Given Hapunda

116 discussions in the questionnaire. When that happens, caution should be taken to plan ahead on how data will be analysed. The choice in terms of the type or number of questions in a questionnaire can move from being simple to complex. This can determine the quality and quantity of data collected, which will inevitably influence whether the research questions will be answered or not. As a way of illustration, a questionnaire can be a very good source of data if you need to find out if small-scale farmers who participated in conservation farming training are utilising the information. In this case, the participants would be the actual farmers who participated in the training but a decision should be made whether a sample or all of them will be interviewed. The sampling decision may be influenced by the availability of human and financial resources, time and actual total population because the smaller the population, the more likely that all of them will participate. Unlike focus group discussions or face-to-face interviews; questionnaires can be administered through telephone, email or post. The tips below provide more guidance on how to go about developing and administering the questionnaires.  Define the purpose and objectives of your study in order to collect correct data from your evaluation questions.  Develop clear and simple questions while avoiding the use of jargon or abbreviations.  Questions should be put in a logical manner to bring about better flow of responses e.g., from general to specific or from less sensitive to more sensitive questions.  Demographic questions focused on age, sex, marital status, education, occupation, etc., are important for establishing casual relationships. However, it is important to only use those relevant for a particular study.  Decide on your sample size that is adequate enough to be representative of the study population. Sometimes this can be a portion or the whole study population.  A pilot study should be conducted to test whether: the questions are clear and capturing the right information, the topic or some of the questions are too sensitive, the time it takes to interview one respondent is good (too long may be boring). The pilot may also act as a gauge for research skills, especially when using research assistants.  The interviewer should have good reading and writing skills, good interviewing skills and have an idea on how data will be analysed as that will help in how questionnaires will be designed.  If a questionnaire is self-administered, the researcher should ensure that the respondent is able to read and write (Burgess, 2001; CDC, 2008c).