Non��response in survey research: a methodological discussion and development of an explanatory model

Non-response in survey research:a methodological discussionand development of an explanatory model

K. Louise Barriball BA RGN

Lecturer

and Alison E. While BSc MSc PhD RGN RHV Cert Ed

Professor of Community Nursing, Research in Nursing Studies Section,

The Florence Nightingale School of Nursing and Midwifery,

King's College London, London, England

Accepted for publication 26 November 1998

BARRIBALLBARRIBALL K.L.K.L. & WHILE A.E. (1999)WHILE A.E. (1999) Journal of Advanced Nursing 30(3), 677±686

Non-response in survey research: a methodological discussion

and development of an explanatory model

This paper demonstrates that the different sources of non-response in survey

research must be considered by researchers in order to minimize the potential for

bias, and that careful planning and management during sample selection, sample

recruitment and data collection can control the extent of response failure. The

®ndings of an interview survey of nurses and nurses' aides illustrate, however,

that the success of strategies designed to reduce the loss of data in survey

research depends upon researchers acknowledging the complex relationships

that exist between non-response phenomena throughout the research process.

Keywords: data collection, data loss, non-response, nursing research,

response failure, sample recruitment, sample selection, survey research

INTRODUCTION

The aim of this paper is threefold: to examine some of the

implications of non-response in survey research; to

discuss how non-response can be contained and managed;

and to highlight the complexity of non-response. The

literature addressing non-response is reviewed prior to a

discussion of the strategies employed by a research team

to reduce non-response in an interview survey exploring

the views and perceptions of continuing professional

education (CPE) among nurses (including midwives and

health visitors) and nurses' aides. An explanatory model

of data loss in survey research based upon the study's

®ndings is presented.

The signi®cance of non-response

While the survey offers a data collection method which

may yield greater validity through the use of more

representative samples, the potential for non-response is

a major disadvantage (Miller 1991). If response rates are

low and/or non-response is systematic and in some way

correlated with the variables under investigation, the

sample from which data are collected becomes unrepre-

sentative (Hartman et al. 1985). In such instances external

validity is threatened and valid conclusions from the data

cannot be drawn (Williamson 1981, Denzin 1989).

The methodological interest in non-response therefore

arises from its potential to introduce bias into the research

process since:

¼ the chances are good that those subjects who ®nd their way into

the research will differ appreciably from those subjects who do not.

(Rosenthal & Rosnow 1969 p. 69)

Correspondence: K. Louise Barriball, Research in Nursing Studies Section,

The Florence Nightingale School of Nursing and Midwifery,

King's College London, James Clerk Maxwell Building, Waterloo Road,

London SE1 8WA, England.

Ó 1999 Blackwell Science Ltd 677

Journal of Advanced Nursing, 1999, 30(3), 677±686 Methodological issues in nursing research

As Moser and Kalton (1971) cautioned, one should

never assume that differences between respondents and

non-respondents do not exist.

Towards a de®nition of non-response

Perhaps the most familiar understanding of non-response

is the degree to which a researcher does not succeed in

obtaining the co-operation of all potential respondents

included in the net sample (i.e. excluding individuals not

meeting the sampling criteria and containing only those

from whom one intends to obtain information). However,

this commonly used and apparently straightforward

de®nition is ambiguous. It also belies the complexity of

non-response phenomena. There are many potential

sources of non-response and it is important that these

are fully considered by researchers (Kviz 1977).

Steeh (1981), for example, identi®ed two major

components of non-response: refusals of potential

respondents to co-operate in the research process and

other reasons why data are not collected from all units in

the sample such as non-contact. In his paper examining

non-response, Aiken (1988) also addressed the problem of

refusals but included in his de®nition the problem of

non-coverage and the potential bias which the exclusion

of individuals from the sampling frame can introduce. In a

comprehensive review of non-response for the Of®ce of

Population Census and Surveys in Britain, Elliot (1991

p. 3) distinguished between three types of non-response:

· Non-coverage: when the sampling frame omits some

units of the survey population either accidentally or

deliberately.

· Unit non response: when no information is collected

from a sampled unit due to, for example, refusal or

non-contact.

· Item non-response: when the sampled unit agrees to

participate in the study but information on all the areas

under investigation is not collected because, for

example, the sampled unit refuses or is unable to

answer a particular question or the researcher fails to

ask the question by mistake.

Using Elliot's threefold typology as the basis for a

standard de®nition of non-response phenomena, it can be

seen that non-response can occur at different stages of the

research process, namely during sample selection, sample

recruitment and data collection. These three sources of

response failure, however, can be manipulated by

researchers to minimize the potential for non-response

bias (Figure 1) and strategies to achieve this are discussed

together with the usefulness of a linear conceptualization

of non-response.

THE STUDY

Two main approaches to address the issue of non-

response phenomena were adopted in the study. First, a

range of strategies were employed which the research

team anticipated would reduce non-response during the

course of the study. Secondly, the impact of the loss of

data at each stage of the research process on the

representativeness and the generalizability of the data

were assessed.

Figure 1 Literature based framework of non-response data loss in survey research.

K.L. Barriball and A.E. While

678 Ó 1999 Blackwell Science Ltd, Journal of Advanced Nursing, 30(3), 677±686

Non-coverage

It is usually not possible to obtain data from an entire

population which means that survey researchers need to

select and recruit samples. However, including some

individuals in a sampling frame while excluding others

runs the risk of introducing bias. A number of approaches

to ensuring the selection of representative samples have

been developed (e.g. strati®ed, random, cluster and census

sampling techniques, see Moser & Kalton 1971) and it is

the responsibility of all survey researchers to consider

carefully the approach or combination of approaches to

sample selection which will minimize non-coverage bias

within a given research situation.

The study designA census sample of randomly selected clinical units in

two district health authorities (Trust status had not been

granted at the time of data collection) was considered to be

the best study design to enhance the representativeness of

the sample and to generate a suf®cient sample size. The

clinical units were randomly selected by the research team

to represent the specialities of paediatrics, general med-

icine and surgery, care of the older adult, mental health

and learning disabilities as well as maternity and com-

munity care. While this is not an exhaustive list of the

clinical specialities available within each study site (with

peri-operative, accident and emergency and outpatients

departments, critical care and specialized units being

excluded from the sampling frame), there is evidence that

different clinical specialities employ staff with different

demographic, professional and employment pro®les, e.g.

age group, ethnicity, type of registration, number and type

of hours worked (Department of Health 1995).

It was important that the sampling frame consisted of a

heterogeneous group of potential respondents since the

evidence in the United Kingdom (UK) suggests that nurs-

ing staff with different pro®les gain differential access to

continuing professional education activities (e.g. Buchan &

Seccombe 1991, Larcombe & Maggs 1991, Chapman & Hall

1992, Rashid 1992). Therefore, the range of specialities for

inclusion in the study were purposively selected in order

to enhance the representativeness of the sample in the light

of the heterogeneity of the total population.

The limitations associated with purposive sampling,

notably the introduction of bias, are acknowledged.

However, decisions regarding the inclusion of different

specialities were informed by empirical evidence (e.g.

Department of Health 1995) and not judgement alone. The

potential for bias was also reduced by the random selection

of units representing each speciality (Moser & Kalton 1971).

In addition, it was appropriate to utilize strategies which

would contribute to the recruitment of a sample represen-

tative of the total population. Purposively selecting clinical

specialities was considered to be one such strategy.

Creating the sampling frame

A census sample approach (excluding students) was

selected for several reasons. For example, to enhance the

reliability of the sampling frame and to make sample

recruitment more manageable. Cartwright (1978) noted

that one of the major dif®culties facing researchers work-

ing with nurses is obtaining lists of staff working in the

areas to be surveyed. The research team was also con-

cerned about maintaining accurate records over several

months of data collection from potential respondents

consisting of an occupational group notorious for mobil-

ity. Following discussions with senior nurses it was

decided that the most accessible records of nursing staff

on each clinical unit were the duty rotas. As these tend to

be amended on a weekly basis it was also felt that the rotas

would provide the most reliable list of potential respon-

dents as well as indicate when respondents could be

visited in person about the study. Each rota was therefore

reviewed regularly by the research team so that, for

example, respondents planning to take annual leave or

commence maternity leave could be recruited to the study

prior to their departure and respondents due to return

from maternity or sick leave during data collection could

be included in the study.

Unit non-response

Several research texts have set out a number of detailed

strategies for minimizing the level of non-contact and

refusals in surveys using both questionnaire and interview

techniques such as the use of pre-noti®cation, personal-

ized and follow-up letters as well as face-to-face contact

with potential respondents (Champion & Sears 1969,

Dillman & Frey 1974, Gordon 1975, Bradburn & Sudman

1979, Miller 1991). Such texts are largely based upon the

assumption that there is much that researchers can do to

reduce unit non-response in survey research.

Non-contact

Contacting the sampleInitial contact with the study population was made

through a personalized letter to each potential respondent

which introduced the research team, explained the pur-

pose of the study, and gave assurances of anonymity and

con®dentiality for study sites and individuals. Secondly,

each clinical unit was given a poster about the study

which included photographs of the two nurse researchers

responsible for the collection of data so that all staff

(including those not in the sample such as student nurses,

ward clerks, doctors) and patients and visitors were

informed that a research study was in progress.

In addition to the personalized letters and posters, the

research team wanted to establish face-to-face contact

Methodological issues in nursing research Non-response in survey research

Ó 1999 Blackwell Science Ltd, Journal of Advanced Nursing, 30(3), 677±686 679

with potential respondents so that the purpose of the

study could be explained in more detail. It was anticipat-

ed that `being in the ®eld' on different days (including

weekends) and at different times (including during the

night) would ensure that personal contact was made with

the majority of the sample; however, relying solely on the

two nurse researchers visiting clinical units in the hope

that potential respondents were on duty would be inef®-

cient in terms of their time and effort. The duty rotas were

therefore used not only to identify individuals to be

included in the sampling frame but also to plan and guide

sample recruitment. The duty rotas provided a detailed

and up-to-date record of when staff might be available and

were adapted by the research team to create record sheets

of contacts with potential respondents, arranged inter-

views (e.g. dates and times), and dif®culties with data

collection (e.g. cancelled and/or re-scheduled interviews,

refusals).

Refusals

Study methodAs personal interviews are more costly and time consum-

ing than questionnaires (Burns & Grove 1987), careful

consideration was given to the selection of the interview

method for use in this study. The primary bene®t iden-

ti®ed was the widely acknowledged lower refusal rates

typically obtained with the personal interview (Kidder

1981, Goyder 1985, Treece & Treece 1986). Indeed, Kidder

(1981) has suggested that a response rate as high as 80% is

not untypical as potential respondents who do not have

the con®dence to write down responses will answer

questions in an interview. Furthermore, face-to-face

contact with a researcher can motivate respondents to

participate who would otherwise not bother with a

questionnaire (Gordon 1975); and respondents who

experience particular dif®culties with research, such as

those whose ®rst language is different from that of the

research team, may need extra support if they are to

participate (Marshall & While 1994).

Sample recruitmentIt has been suggested that conveying the importance and

usefulness of a study is the factor which most in¯uences a

respondent's decision to participate (Heberlein &

Baumgartner 1978, Hox & De Leeuw 1994). It was

considered that interest in the study could be generated

through a high level of face-to-face contact with potential

respondents. It was also considered that researcher

¯exibility during the recruitment of the sample would

help reduce refusals. There is an under-representation of

different `groups' of nursing staff (e.g. those working part-

time and/or night duty) in studies exploring continuing

professional education in British nursing (Clarke & Rees

1989) and an important aim in this study was to address

this under-representation. Arranging interviews at

weekends, in the evening or during the night shift to

accommodate the clinical responsibilities of the sample

was considered to be a potentially valuable strategy

in enhancing the representativeness of respondents

co-operating in the main study.

Acknowledging the rights of individuals to refuse to

participate, members of the sample who wished to decline

their invitation to be interviewed were reassured of their

right to do so. However, Moser and Kalton (1971) have

suggested that people who refuse to participate in the

main study may be willing to answer a few questions

about themselves. All non-respondents were therefore

asked if they would be prepared to complete a brief

questionnaire although each non-respondent was

informed that completion was not an obligation.

Item non-response

The prevention of item non-response is important since

missing information reduces the comparability of the data

and may introduce bias (Moser & Kalton 1971). There are a

number of possible factors which contribute to item non-

response and several strategies were employed in this

study to obtain complete data sets from each respondent.

In particular, the research team were mindful that, in the

light of the depth of data desired and the length of each

interview (approx. 45±60 minutes), interviewer and inter-

viewee fatigue was a potential risk.

Semi-structured formatOne of the many advantages of using the interview

method is that it can facilitate comparability of the data

by ensuring that all questions included in an interview

schedule are answered by respondents (Bailey 1987); it is

also a method that has the potential to generate detailed

data (Gordon 1975, Kidder 1981). To facilitate these

advantages a semi-structured interview format which

permits the use of alternative question wording and

probes was selected for use in this study. The wording

of questions must re¯ect the level of vocabulary used by

respondents to facilitate their understanding of what is

being asked and to maintain their interest and motivation

to answer questions during interviewing (Cannell & Kahn

1968). In the light of the heterogeneity of the sample

which included nurses' aides and respondents for whom

English was a second language, it was anticipated that the

use of the same question wording for all respondents may

be inappropriate as it was likely that not every word

would have the same meaning to every respondent and

that not every respondent would use the same vocabulary.

Since the use of alternative words which leave the

meaning of questions unchanged has been a strategy

successfully employed elsewhere (e.g. Carr 1988), ¯exi-

bility in the wording of questions was considered to be an



appropriate strategy for use in this study to encourage all

respondents to answer all questions.

It was also anticipated that the use of neutral probes

when necessary would enhance data reliability and

validity by: clarifying interesting and relevant issues

identi®ed by respondents; providing cues to respondents

about the level of responses desired; and eliciting

complete information (Nay-Brock 1984, Patton 1990,

Hutchinson & Skodol-Wilson 1992, Smith 1992). Probing

also maximizes opportunities for interviewer±interviewee

interaction and can, thus, contribute to the development

of rapport so that respondents keep talking (Oppenheim

1992) and reduce the possibility of respondents providing

socially desirable answers (Patton 1990). As with any

social interaction, people involved in an interview have

expectations of each other which can be in¯uenced by

such characteristics as age, gender and ethnicity (Bailey

1987). The research team knew that avoiding such

differences between interviewer and interviewee in this

study would not be possible. Therefore, building a rapport

and creating a supportive environment within each

interview was particularly important.

The development of the interview toolThe ®rst stage in the development of the interview

schedule concentrated upon identifying the topic areas

which needed to be covered in each interview based upon

the reviewed literature. The second stage focused upon

question formation when decisions were made about

question content, wording and structure. The wording of

each question was informed by the research team's famil-

iarity with the specialized vocabulary frequently used by

nurses and what seemed an appropriate use of language in

the light of the divergent sample group and was reviewed

for clarity at several stages prior to the main study. By

re®ning the schedule during pre-pilot and pilot work it

was hoped that the schedule used to obtain data from

respondents during the main study would be a robust tool

for eliciting complete data from all respondents.

Interviewer trainingThe development of effective interviewing skills was the

focus of interviewer training which began prior to the

commencement of pre-pilot work. Pilot work was under-

taken: to examine the effectiveness of the interview

schedule as a tool for eliciting information regarding

respondents' views and perceptions of continuing profes-

sional education; and to provide the interviewers with an

opportunity to continue with rehearsal and re®nement of

their interview technique and their accuracy and con®-

dence in using the interview schedule.

The development of the questionnaire schedulesAt the beginning of each interview respondents were

asked to complete a questionnaire requesting details about

their demographic, professional and employment pro®les

which was tested during pilot work and assessed for

clarity and ease of use. It is acknowledged that a potential

weakness of questionnaires is that it is more dif®cult to

ensure that all questions are answered by respondents, as

researchers have less control over the process of data

collection (Bailey 1987). However, it was anticipated that

requesting respondents to complete the questionnaires at

the time of being interviewed would increase the response

rate and reduce the incidence of item non-response since

missing data could be checked at the time of completion.

While each non-respondent was also asked to complete a

questionnaire in person, some non-respondents preferred

to return their questionnaires to the research team at a

later date. Although such ¯exibility had the potential to

increase the level of missing data, particularly through

non-completion, it was considered that this was the best

strategy in the light of this group's history of not

co-operating in the main study.

THE STUDY'S FINDINGS

As part of establishing the external validity of the data the

research team investigated the representativeness of the

sample and the potential impact of unit and item non-

response.

The sampling frame

A number of measures have been developed to estimate

the extent of non-coverage and address the potential bias

which the exclusion of individuals from the sampling

frame may introduce (e.g. population-based weighting

methods, see Elliot 1991). While such detailed and com-

plex analysis of non-coverage was beyond the resources

available to the research team, it was felt that some

evaluation of non-coverage was possible.

The sampling procedures used generated a list of 497

potential respondents with different demographic,

professional and employment pro®les (Table 1). Just over

one-third of potential respondents were non-Caucasian

(n� 168, 34á4%) and over a quarter were nurses' aides

(n� 128, 26á1%). Of the potential respondents with at

least one professional quali®cation, one-®fth were second

level nurses (n� 75) and almost a third held junior (i.e. D

or C grades) positions (n� 107, 30á1%). Nearly a quarter of

individuals included in the sampling frame worked night

or evening shifts only (n� 116, 23á7%) and almost a third

worked part-time hours (n� 155, 32á0%).

Respondents and non-respondents

While face-to-face contact was made with all those

included in the sampling frame (n� 497), one-tenth of

the target population (n� 48) refused to co-operate in the



main study. However, through re-sampling (i.e. obtaining

information from non-respondents, see Hartman et al.

1985) personal data and information about patterns of

continuing professional education activity were

successfully obtained from 43 non-respondents. Five

non-respondents did not return completed questionnaires.

Analysis of the data using the chi-square test revealed

that there were no statistically signi®cant differences

between respondents and non-respondents regarding

quali®ed status, level of registration, clinical grade, the

type of duty hours worked or employment status, although

nurses' aides, second level nurses and staff working part-

time hours were more likely to refuse to be interviewed

(see Table 2). However, there was a statistically signi®cant

difference between the continuing professional education

activities of non-respondents and respondents (P < 0á01).

One-third of non-respondents (n� 13) had not attended a

continuing professional education event in the 3 years

prior to the study, compared with only 16á1% of the

respondent group (n� 71).

Collecting the data

Since other questions answered by respondents may

provide considerable information, it may be possible to

compensate for item non-response by inserting values for

missing responses (Elliot 1991). As it was not possible for

the research team to predict how respondents would

respond to the topics under investigation (i.e. their views

and perceptions of continuing professional education), an

open question format was used throughout much of the

interview schedule. It was considered therefore that using

a method of imputation to compensate for item non-

response would be unreliable and that a more appropriate

approach would be the examination of interview data for

patterns of item non-response during data collection.

Inevitably during a large study there will be some loss of

data through item non-response. However, the

examination of the transcripts and/or audio-recordings of

interviews revealed that this was not a result of the

interviewers accidentally omitting to ask questions but

due to refusals by some respondents to answer particular

questions. An examination of the data set also revealed

that no single question appeared to attract a higher rate of

item non-response although there seemed to be a higher

incidence among some respondents working in particular

clinical units selected for inclusion in the study, namely

nurses' aides working in units caring for the mentally ill.

While information about all the topics included in the

interview schedule was obtained from the majority of

respondents and the views of different `groups' of respon-

dents are well represented, missing information even

when small threatens the validity of data (Moser & Kalton

1971).

Although the scale of missing data from the question-

naires was small (Table 1), Table 2 demonstrates that the

occurrence of item non-response was greatest among non-

respondents. While it would have been possible to insert

values for some missing data (e.g. ethnic background)

based on the knowledge the nurse researchers had

acquired about both respondents and non-respondents

while `in the ®eld', it was decided that the advantages of

compensation would be small in the light of the volume of

data lost. In addition, the insertion of some missing values

could not be undertaken reliably and therefore obtaining

complete data sets on each respondent and non-respon-

dent for whom information was missing would not have

been possible.

DISCUSSION

As research has shown that different `groups' of nursing

staff gain differential access to continuing professional

Table 1 The demographic, professional and employment pro®les

of units in the sampling frame (n = 497)

Sampling frame (%)

Ethnic background

Caucasian 321 (65á6)

Non-Caucasian 168 (34á4)

n = 489

Missing data = 8

Quali®ed status

Quali®ed 363 (73á9)

Unquali®ed 128 (26á1)

n = 491

Missing data = 6

Level of registration (n = 369)*

First level 285 (79á2)

Second level 75 (20á8)

n = 360

Missing data = 9

Clinical grade (n = 369)*

G grade or higher 111 (31á3)

F or E grades 137 (38á6)

D or C grades 107 (30á1)

n = 355

Missing data = 14

Duty hours worked

Days shifts only 373 (76á3)

Night or evening shifts 116 (23á7)

n = 489

Missing data = 8

Employment status

Full-time 330 (68á0)

Part-time 155 (32á0)

n = 485

Missing data = 12

*Question only applicable to quali®ed practioners.



education, the recruitment of a heterogeneous sample was

important to secure the generalizability of the study's

®ndings. The use of duty rotas and a high level of

face-to-face contact with potential respondents ensured

that the majority of those included in the sampling frame,

which contained individuals with divergent demographic,

professional and employment pro®les, were included in

the main study. While visually comparing the

characteristics of potential respondents with other sources

of information does not measure rigorously the extent of

non-coverage bias, external information about different

population groups is often available (e.g. Department of

Health 1995) and successfully provided the researchers in

this study with one indication of the representativeness of

the sample.

The importance of the methods of contact and data

collection in reducing refusals and non-contact was fully

considered by the research team. The ®ndings demon-

strate the success of using personalized letters and the

personal interview, maintaining an accurate index of

sample recruitment and ¯exibility during data collection,

in minimizing response failure, since personal contact

was made with all individuals included in the sampling

frame and refusals were limited to 10% (n� 48) of the total

sample. The analysis of the differences and similarities

between respondents and non-respondents show that

non-CPE attendees were more likely to refuse to be

interviewed. This con®rms Clarke and Rees' (1989) sus-

picions that nursing staff with little or no experience of

continuing professional education are under-represented

Table 2 A comparison of respondents and non-respondents (n = 492)*

Respondents

in the inteview

(%)

Non-respondents

in the inteview

(%)

Sample total

(%)

Level of

non-participation

(%) Statistical result

Quali®ed status

Quali®ed 337 (75á1) 26 (61á9) 363 (73á9) 7á2Unquali®ed 112 (24á9) 16 (38á1) 128 (26á1) 12á5

n = 449 42 491 ± v2 = 2á79789, 1dáf. P > 0á09

Missing data = 0 Missing data = 1 Missing data = 1

Level of registration

(n = 369)**

First level 269 (80á1) 16 (66á7) 285 (79á2) 5á6Second level 67 (19á9) 8 (33á3) 75 (20á8) 10á7

n = 336 24 360 ± v2 = 0á41680, 1dáf. P > 0á52


Clinical grade (n = 369)**

G grade or higher 107 (32á1) 4 (18á2) 111 (31á3) 3á6F or E grades 125 (37á5) 12 (54á5) 137 (38á6) 8á8D or C grades 101 (30á3) 6 (27á3) 107 (30á1) 5á6

n = 333 22 355 ± v2 = 2á89530, 2dáf. P > 0á24


Duty hours worked

Day shifts only 342 (76á3) 31 (75á6) 373 (76á3) 8á3Night or evening shifts 106 (23á7) 10 (24á4) 116 (23á7) 8á6

n = 448 41 489 ± v2 = 0á00000, 1dáf. P = 1á00


Employment status

Full-time 307 (69á1) 23 (56á1) 330 (68á0) 7á0Part-time 137 (30á9) 18 (43á9) 155 (32á0) 11á6

n = 444 41 485 ± v2 = 2á36869, 1dáf. P > á12


CPE attended during the

last 3 years

Yes 370 (83á9) 26 (66á7) 396 (82á5) 6á6No 71 (16á1) 13 (33á3) 84 (17á5) 15á5

n = 441 39 480 ± v2 = 6á73218, 1dáf. P < 0á01


*Excluding non-respondents who did not return completed questionnaires (n = 5).

**Question only applicable to quali®ed practioners.



in many UK studies and suggests that the level of

continuing professional education activity among British

nurses may be lower than that reported in the literature.

In addition to establishing the direction and extent of

bias, the data generated through re-sampling also proved

to be a valuable tool for evaluating the success of strategies

designed to reduce refusals. The trend in part-time staff

being more likely to refuse to be interviewed is disap-

pointing in the light of the effort made by the two nurse

researchers to accommodate the clinical responsibilities

and shift patterns of potential respondents by interview-

ing at weekends, late into the evening and during the night

shift. Nonetheless, it is likely that several more part-time

staff agreed to be interviewed than would have been the

case in the absence of such strategies.

While the level of data loss during interviewing was

not large, it is important to acknowledge that a pattern of

item non-response did emerge during the collection of

data. The preparation of interviewers for data collection

had focused upon managing the dif®culties which may

be encountered during interviewing a divergent sample

group. However, variables which can in¯uence the

performance or motivation of respondents to be `good'

informants may be beyond the control of researchers.

Uncertainty about the effects of planned or current

changes within the organization of clinical units, the

possibility of hospital and unit closures and the threat of

redundancy and new skill mixes were the reasons given

by some in the sampling frame for refusing to be

interviewed. These factors may also have contributed to

why some respondents did not wish to answer certain

questions during the interviews. In addition, researchers

have to accept that there are certain topics about which

respondents may not wish to talk even when these topics

do not appear particularly sensitive. It can be argued that

recurrent refusals by respondents to answer questions

may re¯ect that an interviewer has failed to gain the trust

of interviewees; however, it can also be argued that when

respondents feel able to be honest and indicate that they

do not wish to answer a particular question but are

happy to continue with the interview that a good

relationship between interviewer and interviewee has

developed.

Regarding the questionnaires, the higher level of miss-

ing values in the data collected from non-respondents may

indicate that this group consisted of nursing staff

generally less disposed to co-operating with the study.

Although the questionnaires had been tested for clarity

Figure 2 Explanatory model of data loss in survey research.



and ease of use during pilot work, it may also be an

indication of the bene®ts gained by having researchers `on

hand' to assist with queries which may arise during the

collection of data. Permitting non-respondents to

complete questionnaires in their own time was a strategy

adopted to encourage co-operation and was largely

successful since only ®ve non-respondents did not return

their questionnaires and the possible disadvantages of this

strategy need to be considered in the light of the potential

data yield generated through this approach.

Application of model of non-responsein survey research

Figure 1 attributed data loss in surveys to different stages

of the research process using a linear conceptualization of

non-response. However, the ®ndings from this study

demonstrate that such an approach is too fragmentary

and fails to capture the inter-relationship and

inter-dependence of strategies which can minimize non-

response phenomena. In the study, for example, decisions

made at sample selection directly impacted upon the

success of reducing non-response during sample recruit-

ment. Similarly, generating interest in the study and

establishing rapport with the study sample were not only

invaluable for minimizing refusals during sample recruit-

ment but also for encouraging respondent co-operation

during the collection of data.

To ensure the success of strategies designed to reduce

non-response in survey research, the complex relation-

ships that exist between non-response phenomena must

be addressed. To this end, a more integrated conceptual-

ization of non-response which acknowledges the impact

of measures to manage response failure across different

stages of the research process is necessary. Figure 2

presents an explanatory model of data loss in survey

research which illustrates the interdependence of non-

response strategies and identi®es the extent to which each

strategy can be manipulated. It is important to acknowl-

edge that there are some factors which contribute to the

incidence of non-response which are not amenable to

direct control by researchers (i.e. respondent issues).

However, adopting an integrated approach which

acknowledges that decisions made throughout the

research process have an accumulative effect upon the

different sources of non-response (i.e. non-coverage, unit

non-response and item non-response), it is possible to

reduce the loss of data due to respondent issues.

CONCLUSION

Non-response is an important issue in survey research

since it can compromize the validity of the data set. The

®ndings presented in this paper demonstrate that it is

possible for researchers to minimize non-response and

that an integrated approach to the management of

response failure can be effective. A linear model is rejected

as a useful framework since it belies the complexity of

non-response phenomena and may encourage the

researcher to neglect an ongoing and proactive approach

to minimizing data loss throughout the study.

Acknowledgements

Many thanks to Sara Christian for her help during data

collection. This study was funded by South-West Thames

Regional Health Authority (now South Thames Regional

Health Authority (West)).

References

Aiken L.R. (1988) The problem of nonresponse in survey research.

Journal of Experimental Education 56(3), 116±119.

Bailey K.D. (1987) Methods of Social Research, 3rd edn. The Free

Press, New York.

Bradburn N.M. & Sudman S. (1979) Improving Interview Method

and Questionnaire Design. Josey Bass, San Francisco.

Buchan J. & Seccombe I. (1991) Nurses' Work and Worth. Pay,

Careers and Working Patterns of Quali®ed Nurses. A Review for

the Royal College of Nursing. Report No. 213. Institute of

Manpower Studies, Brighton.

Burns N. & Grove S.K. (1987) The Practice of Nursing Research:

Conduct, Critique and Utilisation. Saunders, Philadelphia.

Cannell C.F. & Kahn R.L. (1968) Interviewing. In The Handbook of

Social Psychology, vol. 2, 2nd edn (Lindzey G. & Aronson E.

eds), Addison-Wesley, Reading, Massachusetts, pp. 526±595.

Carr E.K. (1988) feeding dif®culties after stroke: their assessment

and nursing management. Unpublished PhD thesis, University

of Nottingham, Nottingham.

Cartwright A. (1978) Professionals as responders: variations in

and effects of response rates to questionnaires 1961±77. British

Medical Journal 2, 1419±1421.

Champion D.J. & Sears A.M. (1969) Questionnaire response rate:

a methodological analysis. Social Forces 47, 335±339.

Chapman R.W. & Hall B.V. (1992) Post-Registration Education in

three Fields of Clinical Practice. Centre for Social and Evalu-

ation Research, City of London Polytechnic, London.

Clarke J. & Rees C. (1989) The midwife and continuing education.

Midwives Chronicle 102, 288±290.

Denzin N.K. (1989) The Research Act: A Theoretical Introduction

to Sociological Methods, 3rd edn. Prentice Hall, Englewood

Cliffs, New Jersey.

Department of Health (1995) NHS Workforce in England (1982±

92). Her Majesty's Stationary Of®ce, London.

Dillman D.A. & Frey J.H. (1974) Contribution of personalisation to

mail questionnaire response as an element of a previously

tested method. Journal of Applied Psychology 59(3), 297±301.

Elliot D. (1991) Weighting for Non-response. Of®ce for Population

Census and Surveys, London.

Gordon R.L. (1975) Interviewing: Strategy, Techniques and

Tactics. Dorsey Press, Illinois.

Goyder J. (1985) Face-to-face interviews and mailed question-

naires: the net difference in response. Public Opinion Quarterly

49, 234±252.



Hartman B.W., Fuqua D.R. & Jenkins S.J. (1985) The problems of

and remedies for nonresponse bias in educational surveys.

Journal of Experimental Education 54, 85±90.

Heberlein T.A. & Baumgartner R. (1978) Factors affecting

response rates to mailed questionnaires: a quantitative

analysis of the published literature. American Sociological

Review 43(4), 447±462.

Hox J.S & De Leeuw E. (1994) A comparison of nonresponse in

mail, telephone and face-to-face surveys. Quality and Quantity

28, 329±344.

Hutchinson S. & Skodol-Wilson H. (1992) Validity threats in

scheduled semi-structured research interviews. Nursing

Research 41(2), 117±119.

Kidder L. (1981) Sellitz, Wrightsman and Cook's Research Meth-

ods in Social Relations. Holt, Rinehart & Winston, New York.

Kviz F.J. (1977) Toward a standard de®nition of response rate.

Public Opinion Quarterly 41, 265±267.

Larcombe K. & Maggs C. (1991) Processes for Identifying the

Continuing Professional Education Needs of Nurses, Midwives

and Health Visitors: An Evaluation. English National Board for

Nursing, Midwifery and Health Visiting, London.

Marshall S.L. & While A.E. (1994) Interviewing respondents who

have English as a second language: challenges encountered and

suggestions for other researchers. Journal of Advanced Nursing

19, 566±571.

Miller D.C. (1991) Guides to Methods and Techniques of Collect-

ing Data in Library, Field and Laboratory. Sage, Newbury Park,

California.

Moser C.A. & Kalton G. (1971) Survey Methods in Social Inves-

tigation, 2nd edn. Heinemann, London.

Nay-Brock R.M. (1984) A comparison of the questionnaire and

interviewing techniques in the collection of sociological data.

Australian Journal of Advanced Nursing 2(1), 14±23.

Oppenheim A.N. (1992) Questionnaire Design, Interviewing and

Attitude Measurement. Pinter, London.

Patton M.Q. (1990) Qualitative Evaluation and Research Methods,

2nd edn. Sage, Newbury Park, California.

Rashid C. (1992) Patchy provision. Nursing Times 88(13),

38±40.

Rosenthal R. & Rosnow R.L. (1969) The volunteer subject. In

Artefact in Behavioural Research (Rosenthal R. & Rosnow R.L.

eds), Academic Press, London, pp. 59±99.

Smith L. (1992) Ethical issues in interviewing. Journal of

Advanced Nursing 17, 98±103.

Steeh C.G. (1981) Trends in non-response rates (1952±79). Public

Opinion Quarterly 45, 40±57.

Treece E.W. & Treece J.W. (1986) Elements of Research in

Nursing, 4th edn. C.V. Mosby, St Louis.

Williamson Y.M. (1981) Research Methodology and its Applica-

tion to Nursing. John Wiley & Sons, New York.



Non���response in survey research: a methodological discussion and development of an explanatory model

Documents

Non��response in survey research: a methodological discussion and development of an explanatory model