Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 2

HW #4 Kristina Bryant

1. Research Design- Trying to determine the research design for this study has been no
easy task. I immediately gravitated toward the Pre-experimental since there was no
random assignment and the groups of followers that were tested could not have
equivalence. However, there did not seem to be an option for a research design that didnt
involve a treatment. This study did not give any kind of treatment to either the
leaders in between their pre and post tests nor did they give any treatment to the
followers in their post tests. The Quasi-Experimental research methods were also an
options because they again allowed for no randomization (subjects are not randomly
selected), researchers lack control in avoiding events and manipulators to their
respondents, but again I could not find a method that did not involve a treatment. After
looking into additional research designs this could be turn into a longitudinal study if the
leaders were to have a workshop about leadership and the followers were tested again
after months or years. The leaders with their new found knowledge from the workshop
could be considered a treatment to the followers.
2. Two potential threats to Internal Validity
a. Diffusion- since there were multiple followers surveyed for each leader, it is
extremely possible that those being surveyed could have discussed their answers,
discussed the leadership of their leader, and these discussions could have led to
certain respondents changing their views of their leader and/or strengthening their
views and giving a higher score on the Likert scale than before the discussions.
b. Selection Bias- Since the leaders were asked to solicit 4-6 followers, this can have
all kinds of effects on the data. The leaders potentially asked followers that might
view their leadership more favorably or selected followers that might be more reliable
in finishing the measure. This is not a random sampling, nor is there evidence of a
control group. Also the article did not indicate if any demographic or additional data
was collected about the respondents.
3. Internal Validity Suggestion- My suggestions for the researchers would be to
standardize the conditions in which the followers took their measurements
(Instrumentation), assign a control group and find a way to randomly assign respondents.
I understand that this is not practical for the researchers, but there are far too many
variables and interrupters that could occur in how the researchers collected their data via
surveys handed out by the leaders and returned by mail. Like I mentioned above, having
the leaders choose the followers that would answer the survey questions really brings the
validity of this entire study into question.
4. One potential Threat to External Validity- As the author already noted, the sample
sizes of respondents was small and specific to civil service positions and therefore not
easily generalizable. Additionally, I think that experimenter effects may have played a
role in threatening the external validity of this study. There was no blind data collection
as the leaders themselves administered the measures to their followers. Additionally the

leaders who collected the data were fully aware of the purpose and content of the study
having taken emotional intelligence pre and post tests. Reactive effects of experimental
arrangements- participants were aware that they were participating in a study and this
might have affected their responses.
5. External Validity Suggestion- My suggestion for the researchers would be minimize the
threat of experimenter effects by having and training someone other than the leaders to
administer the measure to the followers.
6. Cronbachs Alphas
a. Altruistic calling- -8.61
b. Emotional Healing- -9.139
c. Wisdom- -7.562
d. Locus of Control- -6.630

(.78)
(.66)
(.85)
(.71)

Well obviously there is something wrong with these figures as they should all be
between 1 and 0. But I followed the formula in the book on page 257 that shows you
how to calculate the Cronbachs alpha using the data from Table 1 in our study.
Since the figures are incorrect I cannot determine if they meet the rule of thumb
criteria. However, if the figures were correct and were between one and zero, I would
be able to see which of the measures were above or below a .70. The higher the
Cronbachs alpha, the more likely that the concepts have internal consistency.
Rule of Thumb citation- (Nunnally, 1978)
I feel like the figures given on Table 1 that intersect might be the correct figuresif
that is the case, all of the items meet the rule of thumb except for emotional healing
which is .04 points shy.
7. Correlational Relationship- The correlational relationship between Organizational
Stewardship and Emotional healing is at a .70, making it statistically significant. That
means that a correlation does exist and those two factors are traits that measure emotional
intelligence and can predict servant leader qualities.
8. Locus of Control- the mean and the standard deviation for both the leaders set of data

and the followers set of data was exactly the same at 2.85 and .59 respectively. However
when comparing the means of the other factors to Locus of Control the leaders rated it
lower on average than the other factors putting it in 5th place among 7 factors. In the
followers group it was considerably higher tying for 3rd place among 7 factors. The
researchers did not do a good job of defining Locus of Control nor did they mention if
they were measuring internal or external Locus of Control. Based on these figures I
conclude that the followers find Locus of Control a more important skill/trait to have than
the leaders do.

You might also like