یک تحقیق تجربی از تاثیر اطلاعات حسابداری بر روند نفوذ دادن فرد

In this study, we experimentally investigate accounting information’s role in the individual donation process. Specifically, we manipulate the presence of service efforts and accomplishments (SEA) information in conjunction with a typical fundraising request. We then investigate whether donors obtain comparative financial accounting information for the purpose of maximizing donation efficiency. In our experiment, potential donors felt that fundraising requests containing SEA disclosures were more informative than typical fundraising pleas. In addition, supplementing a standard fundraising request with summary SEA information significantly increased (1) the quality perception of the requesting charitable organization and (2) the percentage of potential donors who claimed they would donate to the requesting organization in the future. However, SEA information did not translate into increased actual giving in our study. Our experiment next offered actual donors the opportunity to obtain comparative financial information for the purpose of verifying efficient giving (i.e., donors could compare the soliciting organization to a similar organization). Although the majority of actual donors were unwilling to obtain this information, the vast majority of those viewing financial accounting information donated to the organization with the highest program-spending ratio. As such, our results support two seemingly opposed contentions: (1) only a minority of individual donors request and use financial information, however, (2) there is a need for increased accuracy of not-for-profit expense classification.

Charitable donations by individuals have grown from $110 billion in 1990 to $164 billion in 2001 (Byrne, 2002, p. 83). These donations represent approximately two percent of the US gross domestic product (Strom, 2002, p. A19) and are three times greater than the combined charitable donations made by corporations, bequests, and foundations (Byrne, 2002, p. 89). Despite the dollar amounts involved, systematic knowledge of accounting information’s influence in not-for-profit (NFP) settings is undeveloped relative to our knowledge of accounting information’s influence in for-profit settings (Parsons, 2003, p. 104). In the context of individual giving, only recently have accounting researchers (1) outlined a framework to consider when, why, and if individuals might request and use NFP accounting information (Gordon and Khumawala, 1999, pp. 42–45), (2) demonstrated a positive association between favorable financial accounting information and total individual giving (Callen, 1994 and Tinkelman, 1999),2 and (3) shown that potential donors’ exposure to favorable accounting information can directly cause increased individual donations (Parsons, forthcoming). The current study advances our knowledge of accounting information’s influence on individual giving decisions by experimentally modeling a common medium for individual fundraising requests – direct fundraising solicitations. Specifically, in our laboratory setting, we manipulate fundraising requests to either include or exclude service efforts and accomplishments (SEA) information. Broader in nature than pure financial accounting information, SEA information captures mission effectiveness using both financial and non-financial metrics. Although both the FASB and the GASB encourage SEA disclosure, current accounting rules make SEA reporting optional. To our knowledge, the influence of SEA information on individual fundraising requests has not been empirically investigated.3 Consistent with expectations, we provide evidence that SEA information favorably affects potential donors. Specifically, we find that summarized SEA information, which might easily be included in fundraising requests at minimal cost, caused experimental subjects to become (1) better informed about the soliciting organization, (2) more cognizant of the organizations’ past performance, and (3) more likely to donate to the organization in the future. These findings are potentially important to regulators – especially in light of evidence that NFPs rarely provide voluntary SEA information (Brace et al., 1980 and Hyndman, 1991). That being said, our results do not unambiguously suggest that SEA information increases individual donations. Specifically, although the presence of SEA information significantly increased the percentage of subjects who said they would make future donations, we find no difference in actual current donations when experimental subjects were invited to donate their experimental compensation to the charitable organization in question. Our experimental model of the individual solicitation process concluded by asking all subjects who made actual donations if they would be willing to view summary financial accounting information in order to compare the soliciting organization to a similar organization (and possibly change the recipient of their donation). This portion of our study was designed to empirically gauge the extent to which donors would (1) request financial information (Gordon and Khumawala, 1999, p. 31) and (2) allocate resources to the organization that reports a higher proportion of program spending ( Callen, 1994 and Tinkelman, 1999). Fewer than half of actual donors were willing to view summarized comparative financial information to ensure their donation was economically efficient. Although subjects previously exposed to SEA information requested comparative financial information at a higher rate than those without access to SEA information (42 percent vs. 38 percent), this difference was not significant. The relatively low level of donors requesting comparative financial information is surprising because (1) the effort required to make the comparison was quite low, (2) experimental subjects were business students (all else equal, our subjects have a relative proclivity towards using financial information), (3) the specific identity of the soliciting organization was masked to subjects until the experiment was completed (eliminating emotional attachment to the soliciting organization), and (4) students were explicitly informed that the financial comparison might lead them to donate their resources more efficiently. While a relatively low level of financial information requests suggests that accurate NFP financial reporting might be unimportant to individual donors, the vast majority of research subjects who viewed summarized financial information allocated their donation based on reported efficiency measures (as defined by Parsons (2003, p. 114)). Specifically, after viewing common-sized financial summaries, 89 percent of subjects made their donation to the charitable organization with the highest program-spending ratio. As such, our results support the argument that there is an increased need for greater accuracy in expense reporting by functional classification and in preparation of IRS Form 990 (Wing et al., 2004 and Urban Institute, 2004a). The remainder of this paper is organized as follows. In the next section, we develop hypotheses related to the use of accounting information by individual donors. In Section 3, we describe our experimental design while we present our results in Section 4. We conclude in Section 5 by discussing the policy implications of this study.

We analyze background questions in order to assess subject preferences for the charitable cause used in our experiment (i.e., charities benefiting children) and to assess unplanned differences between our experimentally manipulated groups (i.e., SEA vs. no SEA information). Consistent with our pre-test survey (see footnote 7), subjects exhibited an affinity for charities benefiting children. Specifically, on a seven-point Likert scale with four as the mid-point between “least favorite” and “favorite” charitable cause, the average subject ranking for children’s charities was 5.3 (significantly greater than four; t = 17.08, p < 0.001, two-tailed). Although a simple ANOVA reveals that females preferred this cause marginally more than males (5.41 vs. 5.16; F = 3.05, p = 0.08), there was no difference between preferences with regards to groups who did vs. did not receive SEA information (F = 0.47, p = 0.50). As such, we provide some evidence that a general motivation for giving to the charitable cause in our study exists in our subject pool thus increasing the chance that potential donors will use the financial and SEA information in their decision (Gordon and Khumawala, 1999, p. 43). We next examine unplanned differences between our experimentally manipulated groups (i.e., SEA vs. no SEA groups prior to the manipulation). With regards to (1) the frequency with which subjects typically read fundraising requests, (2) the frequency with which subjects make charitable donations, (3) the value placed on the experimental compensation (the engraved pencil), and (4) the proportion of males vs. females, there were no significant differences between experimentally manipulated groups (the smallest p value is 0.19). As such, we analyze the first part of our study, the effects of SEA information on individual giving, with confidence that any observed differences are due to the manipulated factor (the presence or absence of SEA information).4.1. Part 1: SEA information results To test H1, we investigate the value of SEA by analyzing participant responses to the previously discussed Likert scale (where 1 = Not Nearly Enough Information, 4 = The Right Amount of Information, and 7 = Too Much Information). Subjects with SEA information felt the fundraising request was more informative than subjects with only a basic fundraising appeal. Specifically, subjects with SEA information had an average response of 3.61 vs. 2.98 for subjects without SEA information (t = 3.92, p < 0.001, two-tailed). We highlight that subject responses do not trivially reveal “having more information is better” because both groups’ average responses are significantly less than having “the right amount of information” per our scale. In other words, the organizational outcome information (i.e., SEA information) was not considered “information overload” by subjects because these potential donors, on average, would have preferred even more information about the soliciting organization.10 SEA information also emphasized that the soliciting organization had above average past performance. Based on the previously discussed scale (where 1 = A Poor Performer, 4 = An Average Performer, and 7 = A Top Performer), subjects with SEA information felt the soliciting organization’s performance was significantly above average while those without SEA information did not. Respectively, the average responses were 5.11 and 4.06, (t = 6.16, p < 0.001, two-tailed). Given the content of the basic fundraising appeal (Appendix A) and the additional SEA information (Appendix B), this result is not surprising. Specifically, with no outcome feedback, a basic fundraising appeal can reasonably be viewed as providing a two-part message: (1) there is a problem and (2) we need your support to address the problem. With SEA information, a third part of the message is delivered: (3) we are putting funds to good use.11 Barring extant knowledge of the organization, which was not possible in our experimental setting (i.e., the organization’s identity was masked to subjects), potential donors logically need some form of performance information – such as SEA – to assess whether funds are being put to good use. As described above, the presence of SEA information significantly improved subject perceptions of the soliciting organization; however the central question for fundraising purposes is whether disseminating SEA information increases individual donations. When asked whether they would make a future donation to the soliciting organization, a larger percentage of subjects with SEA information reported that they would contribute when compared to subjects without SEA information. As shown in Table 1, average rates of planned future contributions were 84.8 percent and 61.5 percent, respectively (t = 3.39, p = 0.001).12 While the preceding results support H2 and suggest that charitable organizations can benefit by promoting SEA information, as described below, planned future donations did not translate into actual donations.Recall that the final page of the experimental packet invited subjects to forgo their compensation (i.e., the engraved pencil) in exchange for $2.00 being donated to the actual charity represented in the experiment. As shown in Table 1, actual donation rates were 58.2 percent for subjects with SEA information and 61.5 percent for subjects without SEA information (t = 0.42, p = 0.67, two-tailed). Given that charitable organizations can experience somewhat sizable pledge collection loss rates,13 it is perhaps not surprising that subjects’ relatively informal “planned future donations” failed to materialize into actual donations. However, it is surprising that such failures happened differentially between our manipulated SEA groups (resulting in significantly different planned donations, but no difference in actual donations). In an effort to understand the inconsistency between planned and actual donation rates, we grouped subjects into four categories based on the combination of their planned future donations and actual donations. Specifically, Table 2 classifies subjects in the following manner: Group 1: subjects who plan to give in the future and make an actual donation; Group 2: subjects who plan to give in the future but make no actual donation; Group 3: subjects who do not plan to give, but make an actual donation; and Group 4: subjects who do not plan to give, and do not make an actual donation. It is interesting that SEA information consistently increased subjects’ perceptions of the soliciting organization only in Group 1 (see Panel A of Table 2). Nevertheless, the percentage of all subjects who planned a future donation and made an actual donation (i.e., the percentage of subjects in Group 1) was not significantly affected by the presence of SEA information. Specifically, 49 percent of subjects with SEA information are in Group 1 while a comparable 42 percent of subjects without SEA information are in Group 1 (t = 0.88, p = 0.378, two-tailed).14 In contrast to Group 1, subject perceptions of the soliciting organization were not systematically affected by SEA information in Groups 2, 3, and 4. Post-hoc analysis offers some evidence that factors other than the SEA manipulation contribute towards these results. Our method of post-hoc analysis was to construct a series of 4 × 2 ANOVAs (donation groups by SEA) using each background question as a dependent variable. In a sense, we check for unplanned differences in our subject pool now that we know both the exogenously manipulated SEA condition and the endogenously determined donation group. In total, three significant or marginally significant findings result from this analysis. First, donation group is significant (p = 0.024) with respect to the value placed on the engraved pencil. This main effect is driven by Group 2’s relatively large perceived value (i.e., those who planned to give in the future but did not make an actual donation valued the pencil more than other subjects). Second, donation group is marginally significant (p = 0.071) with respect to preference for charities that benefit children. This main effect is driven by other groups preferring the charity’s cause more than Group 4 (i.e., those who neither planned future nor made current donations had a relative dislike for charities that benefit children). Finally, the interaction between donation group and SEA information is marginally significant (p = 0.057) with respect to gender. This interaction is driven by Group 3’s non-SEA condition having a relatively large number of females when compared to Group 3’s SEA condition. Although less intuitive than the preceding results, this finding suggests that females were more likely than males to make an actual donation in spite of not planning a future donation. In total, part 1 of our study offers evidence that including SEA information with fundraising requests is beneficial because SEA data (1) provides desired information to potential donors, (2) increases the quality perception of the soliciting organization, and (3) increases the percentage of potential donors who plan to contribute in the future. In spite of these benefits, however, providing SEA information did not increase actual donations in our study. 4.2. Part 2: Comparable financial information results Recall that all subjects making an actual donation were given an envelope containing summary financial information comparing “For a Child’s Life” to a similar organization.15 Subjects made a choice between (a) returning to their seat in order to make the financial comparison or (b) donating the $2 gift to the charity represented by “For a Child’s Life” and leaving the classroom without making a financial comparison. Only 37 out of 94 subjects elected to make the financial comparison (39 percent) even though the required effort to make the comparison was rather low (i.e., a few additional minutes of participation).16 There is no difference when subjects are separated by those with vs. without SEA information (41 percent and 38 percent, respectively, made financial comparisons; t = 0.374, p = 0.710); however, there is a significant difference between donation groups. Specifically, those who planned to give in the future and made an actual donation (Group 1) were less likely to use financial information relative to those who did not plan to give in the future but made an actual donation (Group 3). Respectively, 33 percent and 59 percent of these subjects used financial information before determining the beneficiary of their gift (t = 2.20, p = 0.031, two-tailed). Reasoning behind this difference can be seen in Table 2. Specifically, Group 3 subjects felt relatively neutral about the organization’s past performance compared to Group 1.17 Consistent with Gordon and Khumawala (1999, p. 47; see Propositions 1, 2, and 4) the subject group with a relative perceived need for more information about organizational performance (i.e., Group 3) tended to use financial information before finalizing their giving decision while the subject group with a perceived information advantage (i.e., Group 1) tended to not use financial information. Of the subjects who used financial information, 33 out of 37 (89 percent) gave to the organization with the highest program-spending ratio and subjects indicated the observable financial difference was the reason for their selection. This finding is consistent with H3. Specifically, 29 out of 33 (88 percent) indicated that the program-spending ratio difference drove their donation decision.18 Of the four subjects who donated to the lower program-spending ratio organization, two gave to “For a Child’s Life” because they “knew something about it” (even though the comparison organization was financially dominant and actual organization identities were masked), one subject apparently did not understand that a relatively high program-spending ratio is better than a relatively low one, and one subject claimed to ‘feel sorry for’ the organization that was relatively inefficient. In summary, part 2 of our study supports the claims of Gordon and Khumawala (1999, p. 43) that financial information can play a role in the giving decision once donors have decided to support a charitable cause and are deciding among competing organizations that support that cause. Further, the fact that donors who used financial information to make their final decision overwhelmingly selected the organization with the highest program ratio confirms findings from earlier research, such as Callen, 1994 and Tinkelman, 1999, that report a positive association between reported program ratios and subsequent contributions. Finally, results from part 2 of our study corroborate the claims of Wing et al. (2004, p. 1) that donors may place “undue reliance” on accounting ratios.