Be sure to employ scientific measures in selecting the sample. Purposive and convenientsampling may not be the best type but some form of systematic random sampling maybe more appropriate.11.
The best way to determine sample size is by use of a formula which takes intoconsideration the standard error, sampling error, confidence intervals etc.12.
Consider the ethics of the research. For example is there any potential harm to therespondents?13.
Determine the implications of this research for the practice of the discipline or field of study such as sociology, psychology or economics. (policy prescriptions too)14.
If you intend to construct your own data collection instrument you can conduct a smallnumber of unstructured interviews and identify the main themes from them. This willadd to the level of reliability of the scales used for measuring the concepts. (See Foddy1993 for more details of structured interview construction)15.
Construct the questionnaire/structured interview or data collection instrument byoperationalizing- this ensures the main topic is deconstructed a number of related sub-themes, topics or issues that will then be used to frame the questions. Remember thatinterviewing beforehand can provide a number of themes and subthemes that may beuseful when operationalizing the topic.16.
When framing the questions the level of measurement should be taken intoconsideration. For instance you make need to use Likert scale type questions formeasuring correlations.17.
Pilot test the data collection instrument by using a smaller sample of respondentssimilar to those being targeted for the actual survey before carrying it out. This will tellwhether or not the wording is unambiguous and the questions appropriate for the typeof data to be used.18.
Never ask for demographics such as age, religion, income and other personal data if they will play no part in the analysis of findings. Avoid wasting data and wastingrespondents time.19.
Clarify the survey administration beforehand. Ask yourself questions such as: will theyinstrument be administered via telephone, house-to-house, mail, internet or someother appropriate means? In this way you will be able to gauge the return rate sincethey affect the quantity and quality of data being collected.20.
Determine if any funding is necessary for the research and who or what agency will beproviding it. Funding obtained from specific agencies may result in changes being madeto the data collect instrument particularly if the funder wishes to have a particularobjective realized.21.
Consult your academic supervisor/advisor any additional measure you may need to takebefore collecting the data.Data Collection Phase1.
Administer the survey using the predetermined means.2.
Ensure that participants are sufficiently apprised of the purpose of the research and arewilling subjects.