Health Education Research, Vol. 19, No. 2, 159-164,
April 1, 2004
© 2004 Oxford University Press
Introductory postcards: do they increase response rate in a telephone survey of older persons?
1 WA Centre for Health Promotion Research, Curtin University of Technology, GPO Box U1987, Perth, Western Australia 6845, Australia 2 Correspondence to: H. Iredell; e-mail: h.iredell{at}curtin.edu.au
 |
Abstract |
|---|
 Top Abstract IntroductionMethodResultsDiscussionReferences |
|---|
The purpose of this study was to assess whether contact by mail via an introductory postcard would increase the response rate to a telephone survey in a selected population of 348 people 60 years and older. Of this group, 208 were sent a postcard introducing the study and informing them that someone would call in the next 1–2 weeks to conduct a telephone interview. The remaining sample of 140 did not receive the introductory postcard. Amongst the potential participants who were contacted and able to participate (n = 240), the response rate for completed interview was 58.0% for those sent a postcard and 42.3% for those who were not. The refusal rate was 25.3% (postcard sent) and 37.2% (not sent a postcard). When the outcome of contact is collapsed to two categories, ‘agreed to participate’ and ‘refusal’, the response rate to participate was 73.5% (sent a postcard) and 59.0% (not sent a postcard) (
2 5.14, d.f. 1, sign. 0.023). The response rate amongst those who recall being sent a postcard rose to 86.0%. Whether agreeing to participate or refusing to be involved in the survey was dependent on being sent an introductory postcard. The conclusion is that mailing a postcard prior to the first telephone contact increases the participation rates of older persons in a telephone survey at very little extra cost (A$0.60 per person). |
Introduction |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Telephone interviewing has become a popular method for data collection in public health research. Many studies of older persons also utilize telephone interviews to investigate varying health issues (McFall and Yerkes, 1998
; Kim, 1999; Kempton et al., 2000; Stevens et al., 2001). Telephone surveys are appealing due to their ability to achieve a cost-efficient and amenable method of delivery at the community level, with the expectation of reasonable response rates (Smith et al., 1995). A criticism of telephone interviews is the potential for selection bias, particularly in populations with poor telephone coverage (Smith et al., 1995). However, in Australia this is not a major problem as it has high telephone coverage, with 94.4% of the population having a telephone in their home (Australian Bureau of Statistics, 1991).
To ensure the validity of any research conducted by survey, the response rate needs to be high enough that the sample truly reflects the characteristics of the target population (Smith et al., 1995; Robertson et al., 2000). Additionally, the recruitment method chosen should be easily applied from one study to another (Smith et al., 1995), be efficient and effective, and minimize delays or difficulties in achieving a sufficient number of participants (Taylor-Davis et al., 1998).
Recruitment methods that utilize a combination of two or more methods enhance recruitment efforts (Taylor-Davis et al., 1998). However, enhancement of response rates using a mixed-mode strategy (mail and telephone) is under-reported in the literature (Smith et al., 1995). Those studies that have reported on strategies to achieve high response rates have reported on the general adult population (Smith et al., 1995; Robertson et al., 2000) and not on specific subgroups such as older persons. Furthermore, they are often based on strategies reported in clinical trials and epidemiological research, which has limited relevance to community-based research (Taylor-Davis et al., 1998).
Robertson et al. (Robertson et al., 2000) conducted a study where an introductory letter was sent 1–2 weeks before making telephone contact. They reported an enhancement of the participation rate of controls by 29%. The interviewers’ feedback indicated that ‘cold calling’ was a major reason for refusals. Smith et al. (Smith et al., 1995) conducted a randomized trial with half their sample receiving an introductory letter before telephone contact and half receiving no advance letter. Their response rate was 76% amongst those who received the letter compared to 60% for those who did not. Both studies concluded that an introductory letter sent in advance to the target population increases participation rates significantly at little extra cost. Other studies have also found that the recruitment of older participants, via an introductory mailed letter and telephone follow-up, was a viable approach and an important factor in obtaining higher participation rates (Taylor-Davis et al., 1998; Sharkey and Haines, 2001).
Dillman (Dillman, 1978) advocates sending an advance letter notifying potential participants of the impending telephone call:
It not only eliminates the element of surprise, but it also provides tangible evidence that the interviewer is legitimate and the telephone call is neither a sales gimmick nor a practical joke. [(Dillman, 1978
People are very skeptical of unexpected telephone calls, particularly when they are accompanied with a request to be interviewed. Consequently, they may be guarded in their responses which can reduce the quality of the answers given to the interviewer (Dillman, 1978).
The purpose of this study was to assess whether contact by mail, via an introductory postcard, before contact by telephone would increase the response rate to a telephone survey, in a selected population of people 60 years and older.
|
Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
A sample of 403 names were randomly selected from the Australian electoral rolls, from three local government authorities in the Perth Metropolitan Area for a pilot telephone survey. The purpose of the survey was to investigate older persons’ perceptions and behavior regarding crossing roads. From these names, a matching telephone number could not be traced for 55 people (13.6%). Reasons for not matching the telephone number may be due to a recent move, a silent number or living with relatives. No contact was attempted for those people. This left 348 potential participants with matching telephone numbers (Table I).
|
Of these 348 people, a random selection of 208 were sent a postcard introducing the study, and informing them that someone would be calling in the next 1–2 weeks to conduct a telephone interview (Table I). The introductory postcard contained information about the purpose of the study, emphasizing the importance of the respondent’s participation and outlining the extent of their involvement. The card invited potential participants to contact the survey coordinator if they had any questions about the research and it pointed out they could arrange an alternative time for the interview if they were called at an inconvenient time. The remaining sample was not sent an introductory postcard (n = 140).
In the subsequent interviews, on establishing contact with the potential participant who had been sent a postcard, the interviewer asked whether they had received a postcard about the study. If they reported having not received the postcard, the interviewer explained that the postcard was to advise them of this phone call and the purpose of the survey. Those who were not sent a postcard did not receive this statement; the interviewer directly introduced the study to the potential participant. The three trained interviewers were given a selection of potential interviewees from each of the three localities.
|
Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Of the 348 names that were selected and matched to a telephone number, 258 could be contacted by telephone (74.1%). The maximum of five callbacks was made with six potential participants. However, there were a number of people who were called unsuccessfully between 1 and 4 times (n = 31). Attempts to contact them ceased when the study had reached the predetermined quota of completed interviews. Additionally, for some potential participants there was no attempt at contacting them at all (n = 53) because the quota had been achieved. Thus, from the sample originally selected for the study, a total of 90 people could not be or were not contacted for an interview.
From these 258 people, 127 completed questionnaires were obtained (49.2%). As we wished to interview only those who reported crossing roads at least twice a week, some potential participants did not meet the screening criteria (n = 38) and the interview was not continued. Seventy people refused to participate in the interview. The outcome of the telephone contacts is given in Table II.
|
Amongst the potential participants who were contacted and able to participate (n = 240), the response rate for completed interview was 58.0% for those sent a postcard and 42.3% for those not sent a postcard. The refusal rate was 25.3% (postcard sent) and 37.2% (not sent a postcard) (Table III).
|
Those who agreed to participate in the survey were comprised of two groups—those who agreed to participate and completed the questionnaire, and those who agreed to participate, but did not meet the selection criteria. The other category was refusal, which was made up of those who outright refused to participate and those who started the questionnaire, but stopped the interview and did not complete. Therefore, when the outcome of contact with potential participants is collapsed to two categories, ‘agreed to participate’ and ‘refusal’, the response rate to participate is 73.5% for those sent a postcard and 59.0% for those not sent a postcard. The Pearson
2 is 5.14, d.f. 1, sign. 0.023, indicating there is a statistically significant difference between the groups (P < 0.05). Whether agreeing to participate or refusing to be involved in the survey was dependent on being sent an introductory postcard.
Not all those who were sent an introductory postcard could recall receiving it. Therefore, when the outcome of contact with potential participants is viewed from the perspective of whether they recall receiving the postcard, the response rate to participate is even higher (86.0%). For those who were sent a postcard, but could not recall receiving it, the response rate to participate is only 56.3%. Therefore agreeing to participate was not only associated with sending a postcard to potential participants, but on them recalling receiving it (P < 0.05) (Pearson 2 21.528, d.f. 2, sign. 0.000).
There were no significant age, gender or country of birth differences between participants who agreed to participate in the survey and those who refused. Nor were there was any significant differences in the characteristics of those who were sent a postcard and those who did not receive a postcard.
The costs associated with the introductory postcard include paper, printing, postage and labels. The total cost for 208 postcards was A$124.81, with an estimated cost of A$0.60 per person. If an illustration for the postcard had to be specifically designed this would be an additional approximate cost of A$150. This would have increased the total cost to A$274.81, with an estimated cost of A$1.32 per person. However, with larger numbers the illustration cost would not be as great per person and printing would also be at a cheaper rate per person.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
An introductory postcard prior to the first telephone contact increases the participation rates of older persons in a telephone survey at very little extra cost. The response rate to the telephone-administered questionnaire was significantly higher among the groups of people who were sent an introductory postcard before telephone contact. Their response rate was 73.5% compared to 59.0% among those not receiving an introductory postcard. The response rate amongst those who also recall being sent the introductory postcard rose to 86.0%. The overall response rate was 68.8%.
While survey methodology textbooks advocate some type of introductory approach to potential participants, our experience has found that at times shortcuts are taken. We found that people were generally more supportive towards our study because we had given them prior notice that we would be calling them.
Mailing an introductory postcard was a relatively straightforward and low cost procedure (A$0.60 per person), with the card being sent out 1–2 weeks prior to the telephone contact. Robertson et al. (Robertson et al., 2000) suggests that it is advisable to post the cards between 7 and 14 days prior to the proposed interview date to prevent participants receiving the advance postcard too early prior to the initial telephone contact.
These findings are applicable to other studies recruiting older people for a telephone-administered survey. High refusal rates can invalidate the results due to selection bias and lead to an excessively long recruitment process. The resulting additional costs are likely to be far greater than the relatively small cost of sending out a postcard.
|
Acknowledgements |
|---|
The authors would like to thank Chantelle Jeffery, who was very caring and professional in interviewing the participants; the older persons who kindly agreed to take part in our study; and the Health Promotion Foundation of Western Australia and Main Roads WA, who provided funding for the project.
|
|
|
References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionReferences |
|---|
Australian Bureau of Statistics (1991) Household telephone Connection, Australia. ABS Catalogue no. 4110.0. Australian Government Publishing Service, Canberra.
Dillman, D.A. (1978) Mail and Telephone Surveys: The Total Design Method. Wiley, New York.
Kempton, A., Van Beurden, E., Sladden, T., Garner, E. and Beard, J. (2000) Older people can stay on their feet: final results of a community-based falls prevention programme. Health Promotion International, 15, 27–33.
Kim, O. (1999) Predictors of loneliness in elderly Korean women living in the United States of America. Journal of Advanced Nursing, 29, 1082–1088.[CrossRef][Web of Science][Medline]
McFall, S.L. and Yerkes, A.M. (1998) Targets, messages and channels for educational interventions on urinary incontinence. Journal of Applied Gerontology, 17, 403–418.
Robertson, B., Sinclair, M. and Forbes, A. (2000) The effect of an introductory letter on participation rates using telephone recruitment. Australian and New Zealand Journal of Public Health, 24, 552.[Web of Science][Medline]
Sharkey, J.R. and Haines, P.S. (2001) The feasibility of using a telephone-administered survey for determining nutritional service needs of noninstitutionalised older adults in rural areas: time and costs. The Gerontologist, 41, 546–552.
Smith, W., Chey, T., Jalaludin, B., Salkeld, G. and Capon, T. (1995) Increasing response rates in telephone surveys: a randomized trial. Journal of Public Health Medicine, 17, 33–38.
Stevens, M., Holman, C.D.J., Bennett, N. and de Klerk, N. (2001) Preventing falls in older people: outcome evaluation of a randomized controlled trial. Journal of the American Geriatrics Society, 49, 1448–1455.[CrossRef][Web of Science][Medline]
Taylor-Davis, S., Smiciklas-Wright, H., Davis, A.C., Jensen, G. and Mitchell, D. (1998) Time and cost for recruiting older adults. Journal of the American Geriatrics Society, 46, 753–757.[Web of Science][Medline]
Received on September 8, 2002; accepted on January 23, 2003
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  E. D. Leeuw, M. Callegaro, J. Hox, E. Korendijk, and G. Lensvelt-Mulders The Influence of Advance Letters on Response in Telephone Surveys: A Meta-Analysis Public Opin Q, September 1, 2007; 71(3): 413 - 443. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||