Farmers’ knowledge, attitudes and practices towards management of cassava pests and diseases in forest transition and Guinea savannah agro-ecological zones of Ghana

Ethical statement

Formal ethical approval was not obtained as the CSIR-Crops Research Institute has not yet put in place an ethical standing committee and according to local regulation is not required for studies not involving the collection of medical samples. The CSIR-Crops Research Institute administration approved the project and all its activities before the project was implemented. Though the Institute has not yet put in place an ethical standing committee, the project met all the institute’s expectations, guidelines and ethics. The study was conducted in line with Ghana’s Council for Scientific and Industrial Research Act, 1996 (Act 521) and the Data Protection Act, 2012 (Act 843). Moreover, respondents were informed about the purpose of the study and use of their responses in a language they understood and willingly gave their consent prior to data collection. Because participants could not read and write, informed consent was sought verbally, which was recorded via audio taping, as approved by the CSIR-Crops Research Institute. The data being presented do not include any personal data via which a specific respondent could be identified or traced.

Description of study areas

The study was conducted in two cassava agro-ecologies of Ghana, Guinea Savannah and Forest Transition, involving two study districts. These two agro-ecologies were originally part of regions selected for a nationwide disease surveillance in 2015/2016/2017 as part of the implementation of the West Africa Virus Epidemiology Project funded by the Bill and Melinda Gates Foundation and the Department for International Development, UK. The chosen agro-ecologies are part of the major cassava growing agro-ecologies in the country. The Eastern region which is within the Forest Transition is ranked 1^st and the Northern region which falls within the Guinea Savannah ranks 5^th. The former region recorded a production level of 4,649,507.58 tons and the latter of 1,428,427.53 tons in 2016 (MoFA, 2016). With a total area of 70,384 km², the Northern region is the largest region in terms of land mass. The total population was estimated at 2,479,461 in 2010 with 50.4% females (1,249,574) and 49.6% males (1,229,887). More than 75% of the economically active population depends on agriculture for survival. The region has a relatively dry climate with a single rainy season beginning in May and ending in October, with rainfall distribution varying between 750 mm and 1,050 mm. With the harsh climate experienced in this region, cassava serves as a resilient and secure food commodity for the populace (Ghana Statistical Service, 2013).

The Eastern region covers a total land mass of 19,323 km² with an estimated population of 2,633,154 in 2010 based on the Population and Housing Census (PHC). 51% are females (1,342,615) and 49% males (1,290,539). The vegetation is tropical with two seasons; dry and wet. The climate and soils support a variety of cash and food crops including cocoa, kola, cassava, rice and oil palm, which account for 70–85% of agricultural output. Cassava dominates in terms of cropped area and total production (MoFA, 2011) and the region ranks 1^st in cassava production (MoFA, 2016). In this region, 53% are employed in the agriculture sector (Ghana Statistical Service, 2013). The study districts fall within the region’s major cassava districts (as shown on the Ministry of Food & Agriculture (MoFA) website and in the map in Figure 1).

Figure 1. District Map of Northern and Eastern Regions of Ghana with study areas in the circle.

Maps were adapted from original versions available from Wikimedia under a CC BY-SA 4.0 license (Northern region map – Rwhaun, 2018, https://commons.wikimedia.org/wiki/File:Districts_of_the_Northern_Region_(2018).png; Eastern region map – Macabe5387, 2017, https://commons.wikimedia.org/wiki/File:Districts_of_the_Eastern_Region_(2012).svg).

Sample size and sampling technique

A two-stage sampling technique was used to select the study districts and sample. The two districts, Yendi Municipal in the Guinea savannah region and Lower Manya Krobo in the forest transition region were purposively selected due to the importance of cassava in these districts and as part of the disease surveillance survey districts. A simple random sampling technique was used to select 10 cassava farmers each from five randomly selected communities within each district, resulting in a total of 100 respondents. Preliminary engagements were made with the district directors of agriculture to inform them of the purpose of the study but not with the respondents. The farmers were drawn from a list obtained from the Root and Tuber Improvement and Marketing Programme (RTIMP) desk officer at the district department of agriculture. The list was inputted into Excel and random numbers were generated to assign to respondents using the RAND function. The INDEX and RANK functions in excel were used to select the 100 participants. The participants selected for a face-to-face interview were informed through district directors of agriculture. Due to incompleteness of some of the questionnaires, 94 valid responses were retrieved, representing a 94% response rate.

Data collection

Data collected was mainly primary and was accomplished in 2017. Data was collected using a semi-structured interview guide (Nsiah Frimpong et al., 2020b) administered to the sampled respondents by interdisciplinary researchers (social scientist: BNF, pathologists: AO, ZAK, LA, JP and breeders: RP, JMA) with MSc and PhD qualifications and trained enumerators with Diplomas and BSc degrees in Agriculture (both males and females) from the districts’ departments of agriculture. The interviews took place at the homes of the respondents, either in the presence of other family members or individually and lasted for a maximum of 45 minutes. Responses were either ticked or transcribed as and where appropriate. A mixed research questionnaire was adopted with both closed and open-ended questions, which elicited both qualitative and quantitative research variables. This integration was deemed appropriate as it provided an in-depth understanding of the questions being posed in the survey. The survey instrument was designed in English and translated to the local language by the trained enumerators during field administration. Prior to field data collection, the research team and trained enumerators pretested the survey instrument to ensure consistency and logical flow of questions.

Individual cassava farmers from selected households were chosen as the unit of analysis and data collected covered their demographics, farm level characteristics, knowledge and experiences with cassava pests and diseases, management practices employed and their effectiveness, sources of information on pests and diseases, knowledge on improved varieties and preferred attributes, willingness to cultivate improved varieties and knowledge of the new devastating cassava disease caused by ‘cassava brown streak virus’. There were several triangulations to check and confirm farmers’ responses on experiences with practices of managing cassava pests and diseases, variety names, characteristics and categorization. Triangulation involves using different methods to collect data on same topic to ensure validity of results. There are four types of triangulation; data, investigator, theory and methodological (DźwigoŁ, 2018) and the study used mostly investigator triangulation. For instance, to confirm respondent’s knowledge on pests and diseases, he/she was asked to name and describe the associated symptoms. Within the multi-disciplinary team, researchers relied on the viewpoints of the subject matter experts to validate pest and disease names/descriptions and variety names, attributes and categorization into traditional and improved during analysis.

At the beginning of the interview, enumerators explained the purpose of the study and content of the questionnaire to the respondents and assured them of the confidentiality of the data collected. Verbal consent to participate in the study was sought from all respondents explicitly in their mother tongue before proceeding to start the interview. Participants could freely terminate the interview at any time if they felt uncomfortable with some of the questions. During the interview, questions regarding pest and disease encounter and management were repeated and further explained to some of the farmers for clarity and to elicit correct responses. Farmers’ responses were constantly read to them for confirmation. Once the responses were read to them, everything was clarified in the field. However, farmers’ contact details were taken in order to make a follow up phone call to clarify issues should the need arise during analysis, but no such call back was done or feedback on findings given.

Data analysis

Data collected was inputted using SPSS 20 software and were analyzed using descriptive and inferential statistics. The descriptive statistics included frequencies, cross tabulations, charts, means, minimums, maximums and standard deviations. The study employed both parametric and non-parametric tests, which included the t-test and Chi-square test (Fisher’s exact). The Chi-square test of independence was first run to test whether any relationship existed between the agro-ecologies and the management measures employed; however, because the expected cell values were less than five, which violated the Chi-square assumption, researchers resorted to the Fisher’s exact test, which qualified for that purpose. Farmers’ attitudes were linked to behavioral control methods such as variety cultivated, awareness of and willingness to cultivate improved varieties, cropping system and choice of planting dates. Practices referred to the actions taken or coping strategies used when they encountered issues with pests and diseases. The effectiveness level of the actions was evaluated on a three-point scale; 1 = “not effective”, 2 = “moderately effective” and 3 = “very effective”. Inter-regional comparisons were carried out to ascertain heterogeneity in the quantitative variables measured using the t-test. These helped to draw conclusions about the population from which the sample was selected. Open-ended questions were coded by reading through the transcripts severally and themes built for textual analysis. Coding was manually done by two of the interviewers (BNF and RP) and perspective of other experts in the team sought to aid in the organization of responses. For instance, codes like “mounds”, “slanting”, “cut stem/sticks” were developed for how the planting of cassava is done and “curled leaves”, “rotten roots”, “yellowing”, “chew leaves”, “white pests”, were developed for knowledge on and experience with pests and disease. At the point where researchers realized they had no more codes or new information, the codes were developed into themes. The textual analysis involved examining the text and analyzing the sentence structure by combining both content and narrative analysis for the open-ended responses. While the content analysis dealt with categorizing responses under and into themes, narrative analysis looked at evaluating farmers’ descriptions for pests, diseases, planting methods, varietal characteristics.

Respondent demographics

A total of 94 respondents were selected for the study; approximately 53% from the Guinea savannah agro-ecology and 47% from the forest transition zone (Table 1). Table 1 shows that 77.7% of the respondents were males and 22.3% females. The majority of the respondents (69.1%) have had some formal education taken at different levels (basic, secondary and tertiary), with many of them clustering around the basic level. A higher percent (91.5%) were married, which presupposes the support farmers could get from their partners. Again, the majority (74.5%) of the respondents were household heads and 77.7% were members of an agricultural organization. Descriptively, the analysis revealed some variations across locations in relation to the categorical variables; level of education, respondents being head of household and membership of agricultural organization, as presented in Table 1.

Test of equality in means between respondents from the two regions

The independent sample t-test was conducted to test for equality in the means of characteristics of respondents in the two study districts. The null hypothesis stated that the means of quantitative variables of respondents from the two agro-ecologies are the same. The F-statistic and p-values of the Levene’s test of equality of variance showed that the test assumes equal variance for all the variables except years in school, experience in cassava farming and household size, which had p<0.05. Details of the Levene’s test results are presented as Extended data (Nsiah Frimpong et al., 2020b). The results showed some differences across the agro-ecologies. Differences in age, years of cultivating cassava, years in school and household size were significant at 1%, while farming experience and proportion of farm damaged by diseases were significant at 5% and the number of adult males actively involved in farming was significant at 10% (Table 2).

Farmers knowledge of cassava pests and diseases

In ascertaining farmers’ knowledge, three aspects were assessed; experience with any pests and or diseases, name of pests and diseases and description of pests and diseases encountered. Results in Table 3 show the regional trend, indicating that majority of the farmers, 74% and 70.5% from Guinea Savannah and Eastern regions, respectively, experienced pests on their farms, while 72.0% and 70.4% farmers, respectively, experienced diseases, with the last period of encounter being quite recent (2014–2017). It was found that 61.8% of the farmers perceived whiteflies (Bemisia tabaci Gennadius), grasshoppers (Zonocerus variegatus), aphids (Aphis gossypii Glover), mealybugs (Phenacoccus manihoti), termites (Isoptera), and grasscutters (Thryonomys swinderianus) as the most common and damaging pests to their crops. The remaining 38.2% could not give specific names but generally described them as “whitish insects”, “flies” or with symptoms such as “create holes or chew leaves” (Questionnaire ID 11 & 12), “web/wax-like patterns on the leaves” (Questionnaire ID 17), “pests clustering beneath the leaves and causing wrinkles” (Questionnaire ID 41), “yellowing of leaves and leaves dropping” (Nsiah Frimpong et al., 2020a). More farmers in the Guinea savannah region were able to give specific names of the pests than those from the Forest transition region.

The predominant diseases known by the farmers were the cassava mosaic disease (22.4%) and root rot (19.4%), which are viral and fungal diseases, respectively. The majority (58.2%) of the farmers could not give specific names but used descriptions such as “black/brown leaf spots”, “curled leaves”, “mottled leaves”, “yellowing of leaves”, “leaves dropping and stunted growth” to demonstrate their knowledge of diseases. Out of the 67 farmers who experienced diseases, only 28 farmers, representing less than half (41.8%), could explicitly give the names of the diseases that attacked their farm. Cassava mosaic disease was experienced more in the North and root rot was experienced more in the Eastern region. The mean damage caused by pests and diseases was 22.84% and 21.02%, respectively, and ranged from 0–100%. Results on the distribution and peak of percentage damage by pests and diseases data showed that pest damage data is slightly skewed and flat, while disease damage data showed normal distribution and flatness; hence not many issues with skewness and kurtosis in the data (Table 5). Out of the 67 respondents who experienced cassava diseases, 74.6% reported single infections and 25.4% observed multiple symptoms. Farmers’ responses on knowledge and experiences of cassava diseases were validated by triangulation using seven statements that described the common symptoms of bacterial, fungal and viral cassava diseases (Table 4). With these prompts, five more respondents confirmed the observance of symptoms on their farm. Multiple symptoms were observed by 21.3% of the farmers; hence the probability of multiple disease incidence. In addition to the cassava mosaic disease and root rot reported by farmers, the symptoms statements revealed that brown leaf spot (35.1%) and white leaf spot (22.3%) diseases could also be common in the districts.

Farmers’ attitudes and practices in cassava pests and diseases management

The majority of farmers (74.5%) intercropped cassava with cereals (maize, sorghum and millet), other roots and tubers (yam and cocoyam), legumes (groundnut, cowpea, pigeon pea and soybean) and vegetables (pepper, okro, tomatoes and garden eggs). Maize (67.1%) and yam (42.9%) were the most common (Table 6). The variety names were verified by technical experts in the team and classified into local or improved varieties; 64% of farmers in the Guinea savannah region cultivated only landraces compared with 22.7% in the forest transition region. Solely improved varieties were cultivated by 40.9% farmers in the forest transition region and 12% in the Guinea savannah region (Figure 2). Table 7 shows the list of local and improved varieties identified in the area. Though the majority (86.2%) of the farmers were aware of improved varieties (Table 8), the local landraces dominated (Figure 2).

Figure 2. Graph of variety types cultivated by farmers.

Most farmers planted cassava around March/April/ May. More than 90% of the farmers planted the cassava by cutting the sticks into pieces and slanting it on mounds. The remainder planted on mounds but in a vertical position. Harvesting was normally done from 6–24 months after planting depending on the variety cultivated.

All (100%) farmers were willing to cultivate improved varieties, locally referred to as “agric bankye”, provided they could access planting materials (Table 9). Qualitatively, the top five desired traits/characteristics included high yielding (big tubers), early maturing (6 months to 1 year), multiple uses/by-products (“gari”, “fufu” and starch), disease resistance and tolerance to harsh weather conditions. There was slight variation in the traits preferred by male and female farmers. Most of the males preferred high yielding (big tubers) and early maturing varieties, but the females preferred early maturing, multiple food uses and disease resistant varieties.

Farmers’ behavior in coping with pests and diseases revealed four options. Similarities existed in both farmers’ immediate and long-term responses. Table 9 indicates that 45.6% immediately contacted an agricultural extension officer, 19.1% employed field sanitation/cultural practices, 26.5% took no action and 8.8% applied pesticides. As long-term measures, 33.8% took no further action, 29.4% sprayed chemicals, 30.8% observed cultural practices with some extension advice and 5.9% invited extension officers to the scene (Table 9). Similar management strategies were employed for diseases. Questions on the effectiveness of pest management strategies indicated that 17.6% believed the practices resorted to were very effective, 36.8% believed they were moderately effective and 45.6% indicated they were ineffective. With a mean score of 2.3 and standard deviation of 0.750, the overall level of effectiveness of the pest management measures was perceived as being moderately effective. With disease management, 20.9% perceived the practices as being very effective, 38.8% as moderately effective and 40.3% as ineffective, with an overall mean score of 1.8 (Table 10).

Testing the relationship between agro-ecologies and farmers’ pest and disease management practices

A more robust test such as the Fisher’s exact test was appropriate to correct for the small frequency values in each cell and to satisfy the assumption of carrying out a Pearson Chi-Square test. The results in Table 11 and Table 12 show that since the P-values (0.357) and (0.697) were greater than 0.05 for pest and disease management practices, we do not reject the null hypothesis of “no association between agro-ecology and choice of practices”. This means that the two categorical variables are not related and thus, agro-ecologies are independent of the choice of pest and disease management practices.

Sources of information on cassava pests and diseases

Figure 3 indicates that the majority of farmers (94.7%) depended on the extension directorate under MoFA for information on pests and diseases, with the least amount of information sourced from research institutions. Other sources included radio (44.7%), fellow farmers (41.5), television (29.8%) and non-governmental organizations (11.7%).

Figure 3. Information channels on crop pests and diseases.

NGO, non-governmental organization.

Underlying data

Harvard Dataverse: Farmers knowledge of cassava pests and diseases. https://doi.org/10.7910/DVN/27YW9V (Nsiah Frimpong et al., 2020a)

Extended data

Harvard Dataverse: WAVE Knowledge Paper_Extended data. https://doi.org/10.7910/DVN/AJZDFN (Nsiah Frimpong et al., 2020b)

This project contains the following extended data:

Data are available under the terms of the Creative Commons Zero “No rights reserved” data waiver (CC0 1.0 Public domain dedication).