TY - JOUR
AU - Meldener,, Vanessa
AB - Abstract In mail surveys and in advance letters for surveys in other modes, it is common to include a prepaid incentive of a small amount such as $5. However, when letters are addressed generically (such as to “Resident”), advance letters may be thrown away without being opened, so the enclosed cash is wasted and the invitation or advance letter is ineffective. This research note describes results of an experiment using a nationally representative sample of 4,725 residential addresses to test a new way of letting mail recipients know their letter contains cash and is therefore worth opening: an envelope with a window revealing $5, so the cash is clearly visible from outside the sealed envelope. We also tested the USPS for evidence of theft, and we compared First Class and Priority Mail postage. We found no evidence of theft. We found no difference in response rates between Priority Mail and First Class, making First Class much more cost-effective, and we found that visible money increased the response rate to a mail survey from 42.6 to 46.9 percent, at no significant cost. 1. INTRODUCTION Survey invitation letters often include small amounts of cash, such as $1, $2, or $5, as prepaid respondent incentives. Such incentives prepaid by mail improve response rates to surveys conducted by mail, telephone, in person, and on the internet (e.g., Singer and Ye 2013; Mercer, Caporaso, Cantor, and Townsend 2015). Interviewers anecdotally report that some respondents do not remember receiving advance letters sent to them that contained cash. Letters may be considered junk mail and may be discarded unopened, with recipients never realizing there was cash inside and never having considered whether to participate in the survey. Cash is obviously wasted if respondents never see it. Various studies have tested envelope designs intended to encourage respondents to open their mail and discover the cash hidden inside, usually without success. Messages printed on the envelope indicating money is enclosed have been ineffective (Dykema, Jaques, Cyffka, Assad, Hammers et al. 2015). Compared with simple business envelopes, the use of large logos, printed messages, and colored envelopes has been ineffective or has even depressed response rates (Dykema et al. 2015; Jans, Park, Rauch, Grant, and Edwards 2015; Williams and Gentry 2017), and the prevailing advice remains that general population survey invitations should be mailed in plain envelopes with, at most, a respected logo as part of the return address (Dillman 2007). This has left survey researchers in the lamentable position of mailing prepaid cash incentives that some intended recipients never see. 2. THEORY For a mail survey response to occur, a chain of five events must take place: the respondent must (1) receive the envelope, (2) open the envelope, (3) choose to respond to the study, (4) complete the questionnaire, and (5) return the questionnaire (Dykema et al. 2015). Prepaid incentives normally take effect at the third stage to motivate the respondent to read and accept the request and follow through on subsequent steps. But the chain may often break at the second stage. It is likely that at least one third of direct mail is thrown away unopened (InfoTrends 2016), and anecdotally, survey invitation letters are sometimes not opened. If prospective respondents could be convinced that survey invitation letters contain money before they decide whether to open the envelope, two helpful things might happen. First, obviously, the incentive would encourage the recipient to open the envelope, which would cut down on the number of invitations that are thrown away without being considered. Second, there are theoretical reasons to expect the incentive to be more influential on cooperation when the respondent becomes aware of the incentive before opening the envelope. Leverage-saliency theory (Groves, Singer, and Corning 2000) supposes that the decision to participate in a survey is a function of multiple factors that are weighted by the degree to which they are made salient during a survey invitation. Making prospective respondents aware of a real cash incentive before they encounter the survey request would increase the salience of the incentive and, therefore, increase the likelihood of cooperation. When a letter, a questionnaire, an incentive, and other materials are all encountered at the same time upon opening an envelope, they compete for attention. When one element of the invitation, such as the incentive, is singled out for presentation before anything else, it necessarily takes on greater salience and has the advantage of being processed into memory in the absence of competing or interfering information (e.g., Waugh and Norman 1965; Greene, Prepscius, and Levy, 2000; Murphy, Hofacker, and Mizerski, 2006). Such first impressions can be persistent and can affect the way the rest of a message is perceived (e.g., Nisbett and Ross 1980, Chapter 8; Tetlock 1983). By presenting cash first, survey researchers would create a more positive initial impression than by presenting an envelope that otherwise looks like it could be junk mail. The elaboration likelihood model of persuasion (Petty and Cacioppo 1986) holds that there are two paths of persuasion. The “central route” involves thoughtful consideration of the merits. The “peripheral route” involves reliance on cues to form an attitude without carefully engaging all the facts. The low-effort peripheral route is used most of the time because people usually lack either the ability or motivation to use the more demanding central route. Someone invited to participate in a survey is likely to use the peripheral route, and peripheral route thinkers look for shortcuts such as first impressions and how they are made to feel. According to Petty et al., the elaboration-likelihood model suggests that “under low elaboration conditions, people will engage in some low-effort strategy that could include processing early rather than late arguments” (Petty, Wheeler, and Bizer 1999, p. 160). Putting the incentive first, ahead of the invitation and questionnaire, would make no difference for careful (central route) thinkers but could gain considerable leverage with peripheral route thinkers by giving them a positive first impression and making them feel good. Based on these theoretical perspectives, if we could make respondents believe envelopes contained cash, they would be more likely to open the envelopes, and they would be more likely to encounter the questionnaire and the request for survey participation while in a positive frame of mind and, therefore, be more likely to respond to the survey. 3. HYPOTHESES To increase the likelihood that respondents would see and be influenced by prepaid mailed cash incentives, we tested a novel mailing approach: an envelope with a large window through which $5 cash was plainly visible. We hypothesized that the response rate to a mail survey that included a $5 cash incentive would be higher when the $5 was visible through the envelope window than when the $5 was placed behind the letter so it was not visible from outside the envelope. Additionally, we tested whether the response rate would be higher for invitations sent by Priority Mail than by First Class mail. Although other experiments have found response rates no better with Priority Mail than First Class (Brems et al. 2006; Mamedova & McPhee 2012), the possibility that letters with visible cash could be stolen motivated us to test a tracked form of mailing, such as Priority Mail. Finally, we tested the hypotheses that the manipulations of cash visibility and postage would not have large effects on the composition of the responding sample. 4. METHOD The sample for this study was a subset of the sample for American National Election Studies (ANES) 2016 Time Series Study. The ANES used two independently drawn address-based samples from the US Postal Service Delivery Sequence File and supplied by Marketing Systems Group. One was a simple random sample of 7,800 residential mailing addresses from all fifty states and the District of Columbia (DC), excluding drop point addresses. The other was a stratified cluster sample of 2,880 addresses in the 48 contiguous United States and DC. For details of the ANES sample, see DeBell, Amsbary, Meldener, Brock, and Maisel (2018). The present study was part of a nonresponse follow-up study (NRFU) to the ANES (DeBell and Maisel 2018), and the dataset is the ANES 2016 Methodology Dataset (American National Election Studies 2018). In this dataset, weights for the NRFU cases were provided only for NRFU respondents. We calculated NRFU weights for nonrespondents as well, applying the same method used by ANES to the nonresponding cases: the analysis weight is the product of the NRFU selection weight and the household weights (the variables weight_ftfbwt0 for the face-to-face sample and weight_hhbwt_o for the internet sample). Weighted results of the NRFU are representative of the US adult citizen population. The NRFU was mailed to a probability sample of 4,725 addresses, of which 2,379 had responded to the ANES and 2,346 had been invited but never responded to ANES. Original ANES interviews occurred between September 7 and November 7, 2016, and most respondents were interviewed a second time between November 9, 2016, and January 8, 2017. The NRFU occurred from March 17 through August 23, 2017, and was not presented to respondents as a follow-up to the ANES; while the ANES was presented as the “American National Election Study” conducted by Stanford University, the University of Michigan, and Westat and sponsored by the National Science Foundation, the NRFU was branded as the “Study of Lifestyles and Attitudes” conducted by the Institute for Research in the Social Sciences in Palo Alto, California (which is a department of Stanford University). Addresses selected for the NRFU study received up to six mailings. All were addressed to “CITY Resident,” where “CITY” was the city in which the address was located. The first mailing was an advance postcard announcing the study, including the statement that “In a few days you’ll receive a letter with the study and $5 enclosed.” The second mailing was an envelope that contained a letter, $5 cash, a one-page questionnaire, and a self-addressed #9 return envelope with a first-class stamp affixed. The letter randomly selected an individual by asking either the oldest or youngest male or female in the household to respond (Hagan and Collier 1983). The third mailing was a reminder postcard. The fourth mailing contained a letter, a replacement questionnaire, and a self-addressed #9 return envelope with a first-class stamp affixed (but no cash). The fifth mailing also contained a letter, a replacement questionnaire, and a self-addressed #9 return envelope with a first-class stamp affixed, but also stated that the study was “ending in a few days, on June 30,” and offered a postpaid incentive of $20 to complete the questionnaire. The sixth and final mailing was a reminder postcard. The six mailings were sent on March 10, March 21, March 28, April 10, June 21, and June 23, 2017, respectively. The second mailing was varied according to a two-by-two experiment design. One dimension was visible cash versus not visible cash. All respondents received $5 in an envelope with metered postage and a clear window that covered most of the front of the envelope. In the “visible cash” condition, the cash was paper-clipped to the front of the letter so that it was clearly visible from outside the sealed envelope. In the “not visible cash” condition, the cash was not clipped and was behind the letter so it could not be seen before opening the envelope, and there was no visible indication that money was enclosed. In this study, all mailings were prepared manually, and there was no cost difference to place the cash in a visible position, other than the cost of paper clips and a few seconds of labor per letter. The second dimension of the experiment was a comparison of First Class mail with Priority Mail. First Class letters were sent in a #10 full face window envelope. Priority Mail letters were sent in a five-by-ten-inch USPS Priority Mail Window Flat Rate Envelope (SKU EP 14-H) with a USPS tracking label. The visible cash experiment was powered to reliably detect a response rate difference of 4.3 points (using standard power assumptions of p < 0.05 and alpha = 0.80) between the experimental groups. (The experiment was not designed to detect response rate differences within subgroups of ANES respondents and nonrespondents, for whom reliably detectable differences would be about fifteen points for many demographic subgroups of respondents.) The postage experiment was powered to reliably detect a difference in undeliverable returned mail rates of about 1.7 points, assuming an average returned mail rate of about 4 percent. 5. RESULTS Of letters mailed to 4,725 addresses, 217 letters were returned as undeliverable and considered ineligible, one was determined to be an ineligible nonresidential address, thirty-two were explicit refusals, nine were implicit refusals in the form of a blank questionnaire returned, 2,026 completed questionnaires were returned, and the remaining 2,440 addresses did not respond. The overall unweighted response rate was 45.0 percent, and the weighted response rate was 44.8 percent. Randomization checks found that random assignment of respondents to experimental conditions performed as expected. For the entire sample, we compared experimental groups on dwelling type, having a telephone number match on the sample frame, and refusal to participate in the ANES study; for ANES respondents, we examined age, gender, race/ethnicity, educational attainment, party identification, having a smartphone, using the internet at home, and attention to politics. With a Bonferroni correction for multiple comparisons, no differences were statistically significant. Table 1 shows response rates by experimental condition. For the cash conditions, the visible cash response rate (46.9 percent by AAPOR response rate formula 1) is 4.3 points higher than the nonvisible response rate (42.6 percent) (t = 2.91, p = 0.004). The visible cash response rate was higher in both postage conditions, but only the 6.7 percentage point difference for First Class mail was significant (48.5 compared with 41.8; t = 3.21, p = 0.001). For the postage conditions, the response rates do not differ significantly between First Class and Priority Mail. Table 1. ANES NRFU Response Rates by Experiment Condition Cash condition Postage Visible cash Not visible cash Difference All cash Condition Percent s.e. n Percent s.e. n Percent s.e. Percent s.e. n First Class 48.5 1.49 1,119 41.8 1.47 1,133 6.7* 2.09 45.2 1.05 2,252 Priority mail 45.2 1.48 1,124 43.5 1.47 1,131 1.7 2.09 44.3 1.05 2,255 Difference 3.3 2.11 — −1.7 2.08 — — — 0.9 1.48 — All postage 46.9 1.05 2,243 42.6 1.04 2,264 4.3* 1.48 44.8 — 4,507 Cash condition Postage Visible cash Not visible cash Difference All cash Condition Percent s.e. n Percent s.e. n Percent s.e. Percent s.e. n First Class 48.5 1.49 1,119 41.8 1.47 1,133 6.7* 2.09 45.2 1.05 2,252 Priority mail 45.2 1.48 1,124 43.5 1.47 1,131 1.7 2.09 44.3 1.05 2,255 Difference 3.3 2.11 — −1.7 2.08 — — — 0.9 1.48 — All postage 46.9 1.05 2,243 42.6 1.04 2,264 4.3* 1.48 44.8 — 4,507 Note.—Percentages are weighted. —, not applicable; s.e., standard error; n, unweighted denominator for the given percent. * p < 0.01. Open in new tab Table 1. ANES NRFU Response Rates by Experiment Condition Cash condition Postage Visible cash Not visible cash Difference All cash Condition Percent s.e. n Percent s.e. n Percent s.e. Percent s.e. n First Class 48.5 1.49 1,119 41.8 1.47 1,133 6.7* 2.09 45.2 1.05 2,252 Priority mail 45.2 1.48 1,124 43.5 1.47 1,131 1.7 2.09 44.3 1.05 2,255 Difference 3.3 2.11 — −1.7 2.08 — — — 0.9 1.48 — All postage 46.9 1.05 2,243 42.6 1.04 2,264 4.3* 1.48 44.8 — 4,507 Cash condition Postage Visible cash Not visible cash Difference All cash Condition Percent s.e. n Percent s.e. n Percent s.e. Percent s.e. n First Class 48.5 1.49 1,119 41.8 1.47 1,133 6.7* 2.09 45.2 1.05 2,252 Priority mail 45.2 1.48 1,124 43.5 1.47 1,131 1.7 2.09 44.3 1.05 2,255 Difference 3.3 2.11 — −1.7 2.08 — — — 0.9 1.48 — All postage 46.9 1.05 2,243 42.6 1.04 2,264 4.3* 1.48 44.8 — 4,507 Note.—Percentages are weighted. —, not applicable; s.e., standard error; n, unweighted denominator for the given percent. * p < 0.01. Open in new tab The response rates among prior ANES respondents (n = 2,379) were 60.8 and 58.0 percent for visible and nonvisible cash, respectively (difference of 2.8, s.e. 2.05, t = 1.37, p = 0.17, not shown in table). The response rates among ANES nonrespondents (n = 2,346) were 32.5 percent and 27.2 percent for visible and nonvisible cash (difference of 5.3 points, s.e. 1.95, t = 2.72, p < 0.01, not shown in table). As noted previously, the experiment was not powered to detect response rate differences within subgroups (especially within subgroups of prior respondents and nonrespondents) unless they were extremely large (not only because subgroups of interest are often small but also because they are nested within ANES respondents or nonrespondents, reducing the numbers). As such, we do not report the nonsignificant differences within further subgroups of survey respondents and nonrespondents. It is worth emphasizing that the 5.3 percentage point increase in response rate among prior ANES nonrespondents shows that visible cash was effective in boosting response among the hardest-to-reach subset of the sample. Table 2 shows the percentage of mail pieces that were returned by the postal service as undeliverable, cross-tabulated by experiment conditions for cash visibility and type of postage. The return rate was 5.3 percent for visible cash and 4.5 percent for nonvisible cash. The return rate was 4.9 percent in both First Class and Priority Mail. All of the mail return rate differences were estimated to be less than one percentage point, and none were statistically significant. Table 2. ANES NRFU Undeliverable Returned Mail Rates by Experiment Condition Cash condition Postage Visible cash Not visible cash Difference All cash Condition Percent s.e. n Percent s.e. n Percent s.e. Percent s.e. n First Class 5.4 0.66 1,181 4.5 0.60 1,182 0.9 0.89 4.9 0.44 2,363 Priority mail 5.2 0.65 1,181 4.5 0.60 1,181 0.7 0.88 4.9 0.44 2,362 Difference 0.2 0.92 — 0.0 0.85 — — — 0.0 0.63 — All postage 5.3 0.46 2,362 4.5 0.43 2,363 0.8 0.63 44.8 — 4,725 Cash condition Postage Visible cash Not visible cash Difference All cash Condition Percent s.e. n Percent s.e. n Percent s.e. Percent s.e. n First Class 5.4 0.66 1,181 4.5 0.60 1,182 0.9 0.89 4.9 0.44 2,363 Priority mail 5.2 0.65 1,181 4.5 0.60 1,181 0.7 0.88 4.9 0.44 2,362 Difference 0.2 0.92 — 0.0 0.85 — — — 0.0 0.63 — All postage 5.3 0.46 2,362 4.5 0.43 2,363 0.8 0.63 44.8 — 4,725 Note.—Percentages are weighted. —, not applicable; s.e., standard error; n, unweighted denominator for the given percent. Open in new tab Table 2. ANES NRFU Undeliverable Returned Mail Rates by Experiment Condition Cash condition Postage Visible cash Not visible cash Difference All cash Condition Percent s.e. n Percent s.e. n Percent s.e. Percent s.e. n First Class 5.4 0.66 1,181 4.5 0.60 1,182 0.9 0.89 4.9 0.44 2,363 Priority mail 5.2 0.65 1,181 4.5 0.60 1,181 0.7 0.88 4.9 0.44 2,362 Difference 0.2 0.92 — 0.0 0.85 — — — 0.0 0.63 — All postage 5.3 0.46 2,362 4.5 0.43 2,363 0.8 0.63 44.8 — 4,725 Cash condition Postage Visible cash Not visible cash Difference All cash Condition Percent s.e. n Percent s.e. n Percent s.e. Percent s.e. n First Class 5.4 0.66 1,181 4.5 0.60 1,182 0.9 0.89 4.9 0.44 2,363 Priority mail 5.2 0.65 1,181 4.5 0.60 1,181 0.7 0.88 4.9 0.44 2,362 Difference 0.2 0.92 — 0.0 0.85 — — — 0.0 0.63 — All postage 5.3 0.46 2,362 4.5 0.43 2,363 0.8 0.63 44.8 — 4,725 Note.—Percentages are weighted. —, not applicable; s.e., standard error; n, unweighted denominator for the given percent. Open in new tab Given that we randomized assignment to the two cash conditions, differences in the composition of the two groups would indicate differential effects of visible cash on survey response. We used t tests to compare the answers from respondents in both cash conditions to fifty-two estimates from all eighteen questions on the NRFU questionnaire (table 1 of the online supplementary material). (These questions addressed talking to neighbors, interpersonal trust, interest in politics, worry about privacy, free time, children in the household, political party identification, presidential election turnout and candidate choice, internet use, education, gender, age, citizenship, Hispanic ethnicity, and attitudes toward professors, reporters, surveys, and the government. See DeBell and Maisel 2018 for the question wording.) The sample is marginally powered for this comparison, and after a Bonferroni correction for fifty-two comparisons, only differences exceeding about nine percentage points would be detectable. The mean weighted difference was 1.6 points (using absolute values, compared with an average sampling error of 1.5), and no differences were statistically significant after the Bonferroni adjustment, although three were in the five-point range. Smaller differences than were detectable could be substantively significant, so this test shows that there were no large differences in responding sample composition between the visible and nonvisible cash groups but does not provide evidence about smaller differences. We used the same procedure to compare the results for the mail conditions. The mean absolute difference between the First Class and Priority Mail groups was 1.9 points (compared with an average sampling error of 1.4), and no differences were statistically significant after the Bonferroni correction. Again, only large difference could have been detected. 6. DISCUSSION Priority Mail postage is considerably more expensive than First Class mail, at $6.65 compared with $0.46 in this study, yet it did not yield a higher response rate. First Class mail, therefore, dominates Priority Mail as a postage option. We were conscious of the potential for theft to thwart delivery of mail with visible cash. We believe that although thefts could occur, the best indicator of the value of visible cash is its effect on the response rate, and this test shows its net effect was to improve the response rate. Therefore, if any thefts did occur, they appear to have been incidental. Moreover, postal employees routinely handle valuable items and may be deterred from theft by the fact that stealing a piece of mail is a federal crime punishable by up to five years imprisonment (18 US Code § 1709). In this context, $5 is a paltry amount that seems unlikely to motivate frequent theft. A second test of the incidence of theft is provided by the undeliverable returned mail rates. Cash visibility does not affect whether an address is valid or mail is deliverable, so we expect equal numbers of mail pieces in each condition to be undeliverable. If postal pilferage is more common with visible cash, we would expect less of the undeliverable mail to be returned, so we would expect a lower returned mail rate for the visible cash condition than for the not visible condition if theft occurs. As shown in table 2, we received more returned mail with visible cash than with not visible cash, giving no evidence that theft occurred at meaningful rates. If anything, this implies that postal workers may have taken special care with the letters containing visible cash. The marginally higher returned mail rate with visible cash may be informative for researchers, who could discover more addresses that are ineligible and therefore save money on future mailings by removing these addresses from the eligible sample. As the use of visible cash is a novel technique, there are many opportunities for further research to establish the extent to which these results will generalize and improve the theoretical understanding of the effects of visible cash. As a practical matter, we found no evidence of theft, but it is worth considering that repeatedly mailing large volumes of visible cash at the same post office might attract thieves and that research designs with a tightly clustered sample, where individual letter carriers deliver multiple letters with visible cash, could also increase the temptation for theft. The letters in this study were addressed to “resident” of the recipient’s city (such as “DENVER Resident”). Letters addressed to named persons look less like junk mail. Therefore, to the extent that the increase in response rates from visible cash is due to motivating respondents to open letters they otherwise would have thought were junk mail, the effect of visible cash might be lower in a study of named persons. On the other hand, if the mechanism of visible cash’s effect is mainly to make the incentive more salient, the effect in a study of named persons might not be reduced. A future study could test this. The effect of an incentive’s salience may depend on how well matched the incentive is to the respondent burden. The $5 incentive we offered was substantial relative to the respondent burden to complete a one-page questionnaire, which we estimated to be a three-minute task. Making the incentive more salient could backfire if the burden-to-reward ratio is much higher. A future study could also ask respondents to report their impressions of the survey invitation. This could confirm or reject the hypothesis that visible cash creates a more positive first impression and makes the incentive more salient. And although we found no evidence of bias from visible cash incentives compared with nonvisible incentives, a future study could include a control group that receives no incentive. Finally, we believe it is worth testing the effects of visible cash in different study designs. The present study was a nonresponse follow-up study, disguised as a new mail study. We believe the effects are likely to generalize to other mail nonresponse follow-ups and to mail studies with fresh sample from the general population. It is worth testing the effects of visible cash in advance or invitation letters for other study designs, such as face-to-face interviews and mail push-to-web. Visible cash improved the response rate by four percentage points in this study. This result is consistent with the hypothesis that visible cash motivates respondents to open their mail and respond to survey invitations, perhaps as a result of making the incentive gift more salient by making respondents aware of it before they read the survey invitation. In many studies that use prepaid incentives, there is little or no extra cost to send incentives visibly in large-window envelopes instead of hidden in regular envelopes, so this is a promising strategy to improve response rates. SUPPLEMENTARY MATERIALS Supplementary materials are available online at academic.oup.com/jssam. The authors thank the principal investigators of the American National Election Studies in 2016 for approving the research: Shanto Iyengar, Ted Brader, and Vincent Hutchings. This research was funded by a grant from the National Science Foundation to Stanford University (grant #SES-1444910). Findings and opinions are those of the authors alone. References American National Election Studies ( 2018 ), ANES 2016 Methodology Dataset. December 10, 2018 Version , Ann Arbor, MI and Palo Alto, CA : The University of Michigan and Stanford University . WorldCat COPAC Brems, C., M. E. Johnson, T. Warner, and L. W. Roberts (2006), “Survey return rates as a function of priority versus first-class mailing.” Psychological Reports, 99, 496–501. COPAC DeBell M. , Maisel N. ( 2018 ), “Methodology of the ANES 2016 Non-Response Follow-Up Study,” accessed February 26, 2019, available at https://electionstudies.org/wp-content/uploads/2018/12/anes_timeseries_2016_nonresponse_follow-up_study_methodology.pdf. DeBell M. , Amsbary M. , Meldener V. , Brock S. , Maisel N. ( 2018 ), “Methodology Report for the ANES 2016 Time Series Study,” accessed June 14, 2018, available at http://www.electionstudies.org/studypages/anes_timeseries_2016/anes_timeseries_2016_methodology_report.pdf. Dillman D. ( 2007 ), Mail and Internet Surveys: The Tailored Design Method 2007 Update with New Internet, Visual and Mixed-Mode Guide , New York : Wiley . Google Preview WorldCat COPAC Dykema J. , Jaques K. , Cyffka K. , Assad N. , Hammers R. G. , Elver K. , Malecki K. C. , Stevenson J. ( 2015 ), “ Effects of Sequential Prepaid Incentives and Envelope Messaging in Mail Surveys ,” Public Opinion Quarterly , 79 , 906 – 931 . Google Scholar Crossref Search ADS WorldCat Greene A. J. , Prepscius C. , Levy W. B. ( 2000 ). “ Primacy versus Recency in a Quantitative Model: Activity Is the Critical Distinction ,” Learning & Memory , 7 , 48 – 57 . Google Scholar Crossref Search ADS WorldCat Groves R. M. , Singer E. , Corning A. ( 2000 ), “ Leverage-Salience Theory of Survey Participation: Description and an Illustration ,” Public Opinion Quarterly , 64 , 299 – 308 . Google Scholar Crossref Search ADS PubMed WorldCat Hagan D. E. , Collier C. M. ( 1983 ), “ Must Respondent Selection Procedures for Telephone Surveys Be Invasive?, ” Public Opinion Quarterly , 47 , 547 – 556 . Google Scholar Crossref Search ADS WorldCat InfoTrends ( 2016 ), “Direct Mail: Integral to the Marketing Mix in 2016,” accessed November 14, 2017, available at http://www.ddprints.com/downloads/infotrends_whitepaper_march_2016.pdf. Jans M. , Park R. , Rauch J. , Grant D. , Edwards S. ( 2015 ), “ Logos and Inserts Can Reduce Survey Return Rates: An Experiment in California ,” Survey Practice , 8 , accessed June 6, 2017, available at http://www.surveypractice.org/index.php/SurveyPractice/article/view/299/html_36 WorldCat Mamedova S. , McPhee C. ( 2012 ), “Return to Sender: Improving Response Rates in Two-stage Mail Surveys,” presented at the 2012 Meeting of the Federal Committee on Statistical Methodology , Washington DC . Google Preview WorldCat COPAC Mercer A. , Caporaso A. , Cantor D. , Townsend R. ( 2015 ), “ How Much Gets You How Much? Monetary Incentives and Response Rates in Household Surveys ,” Public Opinion Quarterly , 79 , 105 – 129 . Google Scholar Crossref Search ADS WorldCat Murphy J. , Hofacker C. , Mizerski R. ( 2006 ). “ Primacy and Recency Effects on Clicking Behavior ,” Journal of Computer‐Mediated Communication , 11 , 522 – 535 . Google Scholar Crossref Search ADS WorldCat Nisbett R. , Ross L. ( 1980 ), Human Inference: Strategies and Shortcomings of Social Judgment , Englewood Cliffs, NJ : Prentice Hall . Google Preview WorldCat COPAC Petty R. E. , Cacioppo J. T. ( 1986 ), “ The Elaboration Likelihood Model of Persuasion ,” Advances in Experimental Social Psychology , 19 , 123 – 205 . WorldCat Petty R. E. , Wheeler S. C. , Bizer G. Y. ( 1999 ), “ Is There One Process or More? Lumping versus Splitting in Attitude Change Theories ,” Psychological Inquiry , 10 , 156 – 163 . Google Scholar Crossref Search ADS WorldCat Singer E. , Ye C. ( 2013 ), “ The Use and Effects of Incentives in Surveys ,” The ANNALS of the American Academy of Political and Social Science , 645 , 112 – 141 . Google Scholar Crossref Search ADS WorldCat Tetlock P. E. ( 1983 ), “ Accountability and the Perseverance of First Impressions ,” Social Psychology Quarterly , 46 , 285 – 292 . Google Scholar Crossref Search ADS WorldCat Waugh N. C. , Norman D. A. ( 1965 ), “ Primary Memory ,” Psychological Review , 72 , 89 – 104 . Google Scholar Crossref Search ADS PubMed WorldCat Williams K. , Gentry R. ( 2017 ), “Mailing Methods Matter,” presented at Annual Meeting of the American Association for Public Opinion Research , New Orleans . Google Preview WorldCat COPAC © The Author(s) 2019. Published by Oxford University Press on behalf of the American Association for Public Opinion Research. All rights reserved. For permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - Improving Survey Response Rates with Visible Money
JF - Journal of Survey Statistics and Methodology
DO - 10.1093/jssam/smz038
DA - 2020-05-01
UR - https://www.deepdyve.com/lp/oxford-university-press/improving-survey-response-rates-with-visible-money-Xim3g0UaDS
SP - 1
VL - Advance Article
IS - 
DP - DeepDyve
ER -