Factors affecting the decision to act dishonestly
Corruption is something everybody has been talking about and some are even upset about. Unfortunately corruption is a subject which people understand very little about. This is because corruption involves dishonest actions that are both immoral and illegal, which prompts people to conceal it. A number of previous studies attempted to look at how corrupt practices happened by investigating backward, which sometimes led to dead-ends as evidence had been removed to avoid prosecution. To better understand and prevent corruption, perhaps we need to look at the beginning -- what makes a person decide to act dishonestly in the first place?
With funding from the Thailand Research Fund, the Thailand Development Research Institute conducted a study on factors affecting the decision to be dishonest, with a focus on transparency and monitoring -- the two factors associated with corruption prevention measures. A controlled field experiment, which allows researchers to observe the natural behaviour of samples in a created environment, was used as a research instrument. Samples came from 425 undergraduate students from various universities in Bangkok.
The research team created a scenario in which each sample was asked to look for spelling errors in a school's typing exam in exchange for a half-day's wage of 200-300 baht, depending on the number of spelling mistakes each of them could find. The exam's text contained general content and a vocabulary that required no specialised knowledge to read. The samples only needed to compare the typing exam side-by-side with the provided correct copy, and circle and report the number of words misspelled. It was a time-consuming task, but should not be difficult for undergraduate students who can read Thai language.
Boonwara Sumano is a research fellow with the Thailand Development Research Institute (TDRI). Policy analyses from the TDRI appear in the Bangkok Post on alternate Wednesdays.
At the beginning, each sample was given an envelope containing a typing exam, a correct paper to check spelling, and 300 baht, allowing 200 baht for showing up and an additional five baht for each spelling mistake they could find. The samples were told if they found 20 mistakes on a paper, they could stop marking, collect the 300 baht and go home. If they found less than 20 mistakes, the samples needed to multiply the mistakes found times five baht, collect their entitled wage and return the rest of the money in the given envelopes. What the samples did not know is every exam had 10 spelling mistakes, so an honest sample would return 50 baht to the envelope, assuming they found every mistake.
Samples were randomly divided into four treatments, which had different levels of transparency and monitoring. Low transparency was when samples were asked to put the marked paper and remaining money in a sealed envelope, and leave the envelope in a box outside the experiment room, whereas high transparency was when samples were told to report the result of their marking with a researcher in another room, who posed as the employer commissioning the task.
Low monitoring was when samples were told the possibility of their marked papers being rechecked later was 10%, whereas high monitoring were told the likelihood was 60%. Every treatment had a combination of transparency and monitoring, e.g. the first treatment had low monitoring and low transparency, the second scenario had low monitoring and high transparency, and so on. Before they started marking, samples were informed about the conditions of their treatment. This information helped shape their decisions while marking. After the experiment, the research team used a survey to collect additional data about the samples such as age, gender, experience and perceptions about corruption. This experiment aimed to see how transparency and monitoring affected the decisions of samples in marking and returning money.
In this experiment, dishonest actions included (a) reporting a number of spelling mistakes other than 10, either to get more money or to quickly collect the money without fulfilling the assignment, and (b) collecting more money than one is entitled.
The results were remarkable. Of 425 samples, only 16 students correctly reported finding 10 spelling mistakes and returned 50 baht. This group was found mostly in the high monitoring and high transparency treatment. The research team used the minimum time these 16 samples took to finish marking (i.e. find all the mistakes) as a benchmark to identify shirking behaviour.
Forty-four samples reported total spelling mistakes of other than 10, mostly in the range of 11-16 words, indicating an intention to collect more money. Some 85 students left earlier than the benchmark time, which demonstrated a tendency to shirk their responsibility and not complete the task for which they had been paid. Another 17 samples were found to take more money than they were entitled to, mostly in the low transparency and low monitoring treatment, with one sample taking all 300 baht.
Statistical analysis shows high transparency has a significant effect on the decision to be honest, while participants respond indifferently to the different levels of monitoring. This might be because samples perceived there was still uncertainty of disclosure in the high monitoring scenario, compared with high transparency, which led to immediate exposure to someone else. This suggests transparency can prevent the decision to act dishonestly and should be promoted in government offices as an anti-corruption priority.
Additionally, some samples reported being offered or receiving money to cheat on marking exams or to cut the queue, which are called petite corruption. This indicates anti-corruption policy should not overlook these small-scale corrupt acts because they are the type youth have experienced.
TDRI Research Fellow
Boonwara Sumano is a research fellow with the Thailand Development Research Institute (TDRI).