Published on in Vol 19, No 3 (2017): March

Costs and Efficiency of Online and Offline Recruitment Methods: A Web-Based Cohort Study

Costs and Efficiency of Online and Offline Recruitment Methods: A Web-Based Cohort Study

Costs and Efficiency of Online and Offline Recruitment Methods: A Web-Based Cohort Study

Original Paper

1Aarhus University Hospital, Department of Clinical Epidemiology, Aarhus N, Denmark

2Boston University School of Public Health, Department of Epidemiology, Boston, MA, United States

3Aarhus University, Research Unit for General Practice, Department of Public Health, Aarhus C, Denmark

4RTI Health Solutions, Research Triangle Park, NC, United States

Corresponding Author:

Tina Christensen, MA

Aarhus University Hospital

Department of Clinical Epidemiology

Olof Palmes Allé 43-45

Aarhus N, 8200


Phone: 45 87168225

Fax:45 87167215


Background: The Internet is widely used to conduct research studies on health issues. Many different methods are used to recruit participants for such studies, but little is known about how various recruitment methods compare in terms of efficiency and costs.

Objective: The aim of our study was to compare online and offline recruitment methods for Internet-based studies in terms of efficiency (number of recruited participants) and costs per participant.

Methods: We employed several online and offline recruitment methods to enroll 18- to 45-year-old women in an Internet-based Danish prospective cohort study on fertility. Offline methods included press releases, posters, and flyers. Online methods comprised advertisements placed on five different websites, including Facebook and We defined seven categories of mutually exclusive recruitment methods and used electronic tracking via unique Uniform Resource Locator (URL) and self-reported data to identify the recruitment method for each participant. For each method, we calculated the average cost per participant and efficiency, that is, the total number of recruited participants.

Results: We recruited 8252 study participants. Of these, 534 were excluded as they could not be assigned to a specific recruitment method. The final study population included 7724 participants, of whom 803 (10.4%) were recruited by offline methods, 3985 (51.6%) by online methods, 2382 (30.8%) by online methods not initiated by us, and 554 (7.2%) by other methods. Overall, the average cost per participant was €6.22 for online methods initiated by us versus €9.06 for offline methods. Costs per participant ranged from €2.74 to €105.53 for online methods and from €0 to €67.50 for offline methods. Lowest average costs per participant were for those recruited from (€2.99) and from Facebook (€3.44).

Conclusions: In our Internet-based cohort study, online recruitment methods were superior to offline methods in terms of efficiency (total number of participants enrolled). The average cost per recruited participant was also lower for online than for offline methods, although costs varied greatly among both online and offline recruitment methods. We observed a decrease in the efficiency of some online recruitment methods over time, suggesting that it may be optimal to adopt multiple online methods.

J Med Internet Res 2017;19(3):e58



Recruiting participants for epidemiologic research is increasingly difficult as the number of projects competing for people’s attention increases and response rates decline [1]. In cohort studies, participant recruitment and data collection are associated with a heavy workload and high costs [2]. Widespread access to the Internet now offers an alternative strategy to recruit participants into cohort studies and to collect data. The Internet offers technical advantages in data collection that can reduce administrative procedures and improve data quality. Examples are incorporating skip patterns that avoid displaying irrelevant questions, building in internal consistency checks, and avoiding errors that occur during manual data entry [3]. Furthermore, the Internet is an effective tool to reach populations who are otherwise challenging to enroll because they have sensitive health concerns, including urinary incontinence, sexual health disorders, or mental health problems [4-6].

Pregnancy planners also constitute a hard-to-reach population as they do not typically announce their pregnancy intentions [7-10]. Among recruitment methods that have proven to be effective in enrolling participants are study-related media publicity [6,11], online advertisements [6,7,10,12-15], printed advertisements [6], and Web-based social groups [6,10].

Some studies have reported the costs of a single online recruitment method [14,16-19], and others have reported on costs of using several online methods [20,21], or both offline and online methods [6,10,12,22,23]. Nevertheless, little is known about how recruitment methods compare in terms of efficiency (number of recruited participants) and costs per participant [20]. Here, we compare the efficiency and costs per recruited participant of online and offline recruitment methods used to enroll women in a Danish cohort study of pregnancy planners that relied on the Internet for enrollment and data collection.


The (Soon-Pregnant) and the (SoonParents) studies are related prospective cohort studies on lifestyle and fertility [8,9,24,25]. In both studies, participants enroll via the Internet and all data are collected by means of Web-based questionnaires. was launched in June 2007. It was succeeded in August 2011 by, which incorporates a dietary questionnaire and includes male partners (Figure 1). Recruitment for is ongoing [26].

The study period for this paper, which focuses on female recruitment in both studies (in the following referred to as one study), is June 2007 through December 2013.

Figure 1. The frontpage of the website.

Study Population

The recruitment area for and covers all of Denmark (total population of 5.7 million on January 1, 2016, and 58,205 births in 2015) [27]. Eligible female participants have a Danish civil registration number, are aged 18-45 years old (18-40 years in, in a stable relationship with a male partner, and attempting to become pregnant. To enroll, participants have to (1) visit the study website, which contains information on eligibility criteria and informed consent and (2) complete a Web-based screening questionnaire to determine eligibility. As part of this process, participants provide their civil registration number, which is a unique, personal 10-digit number assigned to all Danish residents at birth or upon immigration. No incentives are offered to participants, who are followed until they achieve a pregnancy, start fertility treatment, or end of observation (12 cycles). In both studies, enrollment is based on completion of a comprehensive baseline questionnaire, which takes an average of 22 minutes [28].

A substudy,, was introduced to pilot test the feasibility of collecting blood and urine samples to investigate associations between endocrine disruptors and fertility. Participation in the substudy required enrollment in and residence in the Aalborg area (population of 205,407 in December 2013) [29].

Recruitment Methods

During the study period, we employed several online and offline recruitment methods to increase awareness about the study and to encourage enrollment. All advertisements were designed by the same graphics designer and used images, colors, and text phrases identical with or similar to the respective study website to enhance recognition. The most frequently used text was “Planning to get pregnant? Help us find out whether lifestyle influences your ability to become pregnant.”

Offline Recruitment Methods

Three offline recruitment methods were used to attract attention to our study: press releases, posters, and flyers.

A total of 6 press releases with various topics were issued; of which, 3 press releases were written and disseminated with the assistance of contracted, external journalists, 1 was handled by study staff alone, and 2 were issued with help from an internal journalist from Aarhus University. We obtained information from media surveillance companies on the numbers of printed and Web-based articles, and radio or TV features that the press releases resulted in.

The costs of obtaining the services of external journalists and a media surveillance company were €2281.29 and €1761.91 for the first 2 press releases, respectively, and €1772.54 for the fourth press release. Issuing the other press releases involved no direct costs as internal institutional resources were used.

A small A4-size (210x297 mm) poster with information about the study was designed (Figure 2). Each copy was equipped with a block of 50 post-it notes providing study information. The poster included a quick response (QR) code providing a direct link to the study website when scanned with a smart phone. Study staff and colleagues placed 133 posters on notice boards in libraries, hospital canteens, fitness centers, grocery stores, and other public places. In total, €377.76 was paid for the design of the poster, the QR-code, and the post-it notes and for printing.

Flyers advertising, and the substudy, were distributed to pharmacies, a few stores, and 54 general practitioners’ (GP) offices in the city of Aalborg (Figure 3). Costs for designing the flyer and printing 3500 copies were €604.31.

Figure 2. Poster advertising
Figure 3. Flyer advertising and the substudy
Online Recruitment Methods

Online campaigns, primarily consisting of online advertisements, were placed on 6 different websites:,, Facebook,,, and Aarhus University (AU). When participants clicked on an advertisement, they were automatically redirected to the study website. is a popular general health website. Two online advertisements of different sizes were posted 2 weeks after the launch of (Figure 4). agreed to show the advertisement until it had generated 2500 participants at a fixed price of €3.93 each. For the study (Figure 5), committed itself to recruit 10,000 women at a fixed price of €2.99 per enrolled participant. controlled the timing of advertisements to be able to prioritize campaigns by full-price advertisers.

Facebook is a popular social media website. In total, 8 advertisements, lasting 8-10 days each, were posted on Facebook. The advertisements were targeted at potential participants, that is, female, 18-40 years old, speaking Danish, and married or in a relationship. Seven advertisements targeted women living anywhere in Denmark and one targeted only women living in the city of Aalborg.

Advertising on Facebook is managed on the Internet and the price is based on a bidding system. Depending on the number of advertisers aiming at the same target group, the system suggests a price range, within which it is likely that your advertisement will be shown. It is possible to pay per 1000 impressions, that is, the number of times an advertisement is shown on the site, or per click and to choose a daily spending limit. We paid per click and chose a limit of €13.33 per day for 8-10 days at a time. The study staff also created a page on Facebook to promote the study. Postings were made by staff and thus incurred no direct costs. is a general health website. We made an agreement including 100,000 impressions of banner advertisements at an overall fixed price of €1077 over a period of 8 months. is a website for pregnant women, that is, women who want to become pregnant and women with infants. We negotiated an agreement covering 5 one-month campaigns, each consisting of 500,000 impressions of banner advertisements, 50,000 pop up/overlay banner impressions within the category “Fertility,” and short texts in 3 electronic newsletters distributed to subscribers. An overall fixed price of €2019 was paid for each of the 5 campaigns. is a website aimed at parents of infants, pregnant women, and women wanting to become pregnant. We made an agreement including continuous display of a top banner advertisement and a continuous “ownership” picture on a fertility page for 3 months, as well as placement of 2 stories about our study on the website at an overall fixed price of €673.

The Aarhus University Communications Department placed 2 short postings about, including links to the study website, on the AU intranet, which reaches approximately 38,000 students and on the AU Facebook page. The postings were free of charge.

Figure 4. Online advertisement for posted on
Figure 5. Online advertisement for posted on

Assessment of Number of Participants Recruited by Each Method

We defined seven categories of mutually exclusive recruitment methods: online advertisements, press releases, posters, flyers, SnartForaeldre homepage, other homepages, and “other.” We used two data sources, electronic tracking and self-reported data, to determine the recruitment method for each participant. The number of participants enrolled as a result of online advertisements was obtained by electronically tracking participants’ click on the advertisements via Uniform Resource Locator (URL) using different URL codes for each online campaign. URL tracking overruled any self-reported response by the participants. Facebook provided detailed statistics regarding the number of impressions and the number of clicks for each advertisement placed on its website. In addition, we calculated a click-through rate, that is, the number of times someone clicked on an advertisement divided by the number of times it was shown, and a conversion rate, that is, the number of people who enrolled after having clicked on the advertisement (and thus registered by its unique URL) divided by the total number of people who clicked on the advertisement.

The remaining participants were categorized according to their response to the following question on the baseline questionnaire: “How did you hear about the present study?” The response options were: “SnartForaeldre website,” “Facebook,” “Netdoktor,” “Other websites or blogs or chatrooms,” “Poster,” “Flyer from my GP,” “Radio,” “Television or teletext,” “Newspapers,” “Weekly magazines,” “Previously participated in study,” and “Word of mouth,” as well as an open-ended “Other.” It was possible to mark multiple response options. Any open-ended response that could be linked unambiguously to an online advertisement was categorized as such.

The number of women who enrolled as a result of press releases was determined by identifying those who, within a period of 2 weeks after a press release, marked at least one of the following responses: “Other homepages,” “Radio,” “TV or teletext,” “Newspapers,” “Weekly magazines,” or “Other magazines,” without marking any other answers. In a sensitivity analysis, we expanded the time period to 4 weeks.

In a subanalysis, we estimated the average number of participants recruited per 30 days for each online recruitment method.

Finally, we excluded women who did not respond to the question “How did you hear about the study?” or who outside a press release window marked more than one answer. The reason for these exclusions was the inability to assign respondents to one specific recruitment category.


For each recruitment method, costs per enrolled participant were calculated by dividing the total direct expenses (which did not include salary expenses for study staff) by the number of participants recruited from that method. For the recruitment methods “SnartForaeldre homepage,” “other homepages,” and “other,” costs per enrolled participant were not calculated as these methods were free of charge. All costs are reported in euros, applying exchange rates as of March 2012.

Overall Recruitment

In total, 8258 participants were recruited for our Web-based cohort study during the study period (Figure 6). We excluded 534 (6.5%) participants from this analysis as they either did not answer the question on how they heard about the study (n=19) or they provided more than one answer (515 respondents, among whom 452 provided two answers, 59 provided three answers, and 4 gave four answers). Thus, the final study population consisted of 7724 participants. Of these, 3985 (51.6%) were recruited by online advertisements, 803 (10.4%) were recruited by offline methods, and 2382 (30.8%) were recruited from two other online methods, that is, “SnartForaeldre homepage” and “other homepages” (see Figure 6). Other methods generated 554 participants (7.2%), of whom 271 (48.9%) learned about the study from word-of-mouth, whereas 222 (40.1%) heard about the study through radio, TV, newspapers, or magazines outside the time window of a press release. Of the 3985 participants recruited from online advertisements, 3866 were tracked electronically and 119 were categorized based on self-reported data.

Figure 6. Number of participants by recruitment method.

Offline Recruitment Methods

For the offline methods, six press releases yielded 747 participants in total. The press releases generated between 8 and 81 articles/features and recruited between 21 and 506 participants at a cost of €0 to €24.28 per participant. Three press releases were issued without external assistance and were not assigned a price (Table 1). The sensitivity analysis in which the time window of the definition of the “press release recruitment” was changed to 4 instead of 2 weeks did not change the cost per participant for the last four press releases. However, for the press release “Launch of” the average cost per participant decreased from €4.51 to €3.35 and for the second press release “Funding received to continue study,” the average cost declined from €18.35 to €11.82. Distribution of 133 posters resulted in 47 participants at a cost of €8.04 each, whereas the distribution of 3500 flyers for display led to 9 participants, each costing €67.50 (Table 1). The average cost per participant recruited via offline recruitment methods initiated by study staff was €9.06.

Table 1. Costs and number of participants recruited by offline recruitment methods.
Average costs per participant
Press releases

Launch of Snart-Gravid.dk35 articles/features5064.51

Funding received to continue study50 articles/features9618.35

Investigating impact of H1N1 influenza8 articles/features280.00

Launch of SnartForaeldre.dk58 articles/features7324.28

Results: physical activity84 articles/features230.00

Results: oral contraceptives81 articles/features210.00
Posters133 placed in canteens,
supermarkets, and so on
Flyers3500 distributed to GPa offices,
pharmacies, and so on
All offline methods

aGP: general practitioners.

Online Recruitment Methods

Online methods comprised online campaigns including banner advertisements on various websites. A total of 4 years and 7 months of frequent advertisements on generated 2912 participants, each at a fixed price of €3.93 in and €2.99 in Factoring in 59 participants who stated that they heard about the study from, but who were not tracked electronically, the costs were €2.74 per participant in (Table 2). The average number of participants recruited via per 30 days was 73 in and 27 in

Table 2. Costs and number of participants recruited by online recruitment methods.
Online advertising campaignsCharacteristicsParticipants
Average costs per participant
TrackedTracked per 30 daysSelf-reportedTrackedTracked and self-reported

Snart-Gravid.dkOne campaigna, 31 months22737303.93aN/Ab

SnartForældre.dkOne campaigna, 24 months63927592.99a2.74
Minmave.dkFive campaigns, 1 month each689138714.9314.78

October 1, 2012212

January 20, 2013158

March 1, 2013142

May 14, 201388

August 31, 201389
Sundhedsguiden.dkOne campaign, 8 months1111115.12105.53
Baby.dkOne campaign, 3 months196245.4041.08
AU.dkPostings on AUc intranet and AU Facebook page4N/A00.00N/A
FacebookSeven nationwide advertisements, 8-10 days each230121503.442.83
All online methods

aNegotiated, fixed price per recruited participant.

bN/A: not available.

cAU: Aarhus University.

Five one-month campaigns on resulted in 689 participants at an average cost of €14.93 each. Factoring in 7 participants who stated that they heard about the study from, but who were not tracked by a URL code, the average cost was €14.78 each. However, the number of recruited participants decreased from 212 participants in the first campaign to 89 in the last campaign. Consequently, the corresponding cost per recruited participant increased from €9.70 to €23.11 in the course of the campaigns (Table 2). On average, 138 participants were recruited from per 30 days.

An advertisement on yielded 11 participants at €115.12 each over 8 months, whereas a 3-month campaign on generated 19 participants at €45.40 each. When we included participants who reported that they knew about the study from these websites but were not tracked, the cost per participant was €105.53 for and €41.08 for (Table 2).

Seven nationwide advertising periods of 8-10 days on Facebook yielded 230 participants at an average cost of €3.44 each (Table 2). Factoring in 50 participants who stated that they heard about the study from Facebook, but who were not tracked electronically, the average cost per participant was €2.83. It was found that one Facebook advertisement targeted only at women living in Aalborg, that is, a small target population, resulted in 1 participant at the cost of €21.94 (Table 3). In terms of individual nationwide Facebook advertisements, the number of recruited participants varied between 22 and 84 for the first four advertisements, whereas the number was 11 or less for the last three advertisements. The average cost per participant ranged from €1.32 to €22.18. The lowest number of impressions was 16,776 and the highest was 409,129. The click-through rate varied between 0.13% and 1.85%. The conversion rate steadily decreased from 12.07% for the first Facebook advertisement to 1.61% for the last. The average number of participants recruited from Facebook per 30 days was 121.

The average cost per participant recruited via any online recruitment method initiated by study staff was €6.22 for those tracked electronically compared with an average of €9.06 for offline methods.

Table 3. Costs and number of participants tracked by Uniform Resource Locator codes in advertisements on Facebook.
Target areaTimeImpressionsa
Click-through ratec
Conversion rated
Average costs per participant

NationwideMarch 30 to April 7,
NationwideJune 29 to July 7,
NationwideNovember 25 to
December 2, 2010
NationwideSeptember 22-29,
AalborgOctober 9-16,
NationwideMay 3-12,
NationwideOctober 31 to
November 9, 2013
NationwideDecember 21-28,
in total


aThe number of times an advertisement was shown.

bThe number of people who clicked on the advertisement.

cThe proportion of people who saw the advertisement and clicked on it.

dThe proportion of people who clicked on the advertisement and were enrolled.

Principal Findings

In our study population of 7724 pregnancy planners recruited into a prospective cohort study, the majority (6367 participants, 82.4%) were recruited using online recruitment methods. Offline recruitment methods and other methods accounted for 10.4% (803/7724) and 7.2% (554/7724) of participants, respectively. Overall, the average cost per participant was €6.22 for online methods initiated by study staff versus €9.06 for offline methods. Though online methods were the most efficient recruitment tools, costs ranged from €2.74 to €105.53 per participant. Costs per participant for offline methods ranged from €0 (press release issued with institutional assistance, free of charge) to €67.50. We were able to monitor efficiency (the total number of recruited participants), over time for Facebook and, and found that efficiency decreased while cost per participant increased over time.


We lacked access to information on the number and timing of impressions on Thus, we were unable to trace any fluctuation in efficiency over time of a prime recruitment method—the advertisements on We can only speculate whether the efficiency of these advertisements decreased over time as was the case with our advertisements on Facebook and the campaigns on

The average number of participants recruited per 30 days from online campaigns varied from 1 ( to 138 ( and Facebook yielded the highest average number of participants per 30 days. However, the number of impressions for was unknown and it varied both across and within the other recruitment methods, limiting the comparability of the numbers.

The possible imprecision of the estimated efficiency of press releases is also a concern. The number of recruited participants yielded by this method was defined according to a given time period, which may have been too short. As indicated in our sensitivity analysis, more women may have been recruited from this method than we calculated, and thus, we may have overestimated the average cost per participant.

Another limitation is the possibility that participants who reported that they learned about the study from “Other home pages” (2281 participants, 29.5%) may have seen an online advertisement but not clicked on it. Instead they may have signed up later directly via the study website. Thus, the number of participants recruited by each online advertisement may have been underestimated.

Our exclusion of 515 women who marked more than one answer to the question “How did you hear about the study?” may also be viewed as a study limitation, as it may have led to overestimation of the cost per participant. However, had these women been assigned to two or more categories of recruitment methods, they would have been counted twice and costs would have been underestimated.

Comparison With Prior Work

Online Methods

Our special agreement with involved paying a fixed price of less than €4 per recruited participant and resulted in 2971 participants during the study period. Such an agreement seems unusual. However, Ramo et al [20] reported a similar arrangement with a Web-based sampling service that sent out email invitations to people signed up as volunteers for completing Web-based surveys. They paid a fixed price of US $19.24 (€14.71) per completed survey and achieved 67 in 6 months. Thus, their price was much higher than the price we negotiated with

Compared with other studies, the overall average cost of €6.22 per participant recruited by online methods in our study is low. Graham et al [21] reported an average cost per registrant of US $209.34 (€160.06) (ranging from US $73.76 (€56.40) to US $4166.67 (€3185.84)) in their study on the effectiveness of advertisements placed on four different Spanish-language websites to recruit Latinos for a Web-based smoking cessation program. Ramo et al [20] paid US $42.77 (€32.70) per participant recruited through an Internet advertising campaign using various social networking and lifestyle-based websites.

For our North American sister study, PRESTO, Wise et al [10] reported costs of US $43.00 to $96.15 (€32.88 to €73.52) per participant for those recruited via online advertisements placed on websites other than Facebook. This agrees with the costs of the less efficient websites used for advertising our study, which cost up to €115 per participant.

The efficiency of advertisements on Facebook varied greatly in our study and decreased over time. As the number of impressions decreased markedly throughout the study period, the advertisements were shown fewer times at the same cost. Interestingly, the click-through rate was considerably higher for the last two advertisements (1.47% and 1.85%) compared with the first six (0.05-0.23%). However, this did not translate into more participants as the conversion rate decreased throughout the study period. Still, the average cost per participant recruited from Facebook advertisements was relatively low at €3.44.

Ramo and Prochaska [13] compared 20 different advertisements placed on Facebook over a 13-month period to recruit 18- to 25-year-olds for a survey on the use of tobacco and other substances. Efficiency varied widely across the advertisements. In addition, both the number of impressions and the number of clicks increased over time, but not concomitantly. Ramo et al paid an average cost of US $4.28 (€3.27) per completed survey, similar to our costs.

In comparison with the average cost of €3.44 per recruited woman via Facebook in our study, costs ranging from €6.73 to €19.48 have been reported in studies on nutrition [14,30], general health [19], smoking cessation [31], and childbirth preferences [7]. Our sister study, PRESTO, and another study by Richiardi et al, with a topic and target group similar to ours, paid US $27.77 (€21) and €25, respectively, per woman recruited from Facebook advertisements [10,17].

In contrast to our results, Thornton et al [12] reported costs of US $1.86 (€1.42) per participant recruited by Facebook advertisements for a survey about tobacco, cannabis and alcohol use, and mental health, whereas Nelson et al [15] paid US $1.36 (€1.04) per participant who completed a Web-based survey on human papillomavirus (HPV) vaccine uptake.

Bearing in mind that the price per click varies according to the Facebook bidding system and the number of advertisers opting for the same target group, costs from the various studies using Facebook depend either on the target country or area, and may not be directly comparable. The criteria by which Facebook advertisements can be tailored to reach specific target populations also vary from one country to another. Wise et al [10] reported that they used “newlyweds” as a target criterion for the PRESTO study in the United States, whereas this criterion was not available on Facebook in Denmark during our study period. Thus, the efficiency of Facebook advertisements may be influenced by how precisely the advertisement can be targeted at potential participants.

Another factor influencing comparability is the use of incentives for enrolling. Announcing incentives before enrollment may make more people want to sign up for the study in question [12,22]. Among other studies using Facebook advertisements, all used incentives [7,10,12-15,19,30], except for one [17]. Apart from the PRESTO study by Wise et al [10], which offered incentives for completing a required number of follow-up questionnaires, incentives were offered for enrollment and thus could have affected people’s inclination to participate. Some studies have reported the use of incentives worth US $15 (€11.46) [30] and US $20 (€15.28) [15] per participant. One could argue that these costs should be included in the overall recruitment cost per participant. However, this would make it more difficult to compare costs directly.

Some studies using online enrollment methods have reported problems with duplicate entries, in particular studies offering incentives [15,20,30]. We did not offer incentives and our study participants had to provide their unique 10-digit civil registration number to enroll, thus ruling out the possibility of multiple entries. The fact that multiple entries into the study was impossible is a major strength of our study and means that the reported number of participants recruited by each method reflects the actual number of distinct eligible individuals responding to the advertisement.

It may be a concern that online recruitment may not yield a sample which is representative of the background population. However, this concern is of relevance in a cross-sectional study that aims to estimate the prevalence of a disease or a risk factor in a given population at a given time. We do not recruit participants to a cross-sectional study, but into a cohort, where we follow our participants until they achieve a pregnancy, start fertility treatment, or end of observation (12 cycles). On the basis of comparisons within the cohort, we estimate associations between various exposures and time to pregnancy. Some differences are likely to exist between our population and all Danish women trying to conceive, but these differences are unlikely to affect the internal comparisons made within our cohort. In addition, in a recent validation study, we showed that the selection bias was not a major concern in the associations under study [25].

The focus of our study on fertility could explain why our recruitment methods proved more efficient compared with some other studies. We had the advantage of an interesting topic—the expectations associated with possible, impending motherhood. Harris et al speculated that the topic of their study on contraception and pregnancy plans motivated women to enroll. Although their study offered incentives, they found that Facebook advertisements not mentioning the incentive, but rather focusing on generating new scientific knowledge to benefit other women, were highly efficient [6].

Similar to other studies that recruit hard-to-reach populations [13,15], we found that online recruitment methods were indeed a viable method to reach our target population.

As was the case with Facebook, the efficiency of the online campaigns on decreased over time. The average cost per participant thus increased from €9.70 to €23.11. A possible explanation could be saturation of the target population on Another explanation could be the increasing tendency toward the use of smaller devices, like tablets and smart phones, for accessing the Internet [18]. However, due to the length of our questionnaire and the comprehensiveness of the questions, it would be cumbersome to complete it on a mobile phone.

Offline Methods

The offline methods used in our study were less efficient than the online methods in terms of total number of enrolled participants. In a French Web-based study on nutrition, Kesse-Guyot et al [11] included offline methods to recruit participants and found that only 0.61% and 1.69% of participants learned about the study from posters and flyers, respectively. Our findings are similar, as only 0.61% (47) and 0.12% (9) of our participants reported hearing about the study from posters and flyers, respectively. Kesse-Guyot et al also reported varying effects of their mass-media campaigns, which were launched with a press release and a press conference. They experienced the highest peak in enrollment after the first campaign or press release. This was also the case in our study.

When Wise et al [10] used flyers as a recruitment method, they gained 46 participants (out of 2421 in total) at a cost of US $12 (€9.18) each, whereas postcards led to 15 participants costing US $87 (€66.52) each. Thus, their experience with offline methods is similar to ours in terms of poor efficiency and relatively high costs per participant.

In general, studies employing both offline and online methods to recruit participants for Web-based data collection have found offline methods to be less efficient [6,10,22,23]. This difference may stem from the extra participant effort required by offline methods. After seeing a poster in, for example, the gym, you would have to go online and find the study homepage to enroll. In contrast, when an advertisement is provided on the Internet, enrollment is only a few clicks away [21,23]. It should also be kept in mind that offline methods are more labor-intensive for the research team [7]. This factor was not taken into account in the costs reported in our study. In addition, compared with offline methods, online methods seem more familiar to our young target population, which increasingly conducts daily communications [31].


We were able to recruit large numbers of women for our Web-based prospective cohort study on fertility. We found that online recruitment methods were superior to offline methods in terms of efficiency (total number of participants enrolled). Both online and offline costs per enrolled participant showed great internal variation. However, given the higher workload associated with offline methods and the lower efficiency, online methods appear the most appealing. The study topic and the use of incentives are likely to influence the efficiency of recruitment methods and the fact that efficiency of online methods may decrease over time suggests that it may be optimal to use multiple online methods.


The and studies were supported by the NICHD (R21050264), the Danish Medical Research Council (271-07-0338), and the NIH (R01 HD060680-01). The sponsors had no influence on the study in any way.

Authors' Contributions

All authors contributed to the design of the study. TC wrote the first and successive drafts of the paper. AHR, EM, and TC carried out the statistical analyses. All authors contributed to the interpretation of results, reviewed and approved the final manuscript.

Conflicts of Interest

None declared.

  1. van Gelder MM, Bretveld RW, Roeleveld N. Web-based questionnaires: the future in epidemiology? Am J Epidemiol 2010 Dec 1;172(11):1292-1298 [] [CrossRef] [Medline]
  2. Olsen J, Melbye M, Olsen SF, Sørensen TI, Aaby P, Andersen AM, et al. The Danish National Birth Cohort--its background, structure and aim. Scand J Public Health 2001 Dec;29(4):300-307 [Medline]
  3. Ekman A, Litton J. New times, new needs; e-epidemiology. Eur J Epidemiol 2007;22(5):285-292 [CrossRef] [Medline]
  4. Sjöström M, Umefjord G, Stenlund H, Carlbring P, Andersson G, Samuelsson E. Internet-based treatment of stress urinary incontinence: a randomised controlled study with focus on pelvic floor muscle training. BJU Int 2013 Jan 25;112(3):362-372 [CrossRef]
  5. Cook C. Email interviewing: generating data with a vulnerable population. J Adv Nurs 2012 Jun;68(6):1330-1339 [CrossRef] [Medline]
  6. Harris ML, Loxton D, Wigginton B, Lucke JC. Recruiting online: lessons from a longitudinal survey of contraception and pregnancy intentions of young Australian women. Am J Epidemiol 2015 May 15;181(10):737-746 [CrossRef] [Medline]
  7. Arcia A. Facebook Advertisements for Inexpensive Participant Recruitment Among Women in Early Pregnancy. Health Educ Behav 2014 Jun;41(3):237-241 [] [CrossRef] [Medline]
  8. Huybrechts KF, Mikkelsen EM, Christensen T, Riis AH, Hatch EE, Wise LA, et al. A successful implementation of e-epidemiology: the Danish pregnancy planning study 'Snart-Gravid'. Eur J Epidemiol 2010 May;25(5):297-304 [] [CrossRef] [Medline]
  9. Mikkelsen EM, Hatch EE, Wise LA, Rothman KJ, Riis A, Sørensen HT. Cohort profile: the Danish Web-based Pregnancy Planning Study--'Snart-Gravid'. Int J Epidemiol 2009 Aug;38(4):938-943 [] [CrossRef] [Medline]
  10. Wise LA, Rothman KJ, Mikkelsen EM, Stanford JB, Wesselink AK, McKinnon C, et al. Design and Conduct of an Internet-Based Preconception Cohort Study in North America: Pregnancy Study Online. Paediatr Perinat Epidemiol 2015 Jul;29(4):360-371 [] [CrossRef] [Medline]
  11. Kesse-Guyot E, Andreeva V, Castetbon K, Vernay M, Touvier M, Méjean C, et al. Participant profiles according to recruitment source in a large Web-based prospective study: experience from the Nutrinet-Santé study. J Med Internet Res 2013 Sep;15(9):e205 [] [CrossRef] [Medline]
  12. Thornton LK, Harris K, Baker A, Johnson M, Kay-Lambkin FJ. Recruiting for addiction research via Facebook. Drug Alcohol Rev 2016 Jul;35(4):494-502 [CrossRef] [Medline]
  13. Ramo DE, Prochaska JJ. Broad reach and targeted recruitment using Facebook for an online survey of young adult substance use. J Med Internet Res 2012;14(1):e28 [] [CrossRef] [Medline]
  14. Lohse B. Facebook is an effective strategy to recruit low-income women to online nutrition education. J Nutr Educ Behav 2013;45(1):69-76 [CrossRef] [Medline]
  15. Nelson EJ, Hughes J, Oakes JM, Pankow JS, Kulasingam SL. Estimation of geographic variation in human papillomavirus vaccine uptake in men and women: an online survey using facebook recruitment. J Med Internet Res 2014;16(9):e198 [] [CrossRef] [Medline]
  16. Cooper CP, Williams KN, Carey KA, Fowler CS, Frank M, Gelb CA. Advertising campaign on a major internet search engine to promote colorectal cancer screening. BMJ 2004 May 15;328(7449):1179-1180 [] [CrossRef] [Medline]
  17. Richiardi L, Pivetta E, Merletti F. Recruiting study participants through Facebook. Epidemiology 2012 Jan;23(1):175 [CrossRef] [Medline]
  18. Kapp JM, Peters C, Oliver DP. Research recruitment using Facebook advertising: big potential, big challenges. J Cancer Educ 2013 Mar;28(1):134-137 [CrossRef] [Medline]
  19. Fenner Y, Garland SM, Moore EE, Jayasinghe Y, Fletcher A, Tabrizi SN, et al. Web-based recruiting for health research using a social networking site: an exploratory study. J Med Internet Res 2012;14(1):e20 [] [CrossRef] [Medline]
  20. Ramo DE, Hall SM, Prochaska JJ. Reaching young adult smokers through the internet: comparison of three recruitment mechanisms. Nicotine Tob Res 2010 Jul;12(7):768-775 [] [CrossRef] [Medline]
  21. Graham AL, Fang Y, Moreno JL, Streiff SL, Villegas J, Muñoz RF, et al. Online advertising to reach and recruit Latino smokers to an internet cessation program: impact and costs. J Med Internet Res 2012;14(4):e116 [] [CrossRef] [Medline]
  22. Loxton D, Powers J, Anderson AE, Townsend N, Harris ML, Tuckerman R, et al. Online and Offline Recruitment of Young Women for a Longitudinal Health Survey: Findings From the Australian Longitudinal Study on Women's Health 1989-95 Cohort. J Med Internet Res 2015;17(5):e109 [] [CrossRef] [Medline]
  23. Wiist WH, Sullivan BM, Wayment HA, Warren M. A web-based survey of the relationship between buddhist religious practices, health, and psychological characteristics: research methods and preliminary results. J Relig Health 2010 Mar;49(1):18-31 [CrossRef] [Medline]
  24. Wise LA, Rothman KJ, Mikkelsen EM, Sørensen HT, Riis A, Hatch EE. An internet-based prospective study of body size and time-to-pregnancy. Hum Reprod 2010 Jan;25(1):253-264 [] [CrossRef] [Medline]
  25. Hatch EE, Hahn KA, Wise LA, Mikkelsen EM, Kumar R, Fox MP, et al. Evaluation of Selection Bias in an Internet-based Study of Pregnancy Planners. Epidemiology 2016 Jan;27(1):98-104 [] [CrossRef] [Medline]
  26. SnartForaeldre. URL: [accessed 2016-09-09] [WebCite Cache]
  27. DST. Statistics Denmark, Statistical Year Book URL: [WebCite Cache]
  28. Rothman KJ, Mikkelsen EM, Riis A, Sørensen HT, Wise LA, Hatch EE. Randomized trial of questionnaire length. Epidemiology 2009 Jan;20(1):154 [CrossRef] [Medline]
  29. Aalborg Municipality. Statistics URL: http:/​/apps.​​statistik/​webaarbog/​Folketal2014/​struktur/​Hele_Kommunen/​indexlevel1/​HeleKommunen1aarsaldersklasser.​html [WebCite Cache]
  30. Lohse B, Wamboldt P. Purposive facebook recruitment endows cost-effective nutrition education program evaluation. JMIR Res Protoc 2013;2(2):e27 [] [CrossRef] [Medline]
  31. Ramo DE, Rodriguez TM, Chavez K, Sommer MJ, Prochaska JJ. Facebook Recruitment of Young Adult Smokers for a Cessation Trial: Methods, Metrics, and Lessons Learned. Internet Interv 2014 Apr;1(2):58-64 [CrossRef] [Medline]

AU: Aarhus University
GP: general practitioner
HPV: human papillomavirus
QR: quick response
URL: Uniform Resource Locator

Edited by G Eysenbach; submitted 29.09.16; peer-reviewed by K Harris, E Nelson; comments to author 03.11.16; revised version received 16.12.16; accepted 31.01.17; published 01.03.17


©Tina Christensen, Anders H Riis, Elizabeth E Hatch, Lauren A Wise, Marie G Nielsen, Kenneth J Rothman, Henrik Toft Sørensen, Ellen M Mikkelsen. Originally published in the Journal of Medical Internet Research (, 01.03.2017.

This is an open-access article distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on, as well as this copyright and license information must be included.