11 Survey management

In this chapter

By the end of this chapter, you should be able to:-

• state why surveys need to be actively planned and managed;
• estimate the resources – human resources, time, equipment, materials and money – required to conduct a survey;
• describe aspects of mailing procedures for postal surveys;
• identify and select appropriate follow-up procedures for postal surveys;
• outline the nature and use of survey data control systems;
• identify key ethical considerations in conducting survey research.

Active survey management

Good questionnaire and survey design is not enough – the  whole survey process needs to be carefully and actively managed.  The senior researcher generally takes the role of project manager and team leader and as such must take responsibility for:-

• estimating and procuring the resources needed to carry out the project (meaning human and physical resources as well as money to pay for them);
• time-tabling survey operations and processes (making careful estimates of how long must be allowed to complete each) and monitoring  progress against the time-table
• managing staff and resources, negotiating and trouble-shooting as required.

Estimating and procuring resources

Fink (1995) has suggested that the following questions should be asked in determining survey resource needs:-

• What are the main tasks to be carried out?
• What skills are needed to carry out these tasks?
• How much time is needed for each task?
• Who can be used to perform each task?
• What are the costs of each task?
• What additional resources are needed?

Human resource functions and skills may be divided into research and project management functions, field data collection functions and other functions. Of course with small surveys, the project manager may take personal responsibility for carrying out some of the latter.

• Research and project management functions
• designing, planning and managing the research project as a whole;
• ensuring that adequate resources are available in the right place at the right time;
• negotiating with clients, sponsors, gatekeepers, resource managers;
• ensuring that ethical principles and codes of confidentiality are upheld;
• designing and drawing samples;
• designing questions and question sequences;
• drafting questionnaires and other survey documents (including layout) – technical elements of this (such as word-processing) may be delegated to secretarial or clerical staff;
• pre-testing and pilot testing questionnaires;
• organising reprographics and printing – possibly delegated to secretarial or clerical staff;
• designing procedures for obtaining complete and good-quality response;
• checking and coding data (or instructing and supervising others in doing so);
• analysing and interpreting data;
• reporting findings.

• Field data collection functions (more applicable to interviewer-administered and to captive audience self-completion surveys than to postal surveys)
• approaching sampled individuals and securing their co-operation;
• interviewing respondents;
• coding data in the field;
• completing fieldwork returns. 
• Other human functions (in particular, administrative and clerical functions)
• laying out, word-processing and proof-reading documents;
• data-base set-up and management (e.g. for issuing questionnaires and reminders, booking in returns);
• stuffing and opening envelopes;
• constant checking to make sure that numbers tally etc;
• routine data coding, checking and editing tasks;
• handling enquiries from respondents.

Other resources which must be budgeted for and procured include:-

• Printing / copying facilities
• for questionnaires;
• for covering letters and reminders;
• for other survey documents.
• Postage facilities (stamps, business reply envelopes or franking facilities)
• for outgoing questionnaires and letters;
• for return packages;
• for reminders.
• Telephone facilities (including perhaps answering machine or voice-mail facilities)
• for outgoing calls (especially where arrangements to interview are being made by telephone, or telephone follow-up procedures are being employed);
• for incoming queries from respondents (this should not be ignored!).

• Computing facilities
• hardware, including computers (for word-processing documents, data-base maintenance, data analysis etc); printers (for printing questionnaires, letters and other documents); scanners (if OMR / OCR questionnaires are being used – see Chapter 13);
• software, for word-processing, data-base management, statistical analysis;
• data entry facilities (commonly a commercial data preparation agency may be used).
• Stationery
• paper on which to print questionnaires and other documents;
• appropriate letter-head for covering letters;
• envelopes for outgoing post and for return of completed questionnaires (the latter should be stamped or reply-paid, and addressed to the survey organisation;
• labels for addressing envelopes (if “window” envelopes are not being used).

Time-tabling a survey

Adequate time must be allowed for each stage in the survey process.  This includes not only the time it will take to carry out the process from start to finish, but any lead time needed to prepare the staff and resources to carry out the tasks. The timing of what are often laborious tasks depends upon:-

• the magnitude of the task measured in person-hours (taking account of the size of the sample to be processed);
• whether the work to be done will arrive all at once or in instalments;
• the number of staff that can be devoted to it and the skills and experience of those staff;
• whether the questionnaire needs to be developed from scratch;
• the mode of administering the data collection procedures (e.g. postal self-completion, other self-completion, telephone);
• the number of pre-notification and follow-up (reminder) contacts to be made. 

An example of a timetable for a postal survey is presented in Table 12. Of course, not all postal questionnaires will follow an identical schedule. There are no hard and fast rules about the timing of dispatch of reminders, but it is sensible to wait until responses to previous mailings have tailed off (generally after two to three weeks).  It is also useful to specify a cut-off date after which any additional responses will be excluded from the data set.  Nonetheless, it can be seen that a postal questionnaire survey, if carefully designed, prepared and conducted so as to maximise response and obtain good quality data, can easily take eight or nine months from start to finish and is therefore not a way of obtaining information to answer policy or administrative questions overnight.  The need for a survey needs to be anticipated well before the information flowing from it is to be used! 

In planning and time-tabling a survey, careful consideration needs to be given to the inter-dependencies between  the various tasks.  Some activities can be carried out in parallel (at least in principle); for example, the identification of an appropriate sampling frame and selection of the sample can proceed alongside the drafting of questions.  Other tasks have to be carried out in sequence, with the start date for one task depending on a preceding activity being completed; for example, questionnaires cannot be sent out until they have been typed, proof-read and printed, the sample has been drawn and stationery has been procured.  Yet other tasks have a lag time in relation to their antecedents: for example, a number of days must elapse between initial mailing and receiving a completed questionnaire for processing.  The tasks on the critical path (those tasks that must be started or completed by a certain date if the survey is not to over-run) must be clearly identified. As the survey progresses, actual performance should be compared with the projected schedule, and remedial action taken if necessary.

 Costs and costing

As well as time, the resources required to carry out a postal survey will include stationery, postage and staff time. Of these, staff time is likely to be by far the most costly item, even though for in-house surveys it may not be fully recognised in accounting terms and may make its impact as an opportunity cost of taking staff off other tasks. From this it follows that failure to estimate the staff time requirement realistically is likely to lead to difficulties and contention within the organisation and to skimping of survey tasks, particularly those related to survey quality and quality control.

It should also be noted that the costing of postal surveys is not only a question of identifying the operational stages and the types of resource needed at each, but also requires the making of quantitative estimates of unknown factors. For example, estimates must be made of the numbers of sample members who will respond to the initial mail-out and to successive reminders, and of the numbers of case records to be processed (particularly at the manual processing stages). To inform the planning of future surveys, it is good practice to keep careful records of these quantities.  Estimates can also be informed by pilot work (see Chapter 9).

Table 12 Example of a postal survey timetable

The worked example below shows the principles of costing a survey: In a survey, the initial sample size is to be 500.  To boost response rate, two reminders are to be used; the first will be a letter; the second will include a duplicate questionnaire as well as a letter. Reply-paid envelopes will be included with the initial mailing and the second reminder. It is assumed that 40% of those contacted will respond to the initial mailing.  30% of the remaining 60% will respond to the first reminder, while 40% of the residual 42% will respond to the second reminder. 

Anticipated overall unit response rate: \(374 \div 500 \approx 75\%\).

Materials needed:

• questionnaires: \(500 + 210\) for the initial mailing and second reminder
• letterhead: \(500 + 300 + 210\) across all mailings
• large envelopes (A4 size)
• for posting out: \(500 + 210\) for the initial mailing and second reminder
• for return post: \(500 + 210\), enclosed with the initial mailing and second reminder
• small envelopes (DL size): \(300\) for the first reminder

But note that there may be indivisibilities – envelopes tend come in boxes of 250 or 500, and it is not possible to buy ½ a box – which may necessitate purchasing more materials than are strictly required.

Costs:

• printing of questionnaires: unit cost multiplied by \(500 + 210\)
• stationery (envelopes, letterhead - quantities as above)
• postage
• outgoing cost: questionnaire-package cost multiplied by \(500 + 210\), plus letter cost multiplied by \(300\)
• return cost: return-package cost multiplied by \(200 + 90 + 84\)

Don’t guess at the cost of postage – measure and weigh out the appropriate packages and work out costs accordingly!

Note also that these are initial estimates of the amounts of materials required and costs involved.  If response rates to the various mailings are lower (or higher) than anticipated, the resource implications will change.  Therefore it is important to continue to monitor actual performance against predictions and to take action – for example, buying extra stationery – as necessary.

Management of mailing operations

Components of the survey package

The package that the target respondent receives should, as a minimum, comprise a questionnaire, and covering letter (see below for more detailed recommendation on the content and style of covering letter).  In the case of a postal survey, an envelope in which to return the completed questionnaire to the survey organisation should also be included.  In other self-completion questionnaire surveys, provision must also be made for returning questionnaires.  In captive audience surveys, a field-worker or researcher may distribute the questionnaires and then gather them in at the end of the session (e.g. lesson period for surveys of students).  Where questionnaires are self-completed in the absence of a field-worker, a box should be provided in which to place completed questionnaires.  In both of these situations, providing an envelope in which to place the completed questionnaire prior to handing it in helps to preserve confidentiality.

Identifying questionnaires

Managing the dispatch of questionnaires and logging returns needs careful attention.  For reasons of confidentiality (see below), names and addresses of target respondents are not usually placed on the questionnaires themselves; rather a unique identifier (survey number) is used.  Consideration needs to be given to how these numbers are added.  Sophisticated printing and document reproduction systems allow for the production of personalised questionnaires, using a facility akin to the mail-merge function in word-processing packages; however, this is a relatively costly option.  A lower cost alternative is to use a word-processing or database package to produce individualised sticky labels which can be stuck to the questionnaires.  Similarly, self-inking, automatically advancing number stamps can be used to number questionnaires sequentially.  The least satisfactory alternative is to write the survey number on by hand, since this approach is error prone.  Whichever method is used, care must be taken to place the correct questionnaire and any supporting documentation (for example, covering letter) in the right envelope.

Covering letters

In an interviewer-administered survey, the tasks of introducing and explaining the survey are largely the responsibility of the interviewer (although an introductory letter may be sent in advance).  In postal and other self-completion surveys, a covering letter is an essential part of the “package” to be sent out to target respondents. It has several functions, all of which are direct or indirect contributors to obtaining a satisfactory rate of response (see Chapter 10).  The content and tone of the letter can be a major influence on response rates and quality of response. It is best for the covering letter to be a sheet separate from, rather than forming the front sheet of, the questionnaire to be returned. The respondent then has information to keep for future reference.  Meanwhile, the researcher can use the front page of the questionnaire to convey and record additional information and instructions. A good covering letter should:-

• Not be too long – a maximum of one side of A4 paper is appropriate.  If it is felt that additional information about the study or survey needs to be provided, a separate ‘about this study’ insert could be considered.  We have found a ‘frequently asked questions’ format (based on anticipated queries, and those raised in pilot work and in previous similar surveys) to be effective here.
• Be simple – avoid extraneous detail and over-elaborate explanations. 
• Be on letter head paper – use of letter head paper with an appropriate logo gives an impression of seriousness and importance.  It can also indicate who is sending out survey and who the sponsor or client is; if the sponsor and sender are not the same, the text of the letter should convey details of both sponsor and sender.
• Refer to the title of the survey (which should also appear on the cover of the questionnaire); the title should be brief, non-misleading and likely to catch the attention and attract the goodwill of recipients.
• Provide succinct explanations of:-
• the aims of the survey, why it is important and how the findings will be published and used;
• how and why the respondent was chosen, using language and concepts appropriate to the target respondents (e.g. the concept of ‘random sampling’ may need to be explained in lay person’s terms);
• relevant principles of confidentiality (but be careful not to over-stress this, since over-elaborate explanations can raise anxieties that there is something to hide);
• the voluntary nature of participation in the survey;
• the implications of participation (for example, that services or care provided will not be affected by the respondent’s decision regarding participation in the survey; whether there will be follow-up surveys).
• Give clear instructions on who should complete the questionnaire and how (and when) to return it.
• Make an appropriate appeal to the respondent.  Research (McColl et al., 2001) has shown that there is no one style of appeal that is universally superior.  Instead, it is important to consider the likely motivations of the target respondents and to choose the type of appeal accordingly.  Options are: a self interest appeal (i.e. what the respondent himself/herself will get out of participating, or how s/he will personally benefit from the conduct of this survey; an altruistic appeal (i.e. how the respondent’s input will benefit other people); an appeal to extending knowledge in general.
• Avoid giving a sense of obligation / pressure.

• (Possibly) say how long the questionnaire is likely to take to complete – the estimate should be as accurate as possible (e.g. based on pilot work), but it would be appropriate to indicate that some people may take less or more time.  It may also be helpful to specify a deadline for response – it is best to convey that return by this date would be greatly appreciated, without implying that late returns will be disregarded.
• Give contact details (preferably a phone number – with an adequately briefed individual at the other end!) in case of queries or a desire to check on the bona fides of the survey.
• Thank the respondent for co-operation (though it is also desirable to reiterate the researcher’s gratitude on the last page or back cover of the questionnaire itself – see Chapter 6).

Dillman (1978) strongly recommended personalisation of covering letters to “show regard for the respondent”; his recommendations included individually addressed letters and envelopes, and hand-signed letters; he did not appear to consider that this may be infeasible or not cost-effective in large surveys.  Despite these recommendations, empirical evidence on the impact of personalisation is equivocal, especially in relation to specific populations (McColl et al., 2001). With modern database and word-processing packages, personalising the address and salutation through mail-merge facilities is generally feasible if time-consuming.  This level of personalisation is probably appropriate where the method of sampling and tone of the letter implies a personal knowledge on behalf of the sponsor or survey researcher of the target respondent (e.g. in a survey of patients selected through their medical records, who are being contacted by a health care professional involved in their care).  An open (i.e. Dear Patient, Dear Colleague, Dear Householder etc.) letter, combined with a personalised address label on the outgoing envelope is a cheaper and quicker option, and may be adequate for most surveys. In this era of junk mail, however, an envelope addressed simply to ‘The Occupier’ may be less effective than an individually addressed envelope.  Hand-signing letters is probably not an option in all but the smallest surveys – fortunately, high quality replication is now obtainable through rubber signature stamps or by scanning a signature into word-processing software.

Postage options

A range of postal services, to facilitate the dispatch and return of postal questionnaires is available.  Although Dillman (1978)  and others have recommended that stamped rather than franked or reply-paid envelopes should be used for both outgoing and return postage, evidence from recent reviews (McColl et al., 2001; Edwards et al., 2002a, 2002b) has shown little consistent effect of postage type on response rates, non-response bias or quality of response, and evidence of cost-effectiveness of more resource-intensive techniques (e.g. recorded delivery, use of stamps as opposed to reply-paid or franked envelopes) was scant.  Edwards and colleagues (2002; 2009) showed some advantage over first class postage versus other classes of mail, but the primary studies they reviewed included some from outwith the UK, and care should be taken in generalising these findings from other settings.  First class obviously offers a slight advantage in terms of speed of delivery.  However, the increased cost may outweigh the benefits.  What is most important is that the return postage costs are borne by the survey organisation (either through the use of stamped, addressed envelopes or business reply envelopes) rather than the respondent.  This can be achieved through the use of pre-stamped and addressed envelopes, business reply envelopes and Freepost services.  With the first of these options, the cost of the stamp is obviously incurred regardless of whether the questionnaire is returned; with business reply envelopes and Freepost services, the return postage cost is only incurred if the item is posted back, but a small additional cost over and above that of the price of a stamp for an item of equivalent size and weight is charged on each item, and a one-off license fee is also required.

Royal Mail offer a range of facilities of relevance to those conducting postal surveys; these are primarily designed for business users and include business-reply and freepost services, and discounts for pre-sorting bulk posting by post-code.

Royal Mail explains its current Business Reply and Freepost Plus services.

Royal Mail’s delivery-exceptions guidance explains circumstances that can prevent delivery.

Royal Mail’s delivery-exceptions guidance describes circumstances that can prevent delivery and the action taken.

Royal Mail provides a current Track your item service for eligible mail.

Follow-up procedures in postal surveys

As noted in Chapter 10, one of the most powerful means of increasing response rates is through multiple contacts with target respondents.  One approach is pre-notification – sending a letter or making a telephone call to the sampled individuals in advance of mailing the questionnaire to alert them to its arrival.  More usual is the use of reminders.  In order to respond, the sampled individual must have both the will and the means to respond.  In sending reminders, it is important to address both of these issues.

To stimulate the will to respond, points to consider emphasising (although empirical evidence on the most effective approaches is currently lacking) are:-

• Progress to date with the survey – some respondents will be susceptible to peer pressure and may be encouraged to respond by learning how many people like them have already returned a questionnaire. However, there is the risk that others may judge that enough of their peers have responded, and that their individual contribution is therefore not needed!
• The importance of that individual’s response – as just noted, some individuals may feel that they personally do not need to bother responding, especially if response rates to date are good (What difference will my answers make?).  It is therefore important to emphasise the value of each and every response. As with the initial covering letter, it is important to consider the likely motivation of the study population.  It may be worth considering varying the appeal made from that in the initial contact, for instance from an altruistic to a self-interest appeal.
• What to do if the individual has already responded – for example, should they make contact with the survey organisation to check if their questionnaire has been received, or should they simply assume that correspondence has crossed in the post?

But respondents also need the means to respond.  Possible reasons for (apparent) non-response is that the original questionnaire: was never received; has been damaged or mislaid; or has been lost in the return post.  It is therefore a good idea to offer a duplicate questionnaire under these circumstances.  However, research has shown that routinely enclosing a duplicate questionnaire with a first reminder has little effect on response rates to that reminder, but that enclosing it with a second reminder leads to a significantly higher response rate on that occasion (Swan et al., 1980). Mixed mode reminders can also be effective if resources are available – for example, a final reminder in the form of a telephone call (with, perhaps, an offer to complete the questionnaire over the phone). Postcards are a cheaper alternative to letters, but may not be suitable in surveys on sensitive topics, or where the preservation of confidentiality is a major consideration. 

Consider also how many reminders to send and at what time intervals.  Every reminder will attract more responses, but at a diminishing rate and will extend the data collection timetable by a couple of weeks.  Also, repeated reminders may cause significant annoyance to those contacted (in health surveys, some research ethics committees feel that more than three contacts in total – an initial mailing and two reminders - constitutes undue pressure (L Key, personal communication)).  Finally, the more reminders are used, the higher the cost and the longer the time required for the survey.  Two, or at most three, reminders are probably adequate in most social surveys. 

One way of deciding on the timing of reminders is to monitor the rate of return of completed questionnaires and to send the reminders when the return rate falls.  However, this ad hoc approach is at odds with the notion of planning and scheduling which we recommended above.  The point at which the rate of receipt of returns falls to very low levels will depend to some extent on the length of the questionnaire and on whether first or second class postage is used.  Typically, however, the rate of response tends to drop off significantly two to three weeks after a given mailing.  This pattern suggests that a first reminder should be sent two to three weeks after the initial contact, with the second reminder at four to six weeks after the original mailing.

Survey data control systems

Managing dispatch and return of questionnaires, and sending reminders, implies the need to know who has received a questionnaire and who has returned one. For very small surveys, a paper-based control system may be used.  However, a computerised system, using a data-base package is preferable.

The purposes of postal survey control systems are to:-

• initiate action (e.g. original mailings and reminders) for each selected individual;
• provide up-to-date reports for monitoring overall progress against plans;
• allow remedial action to be taken if necessary;
• support reports of survey outcome and quality (e.g. response rates; the extent and nature of any non-response bias).

The inputs to the control system are:-

• the list of sampled individuals, including allocated identification numbers;
• the occurrence and dates of events for each individual (e.g. original mailing, reminders);
• the outcomes of each event (e.g. completed questionnaire returned, notified by Royal Mail that respondent was not known at specified address).

The control system should be used:-

• during the rehearsal pilot stage – at this point, the system should be set up and tested;
• before data collection begins – the details of the sampled individuals should be entered and survey numbers assigned;
• when the initial contact is made – the date of initial contact should be recorded;
• on an on-going daily basis during the data collection period – the outcomes of each event should be recorded; relevant details (e.g. eligibility status, addresses) should be amended as required;
• periodically (e.g. weekly) during data collection period – progress reports (especially response rates to date) should be run, and corrective action taken if needed;
• in line with follow-up protocols – reminders should be generated for non-respondents; the event and date should be recorded;
• after the close of data collection – the data-base file should be reconciled with the data files (from entering questionnaires) and a final analysis of response rates and patterns (including, if possible, non-response bias) produced.

Details should be regularly updated (preferably on a daily basis).  Database fields should generally include:-

• name and address of sampled individuals, to facilitate production of covering letters and address labels;
• personal details of sampled individuals (for example, age, date of birth) to facilitate comparisons of respondents and non-respondents;
• date when initial questionnaire sent / given;
• whether the individual has responded;
• date when the individual responded;
• whether a reminder was required (possibly distinguishing between first and second reminders, to allow comparison of early and late respondents);
• other useful details (for example, if package or letter was returned by Royal Mail as ‘gone away’).

Data quality monitoring and data management

Finally, when questionnaires are returned, they need to be carefully checked before data is entered on to a computer for analysis.  Things to look out for include: data completeness; inconsistent or implausible answers; respondents not following instructions (for example, writing the response in the wrong place; endorsing two response categories when only one is allowed).  Any errors should be noted and, if possible, rectified.  If open questions have been used, these will need to be coded prior to analysis.  Once data have been entered on to the computer, further checks will be required before these data are analysed. Things to check for include: whether skips have been followed correctly; whether the answers given are valid (for example, within sensible range); whether responses are internally consistent (for example,  that a respondent is not claiming to be male and pregnant).  Issues of coding and data processing are discussed in greater detail in Chapter 13.

Ethical issues and confidentiality in surveys

The survey researcher needs to pay attention to the ethical principles laid down by his/her profession and the ethical requirements imposed by the survey topic.  For example, in surveys on health-related topics, especially those of specific patient groups, approval must be obtained from the appropriate Multi-Centre or Local Research Ethics Committee, who will want to see copies of questionnaires, covering letters and other survey documents.  It is also important to ensure that all data holdings are registered appropriately in conformance with the Data Protection Act (1998).

Maintaining confidentiality in survey research is most important.  Confidentiality means that:-

• Questionnaires are only identified by a survey number and  only members of the survey research team can make a link between this survey number and identifiable information (such as names and addresses).
• Access to completed questionnaires is strictly controlled.  All those handling questionnaires should be bound by a written code of confidentiality.  Completed questionnaires should be stored in locked rooms or cupboards.
• Individual, identifiable responses are not revealed to any third party.
• Results of the survey are not presented in such a way as to allow identification of any individual respondent.

It is important to note that confidentiality is not synonymous with anonymity (Zeinio, 1980).  In a completely anonymous questionnaire, no identifier whatsoever (not even a survey number) appears on the questionnaire.  Anonymity precludes record linkage (e.g. relating someone’s questionnaire responses to data on that same individual obtained from another source, such as a scholastic or medical record).  Anonymity also makes sending reminder questionnaires difficult, since there is no way of knowing who has responded and who has not.  This generally means that reminders have to be sent to everyone, which clearly increases costs, can cause annoyance and confusion to those who have already responded, and can result in getting two completed questionnaires (often with different responses!) from some respondents.  One way around this, if anonymity is felt to be essential, is to ask respondents to send back an identifiable letter or postcard under separate cover (Sudman, 1985), so that the survey researchers can keep track of who has and who has not responded (although they will be unable to work out which questionnaire came from which respondent).

Written informed consent for surveys may not be insisted upon by research ethics committees.  Rather, they may allow that agreement to be interviewed or return of a completed questionnaire is indicative of implied informed consent.  Nonetheless, those approached should be properly informed of the implications of participation, and provided with assurances of confidentiality.

Summary of key points

• To ensure timely and cost-effective collection of high quality data, surveys need to be actively planned and managed.
• Resources must be estimated and procured.  The resources needed include human resources, equipment, materials, time and money.
• Human resource functions and skills comprise: research and project management functions; field data collection functions; administrative and clerical functions.
• Other facilities and resources required include: printing facilities; postage facilities; telephones; computer hardware and software; stationery.
• Adequate time must be allowed for survey data collection; the inter-relationship of complex and often laborious tasks must be understood if a survey is to be successfully scheduled and implemented.
• Costing a survey involves the identification of the main operational stages and the type of resources required at each stage, and the making of quantitative estimates of often unknown factors.
• In postal surveys, mailing operations need to be actively managed. Due attention needs to be given to identification of questionnaires and associated documents, production of covering letters and options for outgoing and return postage.
• Adequate follow-up procedures are a key element in achieving good response rates; promoting the will and the means of responding both need to be considered.
• Good survey data control systems are needed to facilitate mailing operations and to allow monitoring of the quantity and speed of response; the use of computerised databases is recommended.
• Survey data should be subjected to rigorous monitoring for quality assurance; manual inspection and editing should be supplemented by computerised checks of validity and consistency of responses.
• Sound ethical principles, including maintenance of confidentiality, should be observed in the conduct of surveys.

Further reading

Dengler R. Organisation and management of postal surveys. Survey Methods Centre Newsletter 1996, 16(1), 14-16.

Kelly K, Clark B, Brown V and Sitzia J. Good practice in the conduct and reporting of survey research. International Journal for Quality in Health Care 2003; 15(3) 261-266. https://doi.org/10.1093/intqhc/mzg031

Mangione TW. Mail surveys - improving the quality. Thousand Oaks: Sage Publications, 1995. (Applied Social Research Methods Series, Volume 40) (Chapters 6 and 7).

• A key web site on data protection issues in the United Kingdom is at https://ico.org.uk/for-organisations/
• Details of the General Data Protection Regulations, which have replaced the 1998 Data Protection Act, may be found at https://ico.org.uk/for-organisations/guide-to-the-general-data-protection-regulation-gdpr
• The Health Research Authority web site is at https://www.hra.nhs.uk/
• NHS codes of practice on confidentiality may be accessed at  https://digital.nhs.uk/data-and-information/looking-after-information/data-security-and-information-governance/codes-of-practice-for-handling-information-in-health-and-care
• Details of the UK Policy Framework for Health and Social Care Research which has a key role in “ensuring that health and social care research is conducted to high scientific and ethical standards” may be found at https://www.hra.nhs.uk/planning-and-improving-research/policies-standards-legislation/uk-policy-framework-health-social-care-research/
• Royal Mail describes its current Business Mail services.
• The UK Data Archive provides guidance on ethical issues, including consent, in surveys at  https://www.ukdataservice.ac.uk/media/428433/ethicallegal_informedconsent_34apr2014_ukds.pdf  and  https://www.ukdataservice.ac.uk/manage-data/legal-ethical.aspx

Current guidance and methodological literature for 2026–27:

• American Association for Public Opinion Research. Code of Professional Ethics and Practices. revised 2026.
• UK Data Service. Introduction to effective and practical research data management. 2024.

References

Dillman DA.  Mail and telephone surveys: The total design method,  New York:  John Wiley and Sons, Inc, 1978.

Edwards P, Roberts I, Clarke M, DiGuiseppi C, Pratap S, Wentz R, Kwan I.  Increasing response rates to postal questionnaires: systematic review.  British Medical Journal 2002;324;1183-1192. https://pubmed.ncbi.nlm.nih.gov/12016181/

Edwards P, Roberts I, Clarke M, DiGuiseppi C, Pratap S, Wentz R, Kwan I. Methods to influence response to postal questionnaires (Cochrane Methodology Review). In: The Cochrane Library, 2009. Oxford: Update Software. https://pubmed.ncbi.nlm.nih.gov/19588449/

Fink A. The survey kit. Thousand Oaks: Sage Publications, 1995.

Jacoby A, Thomas L, Soutter J, Bamford C, Steen N, Harvey E, Garratt A and Bond J. Design and use of questionnaires: a review of best practice applicable to surveys of health service staff and patients. Health Technology Assessment 2001, 5(31) Downloadable from  https://www.journalslibrary.nihr.ac.uk/hta/hta5310/#/abstract .

Sudman S. Mail surveys of reluctant professionals.  Evaluation Review  1985;9:349-360

Swan JE, Epley DE, Burns WL. Can follow-up response rates to a mail survey be increased by including another copy of the questionnaire?  Psychological Reports  1980;47:103-106.

Zeinio RN. Data collection techniques: mail questionnaires.  American Journal of Hospital Pharmacy  1980;37:1113-1119.