9. Surveys
Key points
Get Clinical Tree app for offline access
• Questionnaires enable a large amount of data to be collected quickly and cheaply. They have the advantage of being familiar to a majority of the population, and compared to other forms of data collection, are a non-threatening medium for both the researcher and respondent.
• The response rate is variable. Where it falls under 50%, it is difficult to be certain that the responses received are representative of the sample as a whole.
• Questionnaires have been used so often in the past that people are now less likely to return them. Is this the most appropriate method here? In particular, the ethical issue of harm should be considered where the questions could produce emotional upset, regret, anxiety or confusion.
• Designing a questionnaire involves three elements: the invitation, the instructions, and the body of the questionnaire. There are clear guidelines that should be followed in the construction of questions. The importance of avoiding bias and ensuring reliability and validity must be stressed.
• The basic premise of questionnaire design is that the respondent can read and write, and is fluent in the English language. For several reasons, there is a proportion of the population who will always be excluded in studies using questionnaires. Similarly, with electronic versions, not everyone has access to a computer.
• A pilot study, or the use of a previously used questionnaire, is a good indicator of rigour in the use of questionnaires.
Gathering data for research is exciting. However, its success depends largely on the method used. This must be appropriate to the research question, and should be a reasonable choice for the sample group. In this chapter we examine the use of the survey, which has traditionally been a popular method of collecting research information. A survey is defined as a method of gathering data by directly asking respondents for information in the form of either a questionnaire or an interview.
This chapter will examine some of the principles involved in the use of surveys, and examine the advantages and disadvantages of one method; the use of questionnaires. The principles of questionnaire design will be illustrated at the end of the chapter. The following chapter will examine the use of interviews.
The survey
The survey is a quantitative research design that collects written or verbal information from large numbers of individuals, through either the use of questionnaires or interviews, in an attempt to describe a situation in the form of numbers. It also provides a way of identifying correlation or patterns amongst some of the variables included in the survey. This allows the health researcher to have a better understanding of patterns of behaviour or reactions to the provision of health services that will help in the evidence-based planning and delivery of care. Houser (2008) claims the survey has become one of the most common method of data collection in nursing research, and midwifery researchers have also found this method of collecting data ideally suits many of the questions they wish to answer.
Why are surveys so popular? One reason is that they are very user friendly. They are less intimidating for both the researcher and those participating in comparison to some of the other forms of data collection. They are also a very economic way of collecting a large amount of data, illustrating the very practical reasons for their popularity. They are also an incredibly useful way of filling in gaps in our knowledge, as the number of results from surveys announced in the media every day testifies. Polit and Beck (2008: 324) illustrates this usefulness by saying:
The greatest advantage of survey research is its flexibility and broadness of scope. It can be applied to many populations, it can focus on a wide range of topics, and its information can be used for many purposes.
In summary, surveys are widely accepted as a legitimate foundation for decision making in society. Their use has always been a feature of assessing maternity services, by both health professionals and consumer groups. This pattern is likely to increase in all areas of health care with the importance placed on user involvement under clinical governance.
The basic principle on which surveys are based is that if you want to know what is going on, then the best way to find out is to ask people. The results can include quantitative data in the form of numbers, or qualitative data in the form of words, with many surveys blending the two to give ideas about frequency backed up with comments on interpretations, experiences or opinions.
Surveys can vary in a number of ways, one of the most important being the degree of structure they contain. Very structured surveys are useful if the researcher wants to gain some idea of the kind of pattern involved with a certain type of activity, such as the number of women who intend to breastfeed, or the information needs of first-time pregnant mothers. This type of survey has the great advantage that the results are reasonably easy to analyse, as it usually involves counting the number who said ‘yes’ or ‘no’ or the numbers choosing each of the options in a list. This method is also a common method within audit, where a snapshot of a particular activity in one clinical area is produced.
However, the disadvantage of this method, as Polit and Beck (2008) observe, is that it can be superficial and still leave us asking ‘why’ in relation to a particular pattern of behaviour or opinion. More in-depth information can be gathered using a less structured data collection tool. However, with a less structured approach the researcher can be faced with information overload, and the answers may be very different from each other, making comparisons and summaries difficult.
A further variation is the time period examined by the survey. It can cover a ‘one-off’ collection of data, sometimes referred to as a ‘cross-sectional survey’, which considers the characteristics, experiences or views of those in a variety of clear subgroups in a population. The same survey can be repeated after a period of time with different groups to identify possible changes. For example, the number of women wanting a home birth could be gathered at one point, and then repeated one or more years later to see if there has been a change. This is known as a trend survey, and works in the same way as audit where the same measurement is used at different times on different people to gain an indication of changes.
A different form of the survey is the longitudinal study, which can follow the same group of people over time and keep going back to them to note any changes. This is sometimes called a panel study where the researcher returns to the same group of people over time to identify trends or changes (Burns and Grove 2009). One form of the longitudinal study is the cohort study, which follows one specific group of respondents over time (McKenna et al. 2010). This usually tracks the members of the group to establish possible influences of specific conditions. An example would be the study of a group of babies born to first-time mothers in one maternity unit in one month, who are followed-up over a 10-year period to plot any differences in childhood illnesses that might be influenced by type of infant feeding. The purpose of such studies is usually to identify risk factors for certain conditions (Houser 2008). A major problem with all types of studies that include repeated collections of survey data from the same group is people dropping out of the study (called ‘study mortality’). Where this results in large numbers of people being lost to the study, it is likely to be difficult to develop any clear conclusions because of the amount of missing data.
There are two further problems concerning the use of surveys: the first relates to the representativeness of the sample, and the second relates to validity, or what is actually being measured. The issue relating to the sample is important, as the researcher frequently wishes to generalise the findings to similar people in the wider population. For this reason, researchers must try to avoid bias in the way the sample is chosen. Other methods of sampling will be covered in Chapter 14. Naturally, to get a reasonable picture of what is going on through the survey, the numbers returning questionnaires must be large and should as far as possible be representative of the total group. This is not always the case, which weakens the use of surveys in evidence-based practice, as it is often difficult to know if the results provide a true or accurate picture.
In designing the survey, it is usual to collect some basic ‘demographic’ details such as age, grade of staff, sex or number of children, so that some comparison can be made with those in the wider population to establish if they are a close comparison. This then gives some indication of how far the sample is representative.
The second question relating to validity is a difficult issue. One of the problems in surveys is that we have to accept that the answer given by a respondent is true, despite being unable to test it. There are likely to be some respondents who give socially acceptable answers in surveys, that is, answers that suggest the ‘perfect citizen’, so the accuracy can never be 100% certain. The only action the researcher can take is to try to reduce the amount of distortion produced by the data collection tool. The pilot study is one method by which this can be attempted.
The use of questionnaires
Questionnaires are a tool of data collection consisting of a series of questions that are usually completed by the participants themselves (self-completing) and can take the form of paper, electronic text sent by email, or on-line versions. They collect information on such things as personal attributes, attitudes, beliefs experiences, behaviour and activities (Parahoo 2008), and are probably the most familiar data collection tool in nursing and midwifery research. Most of us are likely to have completed many questionnaires in our working lives, and perhaps deleted or thrown many more away. In this section we will examine some of the advantages and disadvantages of this method and outline some of the important principles of questionnaire design.
A large number of research questions in midwifery have been answered by questionnaire. Why are they such a popular method of collecting research data?Box 9.1 suggests some of the answers.
BOX 9.1
Advantages of questionnaires
Cheap
Quick
Can reach a large geographically spread sample
Can be quite detailed
Low level of embarrassment or threat to both researcher and respondent
Anonymity is protected
Fixed-choice questions easy to answer and to analyse
Familiar method to respondents
Although this list suggests that questionnaires are an ideal data collection method, several factors may discourage a response, as Box 9.2 suggests. There are, then, a balancing number of disadvantages to the use of questionnaires. However, people have now received so many questionnaires that their motivation to complete and return yet another may be low, resulting in a poor response rate, that is, the number of those retuning a questionnaire expressed as proportion of those receiving one. As Burns and Grove (2009) suggest, if the response rate is below 50%, the representativeness of the sample returned can be seriously in question, as there is no certainty that the responses represent the views of those sent a questionnaire. In other words, we may end up with a biased response. Generalisations from this group would then be impossible. Reminders can be sent to increase the final response rate, although the return from this may also be disappointingly low.
BOX 9.2
Disadvantages of questionnaires
Questionnaires have now saturated the population
Response rate may be low dependent on feelings on the topic
Questionnaires are dependent on reading levels, literacy skills and physical ability
Responses will be influenced by the quality of the design and the motivation used to produce a response
There is no opportunity for clarification of questions or answers
Fixed-choice questions may not have an appropriate option for everyone
Similarly, a further limitation to their use is the often overlooked assumption that questionnaire recipients can read and write in English. This is not simply an issue of literacy or English not being a first language; many people have trouble with eyesight or problems with the ability to write because of physical problems. For this reason, surveys involving the public can omit important opinions because certain groups are excluded through the choice of research method.
The disadvantages of questionnaires relate to both the researcher and the respondent. A recurring frustration for the researcher is receiving returned questionnaires containing a number of unanswered questions, or incomplete answers. Once returned, it is not possible to clarify answers or probe further. From the respondent’s point of view, one irritating feature of questionnaires is to be asked to choose from a range of options where none apply to their own situation. This raises questions of validity where a respondent is not truly describing his or her own views or experiences but merely choosing something the researcher has preselected. A final problem is where the researcher uses value-leaden words such as ‘unnecessary’, ‘painful’, ‘appropriate’, ‘satisfied, ‘too long’, instead of phrasing questions in a neutral way. The conclusion is that as a method the questionnaire has a number of advantages and disadvantages. They are surprisingly not an easy option to use in research, and have a number of restrictions and limitations. Their success depends on the design, as this will influence whether an individual will complete them or not. For this reason, the next section will consider some of the principles involved in questionnaire design.
Questionnaire design
The first stage in designing a questionnaire is to ensure that it is an appropriate method for collecting the data. Careful consideration should be given to the advantages and disadvantages outlined above. The study aim should also provide some clue as to whether it is appropriate. If the aim depends on finding out what people say or do, or relates to areas where individuals are in the best position to accurately supply the information, then questionnaires would seem a good option. The review of the literature should help to establish whether previous studies have used a questionnaire, and with what success. Once a final decision has been made to use a questionnaire then the researcher moves into the design stage. This does present a challenge, as Parahoo (2008: 300) observes:
Developing questionnaires is systematic, time-consuming and laborious, requiring much preparation before the first question is phrased.
This can be made easier by following some clear principles. First, divide the questionnaire into its main parts:
• invitation and motivation to take part,
• instructions,
• body of the questionnaire.
Invitation and motivation to take part
This section has traditionally been referred to as the ‘covering letter’ but with the variety of ways in which questionnaires are distributed, including web-based and emailed questionnaires it is better to talk about the invitations and motivation to take part (see Box 9.3). As Polit and Beck (2008)
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree