This paper outlines how I would carry out a survey of 1000 members of the general population in the UK in order to explore their experiences of and attitudes towards the police. Of key importance, I explain the modes of data collection I would use. The importance of the mode of data collection is highlighted. It is also argued that population characteristics are most crucial in determining what mode of data collection to use. The sampling methods to be used are outlined and justification for their use as opposed to others is also well stated.
Survey research is often used to gather information from a homogenous population that is widely spread. In research, the approach taken by the researcher determines the response rate as well as the validity of the responses. One has to weigh the advantages against the disadvantages of using given methods or research approaches. Key to deciding the kind of data collection mode to use or survey type to embrace are the disadvantages and advantages that each method provides in relation to a number of measures or issues. These issues range from privacy or confidentiality needs, amount of visual presentations possible when using the method, possibility, and room for open-ended questions when using the mode, amount of reading or writing needed, possibility of judging the quality of response, level of response rates, personal touch or contact, cost-effectiveness, sample distributions, and time considerations. Often, a number of research methods or modes are infused so as to get a tailored solution that fits the context in question.
Target Population characteristics are very important when designing a survey approach (Sapsford, 7). Some general population characteristic-related considerations revolve around accessibility. There are challenges if the listing of the general population is not clear. There are populations that are not distinct e.g. the general population in the UK. Such populations force challenges in terms of what characteristics to base on in sampling. Literacy in a population is also a critical issue. Literacy does not remain at the level of being able to read and write. One has to be clear or consider the understanding of the technical vocabulary likely to be used. Another consideration is the language used by the target population; the general population in the UK is multilingual thus need to put language issues into consideration in designing the research approach. There is some population that is sensitive i.e. they are suspicion towards and form of investigations. Such special groups require a special approach so as to win their confidence. The geographical dispersion of the population matters a lot; if the population is widely spread, personal administered interviews or questionnaires are not feasible (De Vaus, 122).
Two broad data collection methods are employed in gathering information. Either one uses a questionnaire, interviews, or both. For the purposes of the survey on the experience of and attitudes towards police in the UK, I would prefer to use both questionnaires and interviews. Questionnaires are to help to cover the whole country more cost-effectively, while interviews are to help give the research a personal touch.
A questionnaire provides a set of questions to which a respondent provides answers. Questionnaires are flexible tools that researchers can use (Fowler, 151). By using different kinds of questions, the researcher can control the respondent while at the same time allowing him or her to voice her independent opinion on the issue under discussion. The researcher’s control concern should only be focused on helping the respondent deal with the issue under investigation. The research should not apply leading questions of any sort as that definitely lead to a bias in the research findings(Fowler, 153).
Some question structures commonly employed are either dichotomous, scale-based, contingency-based, or open-ended (Czaja & Blair, 85). Dichotomous questions often require respondents to choose between two options e.g. Yes or No. Scale-based questions often provide a scale or require a respondent to rank given options. Contingency questions are often aimed at gauging respondent’s suitability for answering the following question. Open-ended questions are aimed at getting the respondent to freely comment on the issue at hand. The research objectives or the different aspects of the general objective should guide the kind of questions to be asked. The questions should be sequenced from trivial to more serious or sensitive ones. Question need have interconnectivity and transitory connectors from different issues or topics into others (Czaja & Blair, 85).
Anticipated data is crucial in determining what method or mode of collection is proper to collecting it. Data can either be qualitative or quantitative. Qualitative data requires qualitative coding methods while quantitative data may require methods that allow for numeric presentation. To survey the experiences and attitudes of a population towards a police force, both qualitative and quantitative data and hence methods would have to be used. Both interviews and questionnaires support the use of either qualitative or quantitative approaches. Qualitative approaches help towards exploring aspects inductively while quantitative approaches help toward deductive appreciation thus confirmation of given aspects (Seale, 181). These two approaches would complement each other making the survey more complete and holistic.
There are myriad ways of administering questionnaires just as there are different means and mediums through which interviews can be conducted (De Vaus, 124). The questions are often typed out on papers or electronically presented. Questionnaires utilize both open ended and closed questions to investigate issues and establish trends. Questionnaires are developed with the target audience and research objective in mind. Questionnaires can be administered personally, through snail mail or via electronic media. Questionnaires can be administered to individuals or to groups. Researchers can target different individuals or bring a group together so that collectively they respond to research questions. Group administering of questionnaires is only possible where it is easy to assemble such a group e.g. in a company, school or other institutions (Seale, 194). Another way of administering questionnaires is targeting households. For the purposes of this survey, I would mail questionnaires due the physical distribution of respondents. Mailing questionnaires means that respondent does not receive the personal touch or contact that a researcher is able to provide if he or she delivers the questionnaire (De Vaus, 127). However, it enables the researcher to reach respondents that are far off. Giving questionnaires to individuals is good enough as opposed to group filling of questionnaire as it allows for privacy and confidentiality thus respondents being more open as concerns sensitive issues.
Questionnaires are administered depending on geographical distribution of respondents plus other target population characteristics, time factor, technology averseness and availability, level of confidentiality required and levels of literacy. The geographical or spatial distribution of respondents determines whether researchers can cost effectively reach the respondents for personal contact or not. Population characteristics are varied and are always contextual. For example, a working class population would be more responsive to mail questionnaires than personally administered ones. The amount of time at the disposal of the researcher also determines whether to uses such avenues like electronic administering or send snail mail to respondents. However, in as much as a researcher my want to administer a web survey, if the respondents are not internet savvy or can not be easily reached via internet, then electronic administering of questionnaires is not a possibility. The level of confidentiality required also dictate choice of method through which to collect data. Mailed questionnaires are likely to be answered more honestly than personally administered questionnaires if issue is too sensitive. Matters to do with the police e.g. experience of and attitude towards can not be discussed openly by the general public. Respondents may fear being targeted by security personnel thus not sharing their authentic experience of and attitude towards the police.
Interviews is administering questions but unlike questionnaires where the respondent is handed recorded questions to which he or she responds, interview involves direct asking of questions by the interviewer (researcher) to the interviewee (respondent). Interviews can be done directly or personally or via some indirect medium. In personal or direct interviews, a researcher meets directly with the respondent. In indirect interviews, one may use such tools as telephone. Personal interviews have a great advantage over telephone interviews because they are not limited by costs and the interviewees do not feel bugged. In personal interviews, one is able to have higher control and gathers more information through observing physical cues than in telephone interviews.
Interviews can also be on individual or group based. Individual based interviews are one on one interview. Group based interviews are when a researcher or researchers ask questions to individuals in a group (Seale, 193). Just as it is the case in group administered questionnaires, group based interview suffer lack of privacy and confidentiality is compromised. However, the advantage of group interview is the possibility of interaction between respondents that clarifies issues further. Group based issues are best for issues affecting a general group that are not very sensitive. For example, research on general issues such as the health policy could easily be group based. Researches in a company on say competition could be group based. However, whenever sensitive information that would require privacy or confidentiality is needed, group based administering of questionnaires or interviews is not acceptable.
As indicated, to properly carry out a survey on the general public’s experience of and attitudes towards the police; I would use both questionnaires and interviews. I would mail the questionnaires so as to enable respondent’s privacy and confidentiality. This would enable the respondents to open up and probably share more deeply. I would also do some interviews, particularly personal interviews. Based on sampling that allows for proper representation of all groups in the UK, I would do personal interviews that enables for probing and clarifying of given issues. Telephone interviews would do but given the issue at hand concerns the police, direct personal touch with interviewees would help in winning their confidence better than telephone contact.
Personal interviews and questionnaires administered to individuals serve best when detailed personal information is needed. Interviews are better at deciphering details than questionnaires because the questions can be clarified, the interviewer can crosscheck with the interviewee and observation or eye contact adds more content to the general understanding of what the interviewee is saying. However, to reach many respondents, questionnaires are better because the physical presence of the researcher is not necessary. Interviews are also likely to receive lower reception (low response rate) because an interviewee has to set aside time and be available for the interview. Questionnaires can always be filled at the respondents own time, wherever.
Cost management is often of great concern in a research work (Bryman, 130). The research method or survey mode used has to be within budget. Costs are not just about how much cash is spent but also the material, personnel and facilities necessary for given kinds of data collection modes to be effected. A cost benefit analysis is always necessary before settling down on any mode of data collection. Personal interviews are cost intensive because one has to use money to get to the respondents, book meeting room and somehow entertain the respondent in a way so that he or she is comfortable in the interview. To reach many people, the research would have to spend money on training interviewers. For phone interviews, one would have to consider the cost of access to well installed and maintained phone services that would not cause mishaps or interruptions to the interview process. It is for this reason that questionnaires are used alongside personal interviews. A great percentage of the population would be reached by mail while a special small percentage would be contacted for personal interviews.
To cut cost, often sampling plays a crucial role (Bryman, 137). Sampling helps the researcher to work in a fairly orderly way with greater efficiency and effectiveness. Sampling is the identifying a small number out of the general target population with which to make contact. Sampling helps in making the research easy to execute. The choice of a sampling criterion is dependent on how much information is obtainable about given subgroups or individuals in the target population. Sampling will also depend on availability of the target population and their general characteristics.
Apart from the population characteristics, the research objectives determine what kind of respondents to target (Fowler, 18). If the research objective is to determine attitudes and experience in the general public, then representation becomes the issue in designing the survey. If it were targeting only a specific population in the public, then the consideration would turn to other issue like the expected capacity of respondents in relation to demands of the research objective. Do they have the necessary prerequisite knowledge as to be able to authoritatively answer the research questions? The other consideration about information to be collected is whether some form of referring would be needed in answering. For a survey into attitude and experiences, this may not be an issue but in some other research say into monthly expenditures, some referring into records may be necessary for exactitude.
If information needed is common or can be found from any member of the population, random sampling is feasible (Sapsford, 53). With random sampling, every member in the target population has an equal probability of being contacted. The UK population is not homogenous. There are groups based on race, socio-economic class and general neighborhoods. Treating the whole population (which is one or homogenous in the sense that we are all under one rule) equally would lead to a bias in the findings in favor of majority groups. There are populations in different townships and cities that have some forms of subcultures of their own. The diversity in the population calls for stratified sampling. Stratified sampling is good for target populations that have hierarchy, classes, subgroups or strata of sorts. Therefore, the general population would have to be sampled along lines of discernible strata or sub groups. However to the purposes of getting out 1000 out of the whole population, random sampling would have to be used within the different strata’s. To make the work easier, some form of clustering would have to be done. If homogenous groups exist over some geographical distance, cluster sampling is used. In this survey, clusters would consist of regions in the UK. Once the whole of UK has been divided into regional based clusters, stratified sampling would be done in the regions and random sampling applied in the different strata.
This would generally mean a lot being done in preparation for the survey. A lot of information would have to be accessed on the general population as to be able to identify the potential viable strata in the different regional clusters. This approach would work best due to the complexity and extend of the general public, which is our target population. If the population had been purely homogenous, a simple random sampling would be welcome. If the population had been homogenous in all clusters, we would not have to stratify in the clusters (Punch, 5).
From the fore mentioned reasons, understanding the characteristics of the target population is very important and should inform the sampling procedure. A sample ought to be as representative as possible of the target population (Czaja & Blair, 168); it should carry all characteristics or peculiarities of the target population. Sampling is supposed to make researching easy; it should also improve on the response rate because the population has been reduced to a manageable sample. If sampling is not done properly, the research validity is lost and findings are likely to be biased if generalized to the whole population (Sapsford, 55). For example, if a stratified population that has certain minorities and majorities is sampled on random criteria, the majority group is likely to be more represented than the minority group. For a proper balance, stratified sampling ensures a proper mix.
Works Cited
Bryman Alan. Social Research Methods. 3rd Ed, London: Oxford University Press, 2008
Czaja Ronald & Blair Johnny. Designing surveys: a guide to decisions and procedures. California: Thousands Oaks. 2005
De Vaus D. A. Surveys in Social Research. 5th Ed. USA: Routledge, 2002
Fowler J. Floyd. Improving Survey Questions: Design and Evaluation. California: Sage, 1995
Punch F. Keith. Survey Research: The Basics, Essential Resources For Social Research, London: Sage, 2003
Sapsford Roger. Survey Research. London: Sage, 1999
Seale Clive. Researching Society and Culture. 2nd Ed. London: Sage, 2004