and over half of the respondents (54%)
indicated they have some concern that they have become
sick from being exposed to pesticides [21]. In North
Carolina, Latino farm workers varied noticeably in the
amount of safety training received and the adoption of
safety practices. Perceived lack of control was a relevant
factor that decreased workers’ use of safety practice
[24]. Knowledge is not the only factor controlling the
decisions reached by individuals involved in agricultural
work as far as the opportune adoption of protective
measures and/or in relation to reducing pesticide use.
Other individual factors, including beliefs, attitudes, and
perceptions, inuence protective behavior.

A quantitative analysis of Colombian campesinos’ use of pesticides...
Facultad Nacional de Salud Pública 375
Farmworkers are aware of the risks associated with
pesticide exposure and have varying opinions regarding
the individual’s perception of personal vulnerability [26].
In research performed with farmers and farmworkers in
North Carolina, us, farmworkers believed that they had
been exposed to pesticides because they experienced
symptoms during or after the application. Farmworkers
believed that susceptibility to chemicals was inherent
and beyond the individual’s control. This is an important
nding because ideas about “control” often predict
health behavior [22]. Thus, when farm workers felt they
had “control” over a specic health outcome, they were
more likely to adopt appropriate preventive measures
and behaviors when confronting health threats, as in the
case of pesticide exposure.
Previous work has indicated that farmworkers
experience high levels of perceived risk from pesticides
and low levels of perceived control of pesticide use and
safety measures. In a study of farm workers, Arcury
et al (2002) found that receiving information about
pesticide safety reduced perceived risk and increased
perceived control. However, perceived risk had a limited
relationship to safety knowledge and was not related to
safety behavior. Additionally, perceived control was not
related to pesticide exposure knowledge but was strongly
related to safety knowledge and safety behavior [23].
These results demonstrate that, for pesticide education
to be effective, it must address aspects of control [28].
In the ndings reported by Flocks (2007),
farmworkers attributed extreme weather conditions,
including hot and dry or hot and humid, as affecting their
bodies and making them “weaker” and more vulnerable
to pesticide exposure and absorption [25].
In Colombia, the situation related to pesticide use
for agriculture is multifaceted and difcult to assess. The
country is marked by sociopolitical instability, extreme
poverty, rampant violence and corruption, and a rather
strong inuence of private capital and private interests in
the design and implementation of public policies [19, 29].
This situation translates into weak regulatory frameworks
in several forms, including work safety, labor rights, and
compliance with the management of hazardous materials,
of which pesticides are just an example [30].
Farming is a major sector of the Colombian economy
[31]. The bulk of agricultural production is achieved by
conventional means of cultivation with heavy reliance on
the agrochemical industry [31]. Regulation for pesticide
use in Colombia exists but is not fully or consistently
enforced (30). For example, several pesticides that are
banned in other countries are still used in Colombia [6,
17, 29]. Additionally, the cut ower industry, a strong
component of national agribusiness with high amount
of exports annually, is dependent on the intensive use
of synthetic pesticides and fertilizers, putting a large
portion of the population of agricultural workers and their
communities at high risk [19]. In Colombia, pesticide
use and exposure is widespread [6]. Individual factors
have been scantily studied nding limited knowledge
about the harmfulness of pesticides among campesino
communities, as well as issues of power with the bosses
and social discrimination [19].
This study aims to explore how campesino non
pesticide users differ in their knowledge, attitudes
and beliefs, and perceptions of control and condence
surrounding pesticide use compared to pesticide users.
The research question in this study was: What are the
attitudes, beliefs, and risk perceptions of campesino
pesticide users and agroecological† adherents in San
Cristobal, and how do they differ in these factors?
San Cristóbal is a corregimiento (Rural Division)
in the municipality of Medellin (Central Antioquia,
Colombia). It includes the head of the township and
17 veredas (rural subdivisions) [32], ten of which were
selected for data collection in this project as follows: El
Patio, El Llano, Las Playas, La Cuchilla, La Palma, El
Carmelo, Travesías, Yolombo, El Uvito, and San José de
la Montaña. San Cristobal maintains strong economic,
commercial, and administrative ties with Medellin.
Agriculture constitutes the main economic activity
in San Cristóbal, including cut owers and a broad
range of vegetables. San Cristóbal ranks as the largest
horticultural rural division of Medellin and functions as
a major source of produce to the latter [32].
Methodology
This study was based on quantitative methods. In this
section, we describe the characteristics of participants,
participant recruitment process, steps for the
† Agroecological practices, or agroecology, refer to the application of ecological concepts and principles to the design and management of sustai-
nable agroecosystems. It provides the basic ecological principles for how to study, design, and manage agroecosystems that are both productive
and natural resource conserving, and that are also culturally sensitive, socially just, and economically viable. Agroecology goes beyond the use
of alternative practices to develop agroecosystems with the minimal dependence on high agrochemical and energy inputs, emphasizing complex
agricultural systems in which ecological interactions and synergisms between biological components provide the mechanisms for the systems
to sponsor their own soil fertility, productivity, and crop protection. Agroecology is also the holistic study of agroecosystems, including all
environmental and human elements. It focuses on the form, dynamics, and functions of agroecosystems’inter-relationships and the processes
in which they are involved. By understanding these ecological relationships and processes, agroecosystems can be manipulated to improve
production and to produce more sustainably, with fewer negative environmental or social impacts and fewer external inputs [8]

376 Universidad de Antioquia
Rev. Fac. Nac. Salud Pública Vol. 32 N.º 3 septiembre-diciembre 2014
administration of the instrument, data collection and
analysis, and protection of human subjects.
Participants
All participants involved in this investigation worked in
agriculture, and many of them combined this occupation
with other partial jobs. Most of the participants were
married (74.68%) adult males (54.43%) who have
worked in agriculture for a substantial period (average
~36 years in the eld). The average age was 53 years.
The average amount of years of education was 6.15
(Std. Dev. 3.82). All the demographic information is
described in table 1. The total sample of campesinos was
79. Out of this number, 43 (54%) were only pesticide
users, 15 (19%) were in the transition stage in which
they still used pesticides but were slowly introducing
agroecological practices to a portion of the land they
cultivate, and 21 (27%) were campesinos who had
already changed completely to agroecological practices.
For the analysis we put together the transition group
with the agroecological group in the same non users
group and named it agroecological adherents.
Jose de La Montaña, El Llano, and La Palma in order
to present the project and contact possible participants.
Cognitive Interviews
In this study, we performed two cognitive interviews with
campesinos to rene the instrument. I asked respondents
to thinkaloud as they attempted to answer questions. This
technique helped to identify problems with questions and
indicated possible solutions [34]. The only confusion for
participants was related to the initial ve-item Likert scale
designed to measure attitudes and beliefs about pesticide
harmfulness to human health and the environment and
usefulness of pesticides in crop production. This scale
originally went from strongly agree to strongly disagree.
After completing the cognitive interviews, we reduced the
scale of some questions to three items (yes, neutral, and
no) in order to simplify the options.
Board Meeting
We also attended the board of directors meeting of
Asociación Campesina Agroecológica de la region de
Boquerón, Corregimiento San Cristobal (Campesino
Agroecological Association of the Boquerón) (acab). This
association of campesinos works on different activities and
at different levels with their associates, including education
and training about agroecological practices, the planning of
their harvest, technical support with the management of their
crops, and organization of crop marketing, among others
[35]. We went to this meeting with the aim of presenting
the project to ACAB leaders and obtaining permission to
contact campesinos who belong to this association and
constituted the sample of the non pesticide users.
Questionnaire
We constructed the questionnaire based on the following
components: questions about attitudes and beliefs in a
three-item likert scale [36] (e.g., pesticides affect my
health, pesticides can affect water), questions about
pesticide decision making in a ve-item scale (e.g.,
when you decided to use pesticides, did you think it
was a very good, good, neither good or bad, bad, or
very bad decision?), perceived control about stopping
pesticide use (the response options were on a scale of
1-10, and the question was: if you decide you want to
stop using pesticides, how much control do you think
you have?), perceived condence about stopping
pesticide use(the response options were on a scale
of 1-10, and the question was:if you wanted to stop
using pesticides, how condent are you that you could
stop?), and demographic information. The rst round of
questionnaires was applied to a rst group of campesinos
(seeds) contacted through cier assistants. Following
these initial questionnaires, we asked each contact
to provide names and phone numbers of two or three
Table 1. Demographic information of participants
Demographics (%)
Gender:
Male 54.43
Female 45.57
Marital status
Cohabitates 7.59
Divorced 5.06
Married 74.68
Single 8.86
Widow 3.80
Age (median) 53.00 years
Education (mean) 6.15 years
Education (std) 3.32 years
Sampling
This project used a purposive sampling strategy,
which consisted of purposefully selecting participants
who would best help to understand the problem and
provide information pertaining to the research question
[33]. The rst step was to contact community leaders
through personnel at the nongovernmental organization
Corporaciónpara la Investigación y el Ecodessarrollo
Regional (Corporation for Research and Regional
Ecodevelopment) (cier). We attended several community
action board meetings (Juntas de AccionComunal) in
some of the rural subdivisions including Travesias, San

A quantitative analysis of Colombian campesinos’ use of pesticides...
Facultad Nacional de Salud Pública 377
neighbors. We used snowball sampling techniques [37]
in this portion of the study. The reading level of the
informed consent and questionnaire forms appropriate
for the target audience (fth-grade level).
Data Analysis
This study included 79 questionnaires. First, we lled
out the questionnaires on paper,then we entered the
data into Excel 7® program sheets, and nally we
saved the data on a computer. We stratied data by
pesticide user (n = 43) or non-user (n = 36). The semi-
structured portion had scales that provided ratings.
We processed and analyzed the statistical data using
SAS 9.2® software [38].
Analysis encompassed the following steps: 1)
descriptive statistics for each item, 2) analysis of
principal components of the scales to identify an
appropriated structure, 3) Chi-square tests to compare
responses among the groups (pesticide users and non
pesticide users) across categorical variables, building
a score of probability for questions about attitudes and
beliefs, 4) performing logistic regression analysis to
examine the relationship between pesticide use and the
variables condence and control, 5) calculating internal
reliability using Cronbach’s alpha [39] for attitudes and
beliefs, 6) running Spearman correlation coefcients
to detect monotonecorrelations between the variable
for attitudes and beliefs, and 7) conducting logistic
regression analysis with a dichotomized version of
perceived pesticide harm.
Human Subjects
Participants signed the informed consent form before the
questionnaire started. We conducted all questionnaires
in a private space. Each questionnaire had a unique code
number to protect the privacy of participants. The study
was approved by the Institutional Review Board of the
University of Florida.
Results
Below is a description of the analysis and results from
an exploration of the following individual factors among
participants: attitudes and beliefs related to pesticide use,
decision making about pesticide use, perceived control
about stopping pesticide use, and perceived condence
about stopping pesticide use. I described and compared
the ndings between pesticide users and non- users. I
report the ndings in the following order: descriptive,
bivariate, scale development, and multivariate.
Descriptive
We obtained descriptive statistics of both demographic
information and each item of the questionnaire to assess
the nature of the normal distributions.
Decisions about pesticide use
To evaluate campesinos’ perceptions about how good
or bad the pesticide decision was, we provided a Likert
scale with the following question: When you made the
decision of using pesticides, do you think this decision
was: 1=very good; 2=good; 3=neutral; 4=bad; 5=very
bad. We obtained descriptive statistics to see the
frequencies of responses. Most campesinos expressed
that pesticide use is very good (35.9%) or good (28.1%).
A minority considered it to be bad (16.67%) (table 2).
Therefore, it was clear that pesticide use was accepted in
this community, and it was considered to be something
very good or good by the majority of participants.
Table 2.Frequency of pesticide decision options.
The decision to use pesticide
was
Frequency Percent
%
Very good 28 35.9
Good 22 28.1
Neutral 9 11.54
Bad 13 16.67
Very bad 6 7.69
Bivariate Analysis
Perception about pesticide harm
We wanted to determine if non pesticide users have
a increased perception of the dangers and deleterious
impacts of pesticide use and exposure on human health
and the environment compared to pesticide users. To
answer this question, we built a score of probability
for questions that measured attitudes and beliefs.
This score was based on the questions of attitudes
and beliefs and reected the level of perceptions of
impact of pesticide use on health and the environment.
Small values of this score are related to high levels of
perception of the health and environmental impact of
pesticides (pesticide use affects human health and the
environment). We conrmed this idea using a logistic
regression model to check the variables perceived
pesticide harm(single predictor variable)related to
pesticide use(categorical outcome variable, which

378 Universidad de Antioquia
Rev. Fac. Nac. Salud Pública Vol. 32 N.º 3 septiembre-diciembre 2014
translates into use or not of pesticides). With a p-value
of 0.0007, these variables were strongly associated.
We calculated a Receiver Operating Characteristic
(roc) curve ‡, and the value was 0.776,meaning that the
model is suitable for classication. To assess the t of the
model, we calculated a likelihood ratio test (p-value <
0.0001). The effect size was measured as (0.205-1) x 100% =
-79.5% which means that an increase in one unit in the
score of perceived pesticide harm decreases the logit of
using pesticides by 79.5%.
Perceived control
Campesinos were asked to rate their perceptions on
their level of control to stop pesticide use on a scale
of 1 (no control) to 10 (extreme control). Out of the
total sampled population, 26.5 % responded perceiving
no control to stopping pesticide use. However, 33% of
participants responded perceiving maximum control
(table 3). Participants who perceived maximum control
were all non pesticide users. There is a clear difference
between the amount of perceived control of pesticide
users and non-users.
100% = -74.5% which means that an increase in one unit
in control decreases the logit§ of using pesticides by74.5%.
Perceived condence
To assess how much condence campesinos perceive they
have to stop the use of pesticides, we used a scale of 1(no
condence) to 10 (extreme condence). Most participants
(57.7%) perceived no condence at all to stop pesticide
use. However, some participants (16.7 %) expressed
perceiving maximum condence (table 4). It is important
to clarify that the class that represented perceiving
maximum condence was composed of agroecologists.
In other words, there is a clear difference in perceived
condence between pesticide users and non-users.
‡ A Receiver Operating Characteristic (ROC) curve plots the sensitivity against the false-positive rate (i.e., one minus specificity) for a range of
thresholds to help visualize test performance [40].
Table 3. Frequency of perceived control to stop pesticide use, with1
being no control and 10 being maximum control
§ Logit refers to the logarithm of the ratio of the probability of not using pesticide versus the probability of using pesticide (log(p/(1-p)))
Perceived
control Frequency Percent %
1 21 26.5
2 0 0
3 2 2.5
4 2 2.5
5 11 13.9
6 2 2.5
7 0 0
8 6 7.6
9 2 2.5
10 33 41.7
To explore if pesticide users perceive having less
control over their use of pesticides than non-users, We ran
a logistic regression analysis using the variable pesticide
use as the outcome variable (dependent) and perceived
control as the predictor variable (independent). I treated
the variable perceived control as continuous. The result
showed that the association between pesticide use and
control was signicant (p-value = 0.0004), which means
that higher control decreases the probability of using
pesticides. We measured the effect size as (0.2549 - 1) x
Perceived
confidence
Frequency Percent %
1 45 57.7
2 2 2.5
3 4 5.1
4 2 2.5
5 8 10.2
6 2 2.5
7 0 0
8 1 1.3
9 1 1.3
10 13 16.7
Table 4. Frequency of perceived confidence to stop pesticide use,
with 1 being no confidence and 10 being maximum confidence.
To explore if pesticide users perceive having less
condence to use pesticides than non pesticide users, we
tted a logistic regression model using the variable pesticide
use as the outcome variable (dependent) and perceived
condence as the predictor variable (independent). We
treated the variable perceived condence as continuous.
The analysis showed that the association between pesticide
use and perceived condence was signicant (p-value:
0.0007), meaning that higher condence decreases the
probability of pesticide use. We measured the effect size
as (0.3520 - 1) x 100% = -64.8%, which means that an
increase in one unit in condence decreases the logit of
using pesticides by64.8%.
Pesticide use with other variables
We conducted Chi square tests to compare the
groups of pesticide users and non-users (pesticide

A quantitative analysis of Colombian campesinos’ use of pesticides...
Facultad Nacional de Salud Pública 379
use) with other variables. This variable, pesticide
use, was only signicantly related with occupation
(p-value = 0.0494) and residence (p-value
=<.0001). It was borderline signicant with age
(younger than 53 years, which was the median,
p-value = 0.0561). We dichotomized age in two
categories: lower than 53 and higher than 53 (the median
value).When we ran the test with age being continuous,
we did not nd association with pesticide use.
Scale Development
Perceived pesticide harm
We carried out principal components analysis (PCA)
of the scales to visualize potential structures. We
implemented a principal component analysis to reduce
dimensionality from questions 1a to 1n (attitudes
and beliefs).In order to explain more variability, we
combined components 1 and 2 by using a weighted
average that was used to create an individual score
called perceived pesticide harm. This score went from
42 (maximum score, when the participant answered
no to all the questions) to 14 (minimum score, when
participant answered yes to all the questions). Large
values of this score were associated with people who do
not have a high perception of harm caused by pesticide
use and exposure on human and environmental health.
Because this perceived pesticide harm does not have a
reference value, to assess the denition of how large
is “large,” we mapped the score into the scale 0-1 by
using a logistic transformation. We called it perceived
pesticide harm (logistic).
We wanted to explore if these variables that measure
attitudes and beliefs were inter-related. For this, we used
an internal consistency indicator of reliability called
Cronbach’s alpha for variables that measured attitudes
and beliefs (1a-1n). Results indicated an alpha of 0.64,
indicating that the correlations of each item with every
other item were good, although not very strong. We
checked the values of alpha after deleting each item, and
alpha only increased to 0.67when variable e (pesticides
can affect the soil) was deleted.
To see the correlation between these variables, we ran
Pearson correlation, which detects just linear correlation.
We also ran Spearman correlation, which is very appropriate
because it detects correlations when they are not necessarily
linear (monotonic associations). These correlations can
be curvilinear. The most important correlation of the
variables 1a-1n (attitudes and beliefs) included: pesticide
affects pregnant women’s health with pesticides can harm
children’s health);pesticides can harm good insects with
pesticides affected my health; pesticides can harm the
water with pesticides can harm children’s health; pesticides
can harm children’s health with pesticides can harm the
water; and using pesticides allows me to provide more for
Type of
variable
Name or
variable
p-value Parameter
estimate
Dependent
variable
Perceived
pesticide harm
Independent
variable
Age (younger
than 53)
0.0480 -0.2480
Independent
variable
Residence:
Llano
Other
San José de la
Montana
Travesias
0.0097 0.1805
-0.2054
-0.1805
0.3219
Table 5. Beta regression procedure for estimating the odds of
perceived pesticide harm as a function of variables such as age,
residence, and occupation.
Multivariate Analysis
Because perception of harm includes values between 0
and 1, we implemented a beta regression model using
this score as a response. It is important to observe that
a high perception of harm (closer to 1) means that the
person has an enhanced perception of pesticides being
harmful for health and the environment. Only perception
of harm was signicant with age (p-value: 0.0480)
(younger than 53 years, which was the median) and with
residence (p-value: 0.0097) (table 5).
Additionally, we explored how the variable perceived
pesticide harm depends on the other variables. For this,
we conducted a logistic regression procedure with a
dichotomized version of the variable perceived pesticide
harm as response (0 if perceived pesticide harm < 0 and 1
if perceived pesticide harm ≥ 0). Only perceived pesticide
harm with age (younger than 53 years) as signicant
(p-value = 0.0529, borderline signicant). This is crucial,
as it implied that campesinos younger than 53 years
have less of a perception of the degree of harmfulness
of pesticides for human health and environmental than
campesinos older than 53 years. It is important to clarify
that, in this analysis, logistic regression works with
perceived pesticide harm dichotomized.
Discussion
There was a consistent coherency in the results of pa-
rameters assessed between users and non-users (e.g.,
non-users perceiving more condence and control
than pesticide users). This was an indication that both
groups were well differentiated in terms of what the use
pesticides implies for human health. It also delineated
a characteristic pattern of acceptance of the conditions
imposed from the outside (external, e.g., market pres-
sure, social norm, etc.) in the case of the user sample
my family economically with pesticides make the crop
look cleaner (Table 4 and 5).

380 Universidad de Antioquia
Rev. Fac. Nac. Salud Pública Vol. 32 N.º 3 septiembre-diciembre 2014
vs. a commitment to confront the established dogma in
the case of the non-user.
Pesticide users largely believe that pesticides are
not harmful for human health and the environment. In
contrast, nonpesticide users have a higher perception of
the harmfulness of pesticides for human health and the
environment, especially among older participants (older
than 53 years). Therefore, public health interventions in
this pesticide user population should focus on increasing
awareness about the harmfulness of pesticides for human
health and the environment; this would be benecial
especially when targeted to younger community members.
Non pesticide users perceive that they have more
control in relation to pesticide use than pesticide users.
In the case of perceived condence, most campesinos
expressed having very little or no condence to stop
pesticide use. However, the few participants who
perceived having high condence were all nonpesticide
users. Having higher control and higher condence
decreased the probability of using pesticides.
This study demonstrated that pesticide use for
campesinos represents an environmental justice
concern, as was also found in a study carried out
with farmworkers in the us [23]. A basic principle of
environmental justice is that local communities must
have control over their environment. The environmental
justice framework not only recognizes environmental
injustice as it is associated with humans harming nature,
but it also recognizes that environmental injustice arises
from class, racial, and gender discrimination [41, 42].
For education on preventing/reducing pesticide use
to be successful, it must address the crucial component
of control over pesticide use. In pesticide use and
safety education interventions, campesinos should be
told not only what they should do to reduce pesticide
use and exposure but also why and how this behavior
will decrease exposure and improve their health, i.e.,
building a clearer justication to explain behavior
[23]. Subsequently, future public health interventions
should be aimed at increasing perceived control and
perceived condence mainly among pesticide users.
As a consequence, this could lead to greater behavioral
change of decreasing pesticide use. In other words,
the tools for community empowerment should be built
with local communities instead of being brought in as
an external set of appropriate rules and procedures.
Examples of achievement in the change of paradigm
implied for the non-users can serve as a valuable
preliminary experience and a demonstration of how the
problem of human health alterations and environmental
degradation should have an interdisciplinary approach,
for example in the conjunction of agroecology, public
heath, and human behavior.
The individual factors that played a relevant role in
the decision-making process of pesticide use included
entrenched, and often interdependent, sets of beliefs such
as those related to pesticide use benets (is necessary
for crops, benets crops), low perceived control, low
perceived condence, and low perception of pesticide
harm for human health and the environment. Campesinos
who adhere to these established categories were more
likely to use pesticides. We built a conceptualmodel
that describes the most inuential individual factors
in the decision-making process of pesticide use in San
Cristobal, Colombia. The most relevant components
included beliefs (pesticides are necessary for the
crops, pesticides are not harmful for human health
and the environment, and pesticides benet crops,
making them bigger and increasing their amounts), low
perceived control, low perceived condence, and low
perception of pesticide harm (for human health and the
environment). All these factors increase the probability
of using pesticides (gure 1). Therefore, future public
health interventions must include these components in
order to reduce pesticide use in this community.
In terms of education, more programs and
interventions are needed for campesinos who express the
need and interest to learn more about the harmfulness of
pesticides and the benets of agroecological practices.
Undoubtedly, this education must be accompanied by
more support from stakeholders, institutions, and the
local government in order to provide sustainability in
these processes of change. These interventions must
not be limited to present pre-manufactured information.
Rather, these programs must help affected communities
to build an appropriate local knowledge so as to gain
awareness, control, and condence in the process for
implementing behavioral change on pesticide use and
exposure dynamics.
Figure 1. Conceptual model of the decision-making process of
pesticide use

A quantitative analysis of Colombian campesinos’ use of pesticides...
Facultad Nacional de Salud Pública 381
Moreover, future studies should attempt to use
behavioral, environmental, and psychosocial measures
to build a body of evidence with which to better
understand the risk factors for pesticide exposure among
agricultural workers [20].
Results obtained after statistical treatment of
perceived pesticide harm are more than worrying. They
imply that younger agriculture workers possess less of
a concern on the detrimental effects of pesticides on
their personal health, on the health of the community
around them, and, in general, on the health of the natural
environment on which the local community ultimately
depends. The new generation of campesinos may have
a hard time reorienting its approach to agriculture
in the direction of abolishing pesticide use though
agroecological practices, therefore imposing serious
threats for the health of the rest of the community. This
nding stresses the idea of reinforcing research projects,
educational campaigns, and public health programs
among communities of agriculturalists in rural Antioquia
where the adverse effects of pesticide use-abuse are just
beginning to be documented. Doing so, this will work
in the direction of strengthening our prospects in the
struggle for social and environmental justice.
Research of this kind is limited in Colombia;
therefore, more research is needed on different regions
of Colombia with the aim of exploring in more detail the
level of local knowledge, perceptions of risk and control,
and the beliefs and attitudes related to pesticide use.
Acknowledgments
We thank the compton International Foundation,
which provided the funding for the study, and all the
campesinos who participated.
References
1 Gomiero T, Paoletti MG, Pimentel D. Energy and environmental
issues in organic and conventional agriculture. Critical Reviews in
Plant Sciences. 2008;27(4):239-54.
2 Shiva V. Soil not Oil: Environmental Justice in a Time of Climate
Crisis. Cambridge, MA: South End Press; 2009.
3 Harrison J. Pesticide Drift and the Pursue of Environmental Justice.
First ed. London: Massachusetts Institute of Technology Press; 2011.
4 Pimentel D, Marklein A, Toth MA, Karpoff MN, Paul GS, Mc-
Cormack R, et al. Food Versus Biofuels: Environmental and Eco-
nomic Costs. Human Ecology. 2009 Feb;37(1):1-12.
5 Vergara RA. Intoxicaciones masivas con plaguicidas: importan-
cia de sus efectos socioeconómicos. . III Seminario Nacional
“Aconteceres Entomológicos”. Universidad Nacional de Colom-
bia, Sede Medellín; 2000.
6 Nivia E. Mujeres y Plaguicidas: una mirada a la situacion actual,
tendencias y riesgos de los plaguicidas. Estudio de caso Palmira,
Colombia Palmira: Rapalmira; 2000 October 2012.
7 Jiménez M, Muñoz E. Plaguicidas y la salud. El caso del corre-
gimiento de San Cristóbal. Pag 90. : Publicado por Corporación
Penca de Sábila. ; 1993.
8 Altieri M. Agroecology: the science of natural resource mana-
gement for poor farmers in marginal environments Agriculture,
Ecosystems & Environment. 2002;93(1):1-24.
9 Pimentel D. Sustaining agriculture and the rural environment.
Ecological Economics. 2007 Aug;63(2-3):636-.
10 Ott W, Steinemann A, Wallace L. Exposure Analysis. Boca Raton:
Taylor & Francis group; 2007.
11 Freeman N. Children’s Risk Assessment. In: Robson M, Toscano
W, editors. Risk Assessment for Environmental Health. San Fran-
cisco: Jossey-Bass; 2007. p. 628.
12 Rothlein J, Rohlman D, Lasarev M, Phillips J, Muniz J, McCau-
ley L. Organophosphate Pesticide Exposure and Neurobehavioral
Performance in Agricultural and Nonagricultural Hispanic Wor-
kers. Environ Health Perspect. 2006;114:691-6.
13 Quandt SA, Arcury TA, Austin CK, Cabrera LF. Preventing Oc-
cupational Exposure to Pesticides: Using Participatory Research
with Latino Farmworkers to Develop an Intervention. Journal of
Immigrant Health. 2001;3(2):85-96.
14 Guillette E, Meza M, Aquilar M, Soto A, Enedina I. An Anthropo-
logical Approach to the Evaluation of Preschool Children Expo-
sed to Pesticides in Mexico. Environmental Health Perspectives.
1998;106(6):347-53.
15 Restrepo M, Muñoz N, Day N, Parra JE, Hernandez C, Blettner
M, et al. Birth defects among children born to a population occu-
pationally exposed to pesticides in Colombia. Scandinavian Jour-
nal of Work Environment and Health. 1990;16:239-46.
16 Reigart J, Roberts J. Recognition and Management of Pesticide
Poisonings. 5th ed. Washington DC: U.S. Environmental Protec-
tion Agency; 1999.
17 Ministerio_de_Salud. Pesticidas en Latino América. Colección:
Salud, Ambiente y Desarrollo. Santa Fe de Bogota, Colombia:
Ministerio de Salud; 2010.
18 Montana_State_University. Personal Protective Equipment (PPE)
for Pesticide Applicators. 2014 [cited 2014 February 27]; Availa-
ble from: http://www.pesticides.montana.edu/reference/ppe.htm
19 Sanmiguel-Valderrama O. The feminization and racialization la-
bor in the Colombian fresh-cut ower industry. Journal of Deve-
loping Societies. 2007;23(1-2):71-88.
20 Quandt S, Hernandex-Valero M, Grzywacz J, Hovey J, Gonzales
M, Arcury T. Workplace, household, and personal predictors of
pesticide exposure for farmworkers Environmental Health Pers-
pectives. 2006;114(6):943 – 52.
21 McCauley L, Sticker D, Bryan C, Lasarev M, Scherer J. Pesti-
cide Knowledge and Risk Perception Among Adolescent La-
tino Farmworkers. Journal of Agricultural Safety and Health.
2002;8(4):397-409.
22 Quandt S, Arcury T, Austin C, Saavedra R. Farmworker and Far-
mer Perceptions of Farmworker Agricultural Chemical Exposure
in North Carolina Human Organizations. 1998;57(3):359-68.
23 Arcury T, Quandt S, Russel G. Pesticide Safety among Farmwor-
kers: Perceived Risk and Perceived Control as Factors Reec-
ting Environmental Justice. Environmental Health Perspectives.
2002;110(2):233-40.
24 Elmore R, Arcury T. Pesticide Exposure beliefs Among Latino
Farmworkers in North Carolina’s Christmas Tree Industry. Ame-
rical Journal of Industrial Medicine. 2001;40:153-60.

382 Universidad de Antioquia
Rev. Fac. Nac. Salud Pública Vol. 32 N.º 3 septiembre-diciembre 2014
25 Flocks J, Monaghan P, Albrecht S, Bahena A. Florida Farmwor-
kers’ Perceptions and Lay Knowledge of Occupational Pesticides.
Journal of Community Health. 2007;32(3):181-94.
26 Salazar M, Napolitano M, Scherer J, McCauley L. Hispanic Ado-
lescent Farmworkers’ Perceptions Associated With Pesticide Ex-
posure. Western Journal of Nursing Research. 2004;26(2):146-66.
27 Grieshop J, Stiles M, Villanueva N. Prevention and Resiliency: A
Cross Cultural View of Farmworkers’ and Farmers’ Beliefs about
Work Safety. Human Organizations. 1996;55(1):25-32.
28 Vaughan E, Fridlund-Dunton G. Difcult socio-economic circum-
stances and the utilization of risk information: A study of Mexican
agricultural workers in the USA. Health, Risk & Society. 2006
September;9(3):323-41.
29 Cardenas A. Lineamientos de política sobre uso y manejo mesura-
do de plaguicidas con énfasis en el sector agropecuario y forestal
del departamento de Antioquia. Convenio Interinstitucional COR-
NARE - LA CEIBA - DAMA Consejo Seccional de Plaguicidas
de Antioquia Medellín. Colombia.; 2005.
30 Castro LA. Pesticides in Colombia: their application, use, and le-
gislation. . In: Taylor.M, editor. Pesticides residues in coastal tro-
pical ecosystems: distribution, fate and effects: Taylor and Francis
group.; 2003. p. 375-99.
31 Perfetti J, Balcázar A, Hernández A, Leibovich J. Políticas para
el desarrollo de la agricultura en Colombia. 2014 [cited 2014
February 26]; Available from: http://www.fedesarrollo.org.co/wp-
content/uploads/2013/07/Libro-SAC_Web.pdf
32 Medellin Ad. Atlas Veredal de Medellin. Medellin: Universidad
Nacional de Colombia; 2010.
33 Creswell J. Research Design: Qualitative, Quantitative and Mixed
Methods Approaches. Third edition ed: SAGE Publications; 2009.
34 Drennan J. Cognitive interviewing: Verbal data in the design and
pretesting of Questionnaires. Methodological issues in Nursing
Research. 2003;42:57-63.
35 ACAB. Asociacion Campesina Agroecologica de Boqueron.
2014 [cited 2014 January 22]; Available from: http://www.tien-
dacolyor.com/
36 Jamieson S. “Likert Scales: How to (Ab)use Them” Medical Edu-
cation. 2004;38(12):1217-8.
37 Bernard H. Research Methods in Anthropology: Qualitative and
Quantitative Approaches. Fourth edition ed: AltaMira Press 2006.
38 Delwiche L, Slaughter S. The Little SAS Book: A Primer Fourth
Edition ed: SAS Publishing 2008.
39 Cozby P. Methods in Behavioral Research. 10th ed. Fullerton: Mc
Graw Hill; 2009.
40 Vittinghoff E, Glidden D, Shiboski SC, McCulloch CE. Re-
gression Methods in Biostatistics: Linear, Logitic, Survival,
and Repeated Measures Models. First ed. New York: Springer
Science+Business Media; 2005.
41 Taylor D. The Rise of the Environmental Justice Paradigm:
Injustice Framing and the Social Construction of Environ-
mental Discourses American Behavioral Scientist. 2000 Ja-
nuary;43(4):508-80.
42 London L. Human rights, Environmental Justice, and the Health
of Farm Workers in South Africa. International Journal of Occu-
pational and Environmental Health. 2003;9:59-68.