2.2. Sampling design

The study used a descriptive research design to meet all the research objectives. Secondary data was collected by reviewing legislative documents, forestry-related publications, and project reports to provide an overview of LCF, the communities’ socio-economic status, and how the forest has been managed compared to other forests under similar management regimes in other countries. The sample size was determined Sample size of study was calculated using the formula Z²Pq/d² (Mugenda and Mugenda 2003).

where:

n -: the desired sample size.

Z -: was the corresponding standard score with the probability of error at 0.05 and a confidence level of 95%, which is 1.96

p -: was the occurrence level of the phenomenon under study and is equal to 0.5 where the occurrence level is not known

d -: is the selected margin of error of the study corresponding with 95% confidence level in this case 0.05.

Substituting for the values:

The households were randomly selected from a list of Manyattas (which refers to a group of households living mostly in huts within a common fence but not necessarily related, mainly observed among the Maasai community) in each village provided by the village heads. One Manyatta was randomly selected in each village, and a questionnaire was administered to thirty household heads within each of the selected Manyattas. The manyattas were located in Entasikira, Olorte, Ilkerin, Empupurtia, Ilmaral, and Nkopon zones in the Loita study site. The forest was stratified into six zones based on the discussion held with key informants to address the forest size and variability to represent the forest's social, ecological, and economic situation and the community settlements inside and outside the forest. Stratification also ensured a fair representation and captured a good perception of the forest conservation status and community livelihood.

Snowballing (Mugenda and Mugenda, 1999) was also done whereby a policy-maker (who could be a community leader, a government and NGO officer, or other organization staff) would propose another officer or community who was considered experienced in answering policy questions. Key Informant Interview respondents were selected through purposive sampling (Mugenda and Mugenda, 1999) from the community, relevant government. Additionally, Seven Focus Group Discussion (FGD) were held before and after administering the questionnaires. The FGD meeting were attended by 27 participants representing households selected purposively with the help of KIs and their characteristics defined to the proposal, size, composition, and interview procedures as defined by Freitas et al. (1998); Folch-Lyon & Trost (1981) and Boateng (2012). The FGD participating members representation included; 12 youth, 6 women and 6 men and 3 elders.