Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 3

When experiencing abusive supervision, a subordinate is usually subject to retaliation against the

supervisor who oppresses them. While the scientists have pointed out that this response has some
negative consequences and can result in an unhealthy workplace environment, Liang, through
her research Righting a wrong: Retaliation on a voodoo doll symbolizing an abusive supervisor
restores justice published in 2018, suggests that retaliation, specifically on a voodoo doll, can
help the subordinate reaffirm their sense of justice. With the research interest in workplace
abusive supervision and retaliation as a reaction to injustice, Liang aims to create a series of
experimental studies to propose a functional theory regarding the impact of retaliation on the
subordinate’s justice perception restoration. While the article goes far extent to demonstrate the
prominent role of retaliation in the context of abusive supervision, the research itself could be
enriched and validated by additional methods of research.

Liang quotes a number of research arguing that retaliation could distress the supervisor-
subordinate relationship and potentially lead to other internal conflicts that result in more acts of
violence in the future. In contrast, there is evidence that retaliation, in general, could be
beneficial in helping restore balance in a relationship. There is also a gap in the methods of
experiment where most of the research was conducted in a field setting or in cross-sectional
designs, which could have some drawbacks such as the limit in drawing conclusions and the
susceptibility to potential endogeneity threat. Therefore, Liang decided to carry out this research
to bridge the knowledge and methodological gaps in previous studies. She hypothesized that
retaliation moderates the positive relation between abusive supervision and subordinate injustice
perceptions, such that the relationship is weaker when retaliation is high rather than low.

The hypothesis is tested in two experimental studies in which the manipulation of abusive
supervisor and subordinate’s retaliation was performed and the implicit justice perception of the
participants was assessed afterward. Besides, coupled with two pilot studies regarding the
injustice word fragment development and abusive supervision manipulation, there are five
robustness checks for endogeneity, implicit justice, and probing ordinal interaction.

Regarding the experimental design, they randomly divided the participants into three groups:
abusive supervision/no retaliation, abusive supervision/retaliation, and a control condition. For
the treatment groups (abusive supervision/retaliation and abusive supervision/no retaliation),
participants were asked to recall an incident in which they experienced abusive supervision.
They were then asked to use the materials (pins, pliers, or fire) on the online voodoo doll.
Finally, their perceptions of justice were implicitly assessed through a word fragment measure
test. The control group was simply asked to complete the word fragment.

The number of word fragments was collected as raw data. They were then divided into two
types: injustice-related words and neutral words based on the study of Hafer (2000) and the
English Lexicon Project database. After that, they plotted the ratio of injustice words to total
words and take it as a measure of implicit injustice perceptions. The data was not only used for
the ANOVA test (one-way and two-way) in two studies but also for the instrumental-variable
estimator, the maximal likelihood estimation with robust standard errors estimator, average
treatment effect, local average treatment effect, and the ordinary least squares estimator in pilot
studies and robustness check. Besides, the data was also used to plot three bar charts with a
confidence interval of 95%. The description of all tables and figures in the article is presented
clearly with enough information regarding the specific test/estimator names, the sample size, the
p-value, and the inclusion of non-compilers in the tests.

The research has done a fairly great job of carrying out the experiments to achieve its original
aims. The experiments are relevant to the hypothesis whereby Liang succeeded in providing
concrete evidence and studies to support the efficiency of two main tests used in two studies:
critical incident technique and word fragment measure. Liang also made considerable effort to
validate her experimental studies. While randomization is performed in the treatment and control
group assignment, the author also tried to exclude all the non-complier, which helps strengthen
the research’s internal validity. Although it is admitted that the ecological validity of the study is
inadequate, it is still a theoretically well-built study with strong internal validity (453), which is
good enough for future prediction or upscaling. Liang also conducted several tests to check if
there is a selection problem with the drop-out participants, which is essential to ensure the
external validity of the study. The study’s reliability was ensured with the systematic replicates
of the study 2 in which both supervisory treatment and levels of retaliation are manipulated.
Liang did anticipate potential issues of her design and execution in the limitation section and
provided readers with a multi-dimensional and comprehensive view of the problem when
comparing her work with others in the field.

Having said that, there are several flaws in Liang’s research which she could have done better to
strengthen her paper’s validity and reliability. First, regarding the internal validity, the control
variables between experiments were not mentioned. For example, the temperature, the noise
level in the room where participants complete the experiment, and the period of time in the day
during which the experiment was performed could affect the participans' feelings and the result
of the study. Thus, not reconstructing the identical control variables between studies could result
in a threat to the research’s internal validity. While knowing the treatment could affect the
participants’ response, blindness was also not mentioned in the two main studies. Especially for
the control group who just complete one word fragment test, it would be better if Liang designed
the experiment for the control group that camouflage the treatments to minimize the potential
bias. Moreover, while claiming to exclude the incomplete and incorrect words in the fragment
test, the author failed to demonstrate the number of excluded words and test if those are
significantly important to the research. Second, regarding the external validity, since the
participants, especially in the study 2, all come from one university, the data is then not strongly
representative as the participants in one specific community could share some characteristics that
potentially influence the result. Also in the study 2, the difference in predicated margin between
with and without non-compliers test is quite significant (22.82%), which demonstrates the effect
of drop-out participants on the external validity of the research. Third, regarding the reliability of
the research, results of the systematic replicates of the research are not immensely strong and
convincing with the p-values larger than 0.05. Besides, according to Israel (1992), the sample
sizes of both studies are not large enough, which could lead to an increase in the error margin.

There are some potential issues with the design and execution of the research regarding their
claim and description. In the paper, Liang failed to provide the relation between the implicit
justice perception and the true justice perception in the subordinate’s mind - which is the subject
she is researching on. By pointing out that “perceptions of injustice involve the unconscious
activation of injustice-related concepts”, she admitted that provocation of injustice-related
concepts in mind is just one of the subordinate’s signs. Therefore, if the injustice concepts are
not provoked (which results in the completion of the injustice-related words in the word
fragment measure), it does not necessarily mean that they do not have injustice perception
toward the abusive supervision. This crucial point could also potentially harm the validity of the
research if not carefully taken into consideration. Regarding the voodoo doll assignment, the
author did not provide any research to back up the instruction “label the voodoo doll with their
supervisor's initials”. It is not clear why they used the initials, which potentially results in the fact
that participants did not take the doll as their supervisor in the retaliation process, instead of full
name. Finally, Liang claimed to “yield an optimal target/neutral words balance” between the
injustice and neutral words, however, in Appendix A, the fragment _RON_ has up to 7 neutral
words and just 1 injustice word, whereas the remaining’s ratio is 1:1 or 2:1. This imbalance
between the number of words in two groups could result in a bias towards a target or neutral
words in the absence of an experimental manipulation.

Overall, the paper Righting a wrong: Retaliation on a voodoo doll symbolizing an abusive
supervisor restores justice of Liang was written in a logical and rigorous way, which is an
important premise for future research regarding this topic. That said, other actions could be taken
to improve the validity and reliability as well as mitigate the potential issues of the research.

Reference
Israel, G. D. (1992). Determining sample size.

You might also like