A survey was conducted in 2010 and the first quarter of 2011 in three counties of Nakuru (Njoro and Kuresoi), Nyandarua (Nyandarua South, Nyandarua West and Nyandarua Central) and Meru (Meru Central and Buuri). The three counties are located in in the Rift Valley, Central and Eastern regions of Kenya, respectively. Central region is the leading producer of potatoes in Kenya followed by Rift Valley and Eastern Region. The study counties are the main potato growing areas in their respective regions and together account for approximately 95% of total potato production in Kenya .
The areas studied are all in high altitude- (between 1,400 and 2,700 meters above sea level) and high-rainfall zones, experiencing mean annual rainfalls of 1,000 mm or greater. Nyandarua County has temperatures ranging from a minimum of 2°C to a maximum of 25°C. The rainfall ranges between 700 and 1,500 mm per annum . In Meru County, annual temperatures range from a minimum of 16°C to a maximum of 23°C and rainfall from 500 to 2,600 mm. Temperatures in Nakuru County range from a minimum of 12°C to a maximum of 26°C per year with rainfall ranging from 1,800 to 2,000 mm. Maximum temperatures across all study counties, therefore, are sufficiently temperate, as are minimum temperatures - with the exception of Nyandaura. The highest variability in rainfall is recorded in Meru, where some areas receive less than 1,000 mm per year, which may explain the high usage of irrigation in the county. The predominant soil type is volcanic in Nyandarua and Meru but some parts of Nyandarua have red clay soil. Nakuru mainly has loamy soils.
Since no complete household survey has been carried out in in the last 5 years, we used data from the Kenya Integrated Household Budget Survey (KIHBS) 2005/2006  to estimate the number of households producing potatoes. The total number of such households was 790,752, of which virtually all (97%) were located in the main Central, Rift Valley and Eastern producing-regions.
KIHBS data also provided estimates of the share of potato-growing households in each target county. In Nyandarua, 97% of farmers grew potatoes compared to 34% in Nakuru and 31% in Meru. Together, the three counties accounted for about 33% of all potato growing households in Kenya.
Relevant KIHBS data is aggregated at the level of households. Equally, the respondents targeted in our study were the heads of households. Interviews captured the demographic characteristics of the household head. The household is defined as a place where members 'eat from the same pot’. In the regions studied, this was also synonymous with housing units since independent households in these rural areas do not share the same house.
To be able to generate a random sample from the three regions, we used administrative district-level information gathered through a 2009/2010 enumeration of potato farmers by the Ministry of Agriculture. For some parts of Nakuru (Njoro and Kuresoi), data were incomplete, requiring us to employ a stage-wise stratified sampling approach, estimating the number of farmers in a village and selecting one at a constant interval.
The required sample size (n) was 381 as per the formula below. We however, targeted 419 farmers, assuming a 10% non-response rate and ended up with 402 completed questionnaires.
n = required sample size
t = confidence level at 95% (standard value of 1.96)
p = estimated proportion of farmers growing potatoes - used 55% average as per occurrence in KIHBS 
m = margin of error at 5% (standard value of 0.05)
Using the KIHBS [8
] data for the farmers involved in potato production, the average occurrence was 54% as below:
The survey questionnaire was designed to collect data that could be used to generate additional variables. To increase the reliability of self-reported data, the questions asked were simple and information sought easy to recall. For instance, on yields, farmers were asked about the portion of their land they had dedicated to potatoes in the last season and the production thereof. Total production was divided by the area to generate yield data. Since the study sought to examine production in general, data on the varieties grown were not collected. The specific fertilizers and fungicides used were recorded but the ranges of fungicides were too wide to be analysed meaningfully. Several types of fertilizers were reportedly used but most farmers were unable to recall the specific kind used. As they put it, they simply follow sellers’ advice on the type to purchase. The analysis, therefore, ignores fertilizer distinctions. For irrigation, the data collected were on installed irrigation facilities as opposed to actual use. It was assumed that those who had installed facilities actually used them.
The analysis used the Chi-square and Fisher’s test, regression and logistic regression, where the dependent variable was dichotomous. Stata/SE 10.1 was used for the analysis.