Many scientists seem to agree that there is a crisis in as far as replicability is concerned. In this case, it is becoming increasingly difficult to reproduce scientific experiments with many seeming to yield different results while others use questionable research methods that lead to questionable results. The challenge cannot be narrowed down to a single factor. Rather, a raft of issues and a combination of several factors seem to plague research methodology, especially with regard to replicability. This also appears to be a problem affecting multiple disciplines. Practitioners and researchers are concerned that many experiments cannot be repeated nor can their methodology be extrapolated, transferred, or reproduced. A lot more scientific research is also unable to stand the test of time in both results, structure, and methodology. This has led to the emergence of a host of issues, with the consequences having the potential to affect every level of society.
In their body of work, LaCour & Green (2014) sought to answer several questions . Their main research question was quite simple ; can a single conversation change minds on divisive social issues such as same-sex marriage? The placebo trial sought to examine if either gay or straight canvassers were effective at changing the opinions of voters towards gay marriages. A number of questions were explored in the experiment. In the first question , the scholars aimed at determining who between the two different groups of canvassers had a longer-lasting effect when compared . Likewise, they aimed at exploring the extent of the change of opinion and whether or not it was transmitted within households . A number of theories were put forward and explored. Firstly, the scholars argued that prejudice reduction occurs and outgroup hostility reduces significantly when different groups interact with one another. Secondly, they distinguish ed between active and passive interaction and the effect they had, when compared against one another in the context of yielding a change in opinion.
Delegate your assignment to our experts and they will do the rest.
The experiment involved obtaining a set of both gay and straight canvassers. The groups were each presented a treatment script about gay marriage and a placebo script about recycling. The main criteria for choosing participants for the study was that they had to be registered voters from a precinct that supported a ballot measure banning same-sex marriage in 2008. They also had to be residents of an address that had one other registered voter. Likewise, they had to reside in a neighborhood that had detached dwellings. The canvassers were encouraged to be extremely polite and non-argumentative as they went about speaking with respondents. They were also encouraged not to force their opinions down on the respondents or even show an open bias. The script for both gay and straight canvassers was the same except for one difference, that during the conversation the gay canvassers would reveal that they were gay as far as sexual orientation was concerned. The experiment's sample size was 972 individuals. The number of gay canvassers was 22 while the straight canvassers were 19. The gay canvassers were recruited and trained at the Los Angeles LGBT center.
A baseline first had to be established so that a comparative analysis could be done after data and sufficient evidence had been collected. An online survey that contained 50 questions was sent to every individual with a random selection of topics. But amongst them, there were two questions about same-sex marriage, the idea being not to make same-sex questions the focus of the survey. Also, the scholars sought to divert attention from the same-sex marriage discussion so as to get real sentiments on the matter from respondents. The next ideal situation was to check with respondents after a canvasser had been able to meet them face to face and facilitate a discussion. Once a baseline had been established, a comparative analysis could be done. This would take place once data on the nature of their opinions had been obtained following a visit from a canvasser.
The main question posed was whether or not face-to-face interactions could sway the opinions of respondents and for how long the shift on opinions would last. The authors had proposed that contact with those who were viewed negatively, specifically, those who supported the legalization of same-sex marriage, increased the chance s of shifting public opinion . This is especially if they got a chance to have direct interaction with respondents and potential voters. The study’s baseline was first established by obtaining a basis through an online survey. This would form the basis of a comparison with the information that would be obtained after post-treatment was collected. Evidence from the data collected from respondents showed that gay canvassers managed to obtain a significant shift in opinion almost immediately. In the post-treatment survey that took place after three days, there seemed to be a considerable shift in opinion in favor of same-sex unions. Straight canvassers also reported significant shifts in opinion in favor of same-sex marriages. The change, however, seemed to have lasted longer with respondents who interacted with gay canvassers. In the post-treatment surveys that took place on the 3 rd day, 12 th day, 23 rd day, 27 th day, and 280 th -day mark, the results seemed to remain the same for gay canvassers while differing slightly with straight canvassers . The straight canvassers seem ed to have seen the opinions of their respondents reduce with time in terms of favoring same-sex marriages.
With regard to the evidence presented, while face-to-face interactions are compelling, there is no absolute way to measure an apparent shift in human sentiment. There is also no clear way of showing that one-on-one interaction was the only factor that caused a change in opinion or the only factor that contributed to a shift in opinion. This is because several other factors may have contributed . One example is the landmark ruling by the California Supreme Court allowing the legalization of same-sex unions in mid-2013 which coincided with the survey. The fact that by and large California is seen politically as a liberal state could have played a role especially given that at the time a Democrat president was occupying the white house and he was overwhelmingly in favor of legalizing same-sex unions. Given the dynamic nature of human behavior, opinion is difficult to measure, even when it shifts. It is safer to state that face-to-face interactions with canvassers may have contributed significantly , but not exclusively . For instance, there was a national shift in how people in the United States ( U . S ) had been viewing same-sex unions over a large period of time . This was accelerated by the presence of a Democrat and liberal-leaning president.
Having been impressed by LaCour & Green (2014) and their study results, Broockman, Kalla & Aronow (2015 ) sought to carry out an extension of the study. Thus, they requested and went over the data set. On exploring the information, the scholars began to develop doubts , especially with the parameters that had been used in the study. The study had concluded that a 20-minute conversation with a gay canvasser was enough to change the opinion of a person for over a year. While LaCour & Green (2014) initially had doubts about their study’s methodology and data, they went ahead with their plans. On close examination, it was realized that the study was not replicable. This is especially due to the inconsistencies that were identified in the data.
One of the issues with the LaCour-Green study was that the data seemed not to have been collected as described. Broockman et al . (2015) arrive d at this conclusion by contacting the survey firm used and the lead person. The firm confirmed that the said person had never worked for them and that the firm itself could not execute a study of that nature. Over time, other irregularities began to appear. The baseline data from the study seemed to be indistinguishable from a national survey that had been carried out. Further, over-time , changes that are small and indistinguishable from perfectly normal distributed noise were seen in the data. It was also observed that elements in the dataset were inconsistent with patterns typically found in randomized experiments and surveys.
LaCour & Green (2014 ) claimed that the study had been drawn from a distinct sample of non-random voters in Los Angeles . However, the distribution in the gay feeling thermometer used in the study was identical to one in a national survey data set. The use of a feeling thermometer was also seen as not ideal since feeling thermometers have a reputation of being unreliable when it comes to the representation of data in surveys. It was also observed that the changes in the feeling thermometer seemed consistently perfect which was quite unusual given the fact that the sample size was almost 1000 respondents. It was also noted that the voters the data set claimed had been successfully reached had identical attitudes to the voters in the treatment group that the campaign had not managed to reach successfully . Broockman et al . (2015) argued that this was questionable since traditionally, a difference is often seen.
A summary report was done by Broockman et al . (2015), and the findings were presented to Donald Green , who agreed with the evidence presented. While waiting for LaCour to present his counter-evidence, Green had no choice but agree with the overwhelming amount of evidence that clearly showed that a lot of the work in the survey had been replaced with datasets from previous surveys. This helped to explain the fact that the dataset used by LaCour & Green (2014) was similar to datasets used in a number of National Surveys. When confronted with the issue by his graduate school supervisor, LaCour vehemently denied any wrongdoing. When he was requested to produce the evidence to support his claims, he refused to offer the evidence that would exonerate him from any research malpractice. Thus, Green was left with no choice but apologize for the mistakes and the embarrassment he had caused the field of science.
T he shortcomings identified in the study by LaCour & Green (2014) are increasingly becoming common . Many scholars are dealing with the increased need to publish consistently in order to remain relevant in the scientific world. This has come at a cost due to the increasing use of questionable research methods and other forms of research malpractice. In the above case, there were irregularities in the dataset used by LaCour & Green (2014). Owing to evidence of data replication, it is clear that the scholars did not collect the data as they claimed. These include the baseline outcome data and small over-time changes. The inconsistencies identified with regard to conventional randomized experiments and responses from the survey point towards malpractices on the part of the authors. It is due to such anomalies that the study by LaCour & Green (2014) was not replicable. These anomalies point towards the challenges facing the scientific community today.
References
Broockman, D., Kalla, J., & Aronow, P. (2015). Irregularities in LaCour (2014). UC Berkeley. Retrieved from http://stanford. edu/~ dbroock/broockman_kalla_aronow_lg_irregularities. pdf
LaCour, M. J., & Green, D. P. (2014). When contact changes minds: An experiment on transmission of support for gay equality. Science , 346 (6215), 1366-1369.
