What do campaign contributions buy? Lobbyists’ strategic giving

Scholars looking for evidence of corruption in Congress have focused on the data available to them: the limited contributions of political action committees. This literature has largely failed to identify systematic money-induced legislative behavior. But what if lobbyists are using their personal funds to contribute to congressional campaigns? I use newly available data to show that individual lobbyists make contributions in predictable ways, favoring key members at key times. In particular, healthcare lobbyists were significantly more likely to give to members of the committees drafting the Affordable Care Act relative to other members and other times. The findings represent an important step forward in understanding what actors who are interested in legislative decisions might expect in return for their campaign contributions.


Introduction
What do campaign contributions buy? Lobbyists meet with, inform, and attempt to persuade legislators and their staff to adjust legislation to their liking, while on the same day they are giving money to the legislator's political campaign-seemingly as close to legalized bribery as one can get. Scholars looking for evidence of corruption have previously focused on the data available to them: the contributions of political action committees (PACs). An extensive literature on PAC contributions has largely failed to identify systematic evidence of PAC-money-induced legislative behavior (for reviews of this literature, see Baumgartner and Leech 1998;Smith 1995;Richter et al. 2009).
But what if interest group lobbyists are using their personal funds to contribute to congressional campaigns? Would these efforts be possible to detect? Until the Honest Leadership and Open Government Act (HLOGA) of 2007, there was no way to observe the contributions of registered lobbyists (except as members of the general public; see Koger and Victor 2009). We now have for the first time the ability to connect the efforts and success of individual lobbyists with their contemporaneous contributions to Members of Congress.
This paper uses newly available data to make inferences about what lobbyists' financial contributions are intended to do. A dyadic analysis of lobbyists and senators makes possible models that explain whether-or not-a particular lobbyist gives to a particular senator at different points in time. The findings indicate that lobbyists respond in predictable ways to legislative events in Congress by contributing more to key members during key periods. When a lobbyist's issue is at the top of the agenda, as was health reform in 2009, that lobbyist focuses contributions on key committee members and puts less emphasis than at other times on constituent ties, majority party status, and electoral proximity in deciding to whom to donate. The results suggest worryingly that campaign contributions are intended, at least, to draw attention to a lobbyist's current concerns, and at most, to influence the legislator to act in the lobbying group's interests.

Theory and literature
Studies investigating the rationale for citizen contributions to electoral campaigns focus on two principal reasons, one electoral and the other transactional. The electoral explanation holds that donors give to candidates whom they would like to win the next election. If this explanation is correct, we might expect individuals to support their own representatives (Miler 2010;Hojnacki and Kimball 2001;Wright 1985), regardless of committee membership (Bennett and Loucks 1994) or leadership status (DeGregorio 1997). The transactional explanation holds that individuals who contribute to candidates do so in the hopes of buying some legislative benefit with their contribution, such as votes (Brooks et al. 1998;Stratmann 1991;Wawro 2001;Fellowes and Wolf 2004;Gordon 2001), meetings (Austen-Smith 1995Langbein 1986;Lohmann 1995), earmarks (Rocca and Gordon 2013), tax breaks (Richter et al. 2009;Gupta and Swenson 2003), or bureaucratic outcomes (Godwin et al. 2012;Claessens et al. 2008;Drope and Hansen 2004). If this explanation is correct, we should see higher contributions when Members of Congress are able to focus more on the lobbyist's interests, for example, when the member is addressing the issue in committee (Hall and Wayman 1990;Bennett and Loucks 1994;Loucks and Bennett 2011;Rozell et al. 2012) or in a vote. We should also see contributions flowing to members of the majority party (Cox and Magar 1999;Rudolph 1999) and to the leaders in that party. This paper investigates a particular kind of campaign donor: those who are also registered lobbyists. The purpose of lobbying is to influence public policymaking. The principal function of Capitol Hill lobbyists is to secure meetings with legislators and their surrogates and, in these meetings, to persuade legislative offices to do what is in the interest of the lobbying group. Given this objective purpose of lobbying, I propose the theory that lobbyists who donate to legislators are doing so for transactional reasons-to influence legislative decision making, not to influence electoral outcomes. If so, lobbyists strategically contribute to the legislators with the most power over their issue area, and they time these contributions to have maximum effect.

The health reform debate of 2008-2010
To test this theory about the rationale behind lobbyists' contributions, I use the case of health reform legislation in the US Senate. The high salience of health reform legislation gave a great number of lobbyists the motivation and opportunity to attempt to influence the policy process. It was the first major policy debate following the 2008 requirement that lobbyists disclose to the House and Senate information about their individual federal campaign contributions. Every lobbyist in Washington who dealt with healthcare issues wanted "to be a part of health reform," making sure that their issue/industry/disease benefitted from the new law. 1 A new Democratic president was making healthcare reform his top priority, and the Democratic Congress was ready.
In particular, US senators, especially those on the Committee on Finance and the Committee on Health, Education, Labor and Pensions, or HELP, were in a position to be the most influential members of Congress working on health reform. The supermajority required to pass a bill over a filibuster in the Senate, and the Democratic majority in the House, meant that the Senate would be the chamber where most of the decisions and negotiations on the bill would occur. As tax-law-writer, the Senate Committee on Finance holds all jurisdiction in the Senate over Medicare and Medicaid-the vehicles through which the Congress would shape the reformed healthcare system. In November of 2008, Senate Finance Committee Chairman Max Baucus released a White Paper proposal for health reform legislation, which started the legislative discussion and remains very similar to today's public law. 2 If lobbyists give for transactional reasons, then they should pay particular financial attention to the politicians best positioned to influence the content of the bill of concern to the lobbyist. Thus after the release of the White Paper, the US Senate, and the Senate Finance committee in particular, should have been the obvious venue targeted by lobbyists working on healthcare reform (which I abbreviate to "health lobbyists"). To a lesser extent, health lobbyists might also attend to members of the HELP committee, whose jurisdiction covers discretionary spending on health; they should also give to those with greater political power, e.g., party leaders and the majority party, i.e., Democrats.
To maximize influence, lobbyists' contributions should be numerous and large. We might also expect contributions to be timed to have greatest effect: health lobbyists should increase their donations as health reform legislation is written, and they should specifically concentrate contributions on key committee members while that committee is marking up the bill. For these reasons, I propose three hypotheses: H1 Healthcare lobbyists give more money to more members of Congress during health reform, relative to before and after health reform, and relative to nonhealth lobbyists.
H2 Healthcare lobbyists give more monetary attention to members of the two key senate committees-Finance and HELP-relative to other senators and relative to nonhealth lobbyists.
H3 Healthcare lobbyists pay more attention to Finance members than to HELP members, especially later in the debate, as it becomes clearer that Finance would write the majority of the bill as well as manage it on the senate floor.

Data
Starting in mid-2008, lobbyists began filing forms, known as LD-203s, disclosing the date, amount, payee, and honoree of each contribution to a federal candidate, along with the name of the lobbyist-contributor and the organization for whom they lobby. 3 Lobbyists who do not make any contributions are also required to file a report. The Government Accountability Office estimated in 2010 and 2011 that as many as 98% of LD-203 reports accurately recorded all contributions as required, and fewer than 11% of registered lobbyists failed to file  In 2009-2010, the Federal Elections Commission (FEC) limited individual contributions to $2400 per donor per election and limited PAC contributions to $10,000 per year. While it is a common practice for individuals to contribute to a candidate's primary and general elections at the same time, there are numerous examples of contributions that reportedly exceed $4800, and some that exceed $10,000. Since filers frequently do not properly list the PAC as the contributor (e.g., listing it in the "payee" box), it can be difficult to determine whether the contribution exceeds the federal limits. I assume that if a reported contribution specifies an individual's name, and this person is not the candidate-recipient, then the individual is responsible for the contribution. For contributions that list an organization, such as a PAC or lobbying firm, and do not specify an individual donor's name, I assume it is a PAC contribution. I then assume that reported contributions that exceed their respective limit are errors, either in reporting the data, in which case they were probably amended, or in understanding the law, in which case they were likely refunded, and I drop these contributions from the data (0.27% of observations). I also drop negative contributions (0.21% of observations), since these are most likely refunds for erroneous contributions.
To identify contributions to incumbent senators, I searched for every conceivable variation on each senator's name and then standardized the names of the recipients. An assistant searched individually for information about candidates who were not easily identifiable as House or Senate candidates. I then merged the names with information about each incumbent senator, including their class, party, and committee membership, using information from the official Senate website. I found that about 30% of all the contributions went to elected senators, which is the subgroup of interest here. All contributions to Senators Biden, Clinton, McCain, and Obama were dropped from these data, since these candidates were running for president or vice president in 2008.
We gain additional information on lobbying as a result of the Lobbying Disclosure Act (LDA) of 1995, which requires that lobbyists report information about their substantive lobbying efforts and associated expenditures. 5 From these data I derive the number of clients the organization has in each US state (or, for in-house lobbyists, the state in which the organization is headquartered). In addition, the Center for Responsive Politics (CRP), a nonprofit government watchdog organization, uses the LDA forms to create a detailed, standardized database (found at OpenSecrets. org) that includes the names of individuals who lobby for a registrant on behalf of a client, as well as the issues they work on. CRP created from these data a list of all lobbyists who worked on healthcare reform issues, which I matched to LD-203 data. I also use CRP's cleaned expenditures data. Matching organizations and individuals 5 Individuals who work for a lobbying firm and spend 20% of their time and $3000 or more in a quarter lobbying on a client's behalf must submit quarterly information about these expenditures, most of which comprise the lobbyists' salaries or fees. The firm as a whole must file separately as well. In-house lobbyists must file a disclosure form for the organization if their lobbying expenditures exceed $11,500 in a quarter. The information required includes the names of each individual lobbying for the organization (and whether new lobbyists are "covered" under the lobbying restrictions on former government employees) and the policy areas (of nearly 80 choices) and specific issues on which each registrant lobbied for each client, along with the dollar expenditures each registrant spent on a client's (or their employer's) behalf during the quarter. between the LD-203 data, the LDA data, and the OpenSecrets data required text matching to identify the same people and organizations. After a lengthy text-matching process, I matched 89.4% of OpenSecrets' list of healthcare lobbyists with LD-203 names and organizations. 6

Describing contributions activity
During the health reform debate period, defined here as starting with President Obama's November 4, 2008, election and ending when he signs the legislation into law on March 23, 2010, about 2800 individual lobbyists who filed LD-203 forms were also listed in LDA data as lobbying regarding healthcare issues, including Medicare, Medicaid, and prescription drugs. These health lobbyists, who account for 34% of the LD-203 filers, provided about 60% of all the money contributed by lobbyists to federal candidates and about 62% of the money going to incumbent senators. Figure 1 highlights the disproportionate giving of health lobbyists during consideration of health reform legislation. As shown, healthcare lobbyists made more contributions to senators during this period than did lobbyists working on all other issues combined. The level of activity surrounding health care is remarkably high, since two-thirds of lobbyists do not work on healthcare issues at all during this time, and since there are more than 70 other issues about which people could lobby. Yet we know from Baumgartner and Leech (2001) that lobbying activity across policy issue areas is highly skewed.
The average health lobbyist contributed more to senators than did the average nonhealth lobbyist in 64 out of 73 weeks, as shown in Fig. 2. Similarly, the average dollar amount of a contribution from a health lobbyist to an incumbent during the health reform period, $942, was significantly higher than for the average nonhealth lobbyist, at $686. These findings support Hypothesis 1. Further, the data show that the same lobbyists increased their lobbying activity on health reform during this period. Paired t tests of LDA data show that between September 1, 2008, and December 31, 2009, the same firms increased their spending on health issues compared to the previous quarter (and this is significantly different 3 out of 4 times).
In addition to their heightened activity during this time, health lobbyists during the health reform debate gave in ways that support the notion that they were  Figure 3 shows that in an average week in 2008 before the election, health and nonhealth lobbyists contributed in similar numbers to senators not on Finance and HELP, though health lobbyists contributed more to members of those committees than did nonhealth lobbyists. Later, during consideration of what became the Affordable Care Act, health lobbyists nearly doubled their weekly contributions to key committee members (from $21,000 to $39,000) while reducing their contributions to other senators by 20% (from $44,500 to $37,000). In the same time period, nonhealth lobbyists decreased their weekly contributions to other senators by 34% (from about $38,000 to about $25,000) and increased their contributions to the two health committees by 46% (from $13,000 to $19,000)-perhaps in an effort to retain those senators' attention. After the Affordable Care Act was signed on March 23, 2010, both groups decreased their weekly contributions, though health lobbyists' contributions remained higher than nonhealth contributions, and their focus on key committee members persisted, if not quite at the levels of 2009. While Fig. 2 distinguishes health lobbyists from nonhealth lobbyists over time, Fig. 4 distinguishes health lobbyists' contributions to key committee members from health lobbyists' contributions to other senators. Hypothesis 2 argues that health lobbyist contributors give more money to key committee members than nonhealth lobbyists, and this expectation is supported in Fig. 4. Just after each key committee releases its proposed versions of the bill, and again leading up to the vote of the full Senate, there are peaks of contributions, and these peaks are higher for key committee members. (These peaks also correspond loosely with filing deadlines just before which candidates step up their fund-raising efforts.) During periods of low Finally, Hypothesis 3 holds that health lobbyists will increasingly prefer Finance members over HELP members as the health reform debate develops. Figure 5 shows that over the course of consideration of health reform, health lobbyists' contributions to Finance members increased more steeply than their contributions to HELP members. We also see in Fig. 5 that contributions to both committees peaked during their respective markups of the bill. Unexpectedly, though, during the Finance markup the peak for HELP members was higher than to Finance members. To draw more reliable conclusions, statistical analysis is needed.

Modeling contributions activity
To better evaluate the effects of legislation on lobbyists' contributions, I create a statistical model of lobbyists' decisions whether, when, to whom, and how much to contribute. Since FEC contributions data do not indicate whether a donor is a registered lobbyist, previous researchers have typically been unable to control for those lobbyists who choose not to contribute, either at all, or to a particular candidate. Importantly, the present study avoids selecting on the dependent variable by creating a data cell for every lobbyist who files an LD-203 contributions disclosure form paired with every incumbent senator during the health reform debate of 2008-2010. During the health reform debate, I argue, health lobbyists are more likely to make contributions than nonhealth lobbyists, and they direct them to senators on the two key committees, Finance and HELP. Further, I expect these trends to increase during the health reform debate relative to before, and to diminish after the new law is passed. To compare the effects of health and nonhealth lobbyists' contributions, I create two-way interaction terms multiplying whether a lobbyist works on health issues by each of the four variables describing senators' position in the Senate (Finance member, HELP member, Democrat, and party leader). Following Brambor et al. (2006), all constituent parts of these interactions are also included in the models.
Several important control variables are included in the models as well. First, the models control for the distinct possibility that senators running in the next electionwhether in 2008 or 2010-ought to receive more contributions than senators not running for re-election. Second, the models include a variable that captures whether the lobbyist is employed by a known lobbying firm, as reflected in Lobbyists.info's "Factors of Influence" list, 7 because lobbyists who work for large, established lobbying firms may consequently be interested in targeting more senators than would Third, the models include a variable that captures constituent effects. This variable records whether any of the lobbyist's clients are from the same state as the senator, as indicated in lobbying expenditures reports. (For in-house lobbyists, this variable reflects whether the organization is from the senator's home state.) By including constituent ties in the models, we can judge whether or not lobbyists are guided in their contributions by members' power in the chamber, beyond the known tendency for donors to give to their own congressional representatives (DeGregorio 1997;Wright 1985). 8 The fourth control variable is the organization's expenditures on lobbying. Since the vast majority of these expenditures goes to lobbyists' salaries, I use lobbying expenditures as an indicator of a lobbyist's capacity to give campaign contributions. These data also come from lobbying disclosure reports. For each employer, spending is summed across the six-quarters; the natural log of this number is used in the models, since it is reasonable to expect a curvilinear relationship.
The unit of analysis for the models is the lobbyist-senator dyad, and the dependent variable is an indicator variable capturing whether any contribution occurs in the dyad or not. (I use a dichotomous dependent variable because the total dollars contributed or the frequency of contribution is correlated with other variables such as lobbying firm size and group type.) I report odds ratios, which describe the odds that a positive value (i.e., 1) for an explanatory variable results in a positive dependent variable (i.e., 1), relative to a zero value for the explanatory variable. Since the same senators appear repeatedly in the dataset, robust standard errors are clustered around the senator. 9 Looking at Table 1, we see first that health lobbyists, whether in 2008 before the election, during the health reform debate, or in 2010 after the bill is signed, are significantly more likely to contribute to any given senator than are nonhealth lobbyists (H1). During the debate, the odds that a healthcare lobbyists contributes to a given senator are about 81% greater than the odds that a nonhealth lobbyist gives to a senator, controlling for the other variables in the model. Before the debate, the odds that a health lobbyist contributes to a given senator are about 53% greater than the odds for a nonhealth lobbyist, and after the debate the odds are about 76% greater. It thus appears that health lobbyists are sensitive to legislative events and are more likely to make contributions when their clients or employers have more at stake.
Further down the list of variables we see the interactive effects of health lobbyists with respect to particular senators (H2). During the consideration of health reform, model predictions indicate that the odds that a health lobbyist contributes .791 .803 The unit of analysis is the lobbyist-senator dyad. The first model describes contributions in 2008 before the election of President Obama, and the release that same week of Senator Baucus's broad plan for health reform (known as the white paper). The second model describes contributions activity during the health reform debate (defined as 11/4/08-3/23/10). The third model covers 2010 after the bill was signed into law. Errors are clustered on the senator. Coefficients in bold are significant at p < .05 to members of the Senate Finance committee are almost 2-1/2 times the comparable odds for non-Finance members. Likewise, the odds of a contribution from a health lobbyist to a member of the HELP committee are also 2-1/2 times the size of the odds if the senators are not on the HELP committee (2.412 and 2.415, respectively; both are statistically significant 10 ). The third hypothesis (H3) holds that healthcare lobbyists' contributions focused more on Finance members relative to HELP members, especially later, as the chair of the Finance committee managed the bill on the floor and the Finance committee began to oversee implementation of the law. The table shows that before the debate, health lobbyists were more likely to give to members of the HELP committee than to members of the Finance committee. After the debate, the odds that a Finance member receives a contribution from a health lobbyist are 41% more likely than if either or both of Finance and health lobbyist are zero, while for HELP members, the odds of receiving a contribution from a health lobbyists are only 3% greater than if HELP and/or health lobbyist is zero, and the latter is not significant. It seems likely that the full set of health lobbyists-some experienced, some newly mobilized-did not immediately appreciate the overwhelming role of the FFF Finance committee in writing the bill, and gave liberally to the committee whose name starts with the word "health." Alternatively, other factors, including the timing of the two committees' markup sessions, and the death of HELP chair Senator Kennedy, drew lobbyists' attention away from HELP and toward Finance later in the debate. (Dropping Senator Kennedy from the analysis has very little effect on the models, however, and does not change my conclusions.) Comparing the three models, we see that lobbyists pay considerable attention to home-state connections: when one or more of the lobbyist's clients have addresses in a given senator's state, the odds that the lobbyist contributes to that senator are between 11 and 17 times greater than if the lobbyist has no clients in the senator's state. Interestingly, these constituent effects are greatest after the debate and weakest during it, suggesting that issue pressures supplant constituent pressures somewhat when an issue is at the top of the legislative agenda. Likewise, we see in the "after" period increased attention to senators running for re-election and greater contributions to the majority and minority leader, relative to during the debate.
Lobbyists employed by major lobbying firms are more likely to contribute, and this effect is stable over time, supporting the use of this variable as a control. In addition, the more a lobbying organization spends, the more likely are the organization's lobbyists to make personal contributions to senators: During the health reform debate, with each standard-deviation increase in the natural log of an organization's lobbying expenditures, the model predicts a 31% increase in the odds that one of its lobbyists will donate to a given senator. These results are consistent with previous research that notes that for interest groups, making campaign contributions and spending money on lobbying are complementary activities (Ansolabehere et al. 2002). The findings are also in keeping with anecdotal accounts that well-paid lobbyists are often expected by those paying them to make political campaign contributions (Leech 2013).

Discussion and conclusions
Newly available data on the personal contributions of registered lobbyists to federal candidates show that lobbyists vary their contributions as a function of the issue at the top of the legislative agenda. Although only about 34% of lobbyists in these data worked on healthcare issues, these lobbyists contributed 62% of the money given to senators during the Senate's consideration of health reform. As the health reform debate began, health lobbyists nearly doubled their contributions to members of the two committees writing health reform legislation and simultaneously decreased their contributions to senators not on those key committees. After the bill became law, health lobbyists continued to pressure the committee that would oversee its implementation, while they significantly decreased their contributions to the other committee whose jurisdiction includes health but which would not oversee the law's implementation.
This paper makes at least three advances over previous scholarship. First, it uses lobbyists' personal contributions, rather than those of interest group PACs. Doing so responds to critiques (e.g., Milyo et al. 2000) of the PAC literature that low dollar limits may obscure any links between contributions and policy outcomes. Since there is no limit on the number of individuals interest groups may employ to lobby for them, interest groups may use lobbyists to greatly multiply the dollars they can contribute to the campaigns of members of Congress, and the present study supports this possibility. Moreover, the new data allow us to differentiate between ideologically minded contributions from the general public and transaction-oriented contributions from lobbyists, since by definition lobbyists seek to influence politicians.
Second, the study employs a dyadic design that includes all instances of contributions between a lobbyist and a senator, as well as all potential instances of contributions from lobbyists to a senator. This method allows inferences about the conditions under which lobbyists choose to make a donation or choose not to donate. We can observe contributions from lobbyists working on the central issue of the time period as well as those not working on it; we can also observe the likelihood of contributions by interested lobbyists at different points in time.
Third, this research suggests that lobbyists whose clients have a stake in the specifics of highly salient legislation do sometimes use their own money to increase their access to, enhance their influence over, or show their appreciation of elected officials who are in a position to help lobbyist-donors' clients. While it can be very difficult to adjudicate among these three reasons for lobbyists' contributions (Lowery 2013), the predictability of lobbyists' contributions based on their issue interests supports the notion that financial contributions from lobbyists are intended as transactions, rather than merely efforts to keep ideologically similar politicians in office.
There is scope for future investigation of lobbyists' contributions to members of Congress. In particular, researchers should investigate lobbyists' contributions surrounding additional issues. It may be, for example, that enthusiasm for the new president and his choice to make health reform his top priority was largely responsible for the sharp increases in bipartisan donations from health lobbyists-though this does not explain health lobbyists' demonstrated preference for key committee members, nor their tendency to give to Finance members over HELP members as the months pass, nor the decrease in their contributions to their clients' own senators in the midst of the health reform debate. Similarly, researchers might investigate lower-salience issues, since it is likely that the dominance of health reform legislation in 2009 allows us to observe a significant pattern while strategic contributions surrounding lower-salience issues tend to wash out in statistical analyses (also see Jones and Keiser 1987 in support of this point).
For several reasons, I stop short of arguing that this analysis proves that lobbyists make campaign contributions in a strategic effort to "buy" desired policy outcomes. In the models presented, variables that are not related to an interest in-or jurisdiction over-health issues also matter, and sometimes matter more. While the data show that the effect of constituent ties is lowest during the health reform debate and higher before and after the debate, the existence of one or more clients in the senator's home state is still the most powerful predictor of a contribution in the models. Other control variables matter too: the odds that a lobbyist who works for a known lobbying firm contributes to a given senator are 2-3 times greater than for other lobbyists, and the odds that the Senate's majority and minority leaders receive lobbyists' contributions are 2.5-3.8 times greater than for other senators. Still we see that just being a healthcare lobbyist during 2008-2010 is a consistent and statistically significant predictor of greater likelihood of contributions, especially when this variable is interacted with those senators on the two health-writing committees.
An additional reason I do not argue that lobbyists have succeeded in purchasing influence with their contributions is that I have not identified here any specific link between lobbyists' contributions and enhanced access or influence for contributing lobbyists. Previous scholars have argued that campaign contributions cannot buy legislative favors, because, for example, the "campaign contribution contract," if it exists, is unenforceable (McCarty and Rothenberg 1996), or because contributions on both sides of an issue tend to balance out and nullify any effects (Baumgartner et al. 2009). But the findings of this paper do suggest that the lobbyists are at least trying to purchase something, supporting recent research that provides significant and troubling links between campaign contributions and donors' access to members of Congress (Kalla and Broockman 2015;Vidal et al. 2012;LaPira and Thomas 2014;Miler 2007). In short, the data analyzed here provide compelling evidence of a phenomenon that is notoriously difficult to document (Lowery 2013)-that groups seeking particular, often financial, gains from legislative actions do seem to use money in an effort to help achieve these gains.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creat iveco mmons .org/licen ses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.