Overall, I thought this paper was a more
focused look at a specific topic relating to paternal neuroendocrine function and
had fewer experiments that were tailored to the researchers’ hypothesis. The assays
were more specific compared to the last paper we read and had fewer variables
to manipulate and test for. This made for a slightly more in-depth study, with
specific conclusions drawn from limited assays, as opposed to the last paper,
which was broader and had more of a wider range of results and experiments. I
thought that this was appropriate, as the previous paper seemed a bit more
exploratory, while this paper was supported by previous experiments with
similar results.
However, one issue I had with this paper was
its use of figures and tables – specifically, in the way they listed some statistical
information regarding their behavioural assays in table format. It made it very
difficult to read and understand, and I think graphs with error bars would have
made for easier reading rather than having to parse through the columns of the
tables. In addition, I thought their results section was very statistic-heavy,
using words like ‘significant interaction’ and ‘main effect’ that I thought
were harder to put into context. This made it hard to interpret the results of the
experiments, leaving me confused about what they were actually able to prove. I
understood eventually that this was just the nature of the experiment, and that
some of the assays they ran had a different effect between virgins and parents –
for example, in the administration of the tamoxifen drug, there were different
responses between parents who did and did not receive the drug as well as
differences between controls and parents who did receive the drug. This ‘double
result’ assay made for a little bit harder interpretation, but eventually I was
able to understand what the authors were talking about (I think, I could be completely
wrong).
Another issue I had
with the study was the number of animals that they needed to exclude from
analysis for a variety of reasons, whether it was brain slices that needed to
be excluded from assays, or poor behavioural responses from animals that led to
them being removed from analysis. While I understand that problems may occur
throughout experimentation, they had n values as high as 13 being excluded from
particular studies, which made me wonder whether there was a bigger issue to be
addressed with the assays that were being used. In addition, when it came to
their estradiol concentration assay, the authors write that ‘concentrations did
not meet assumptions of normality’ and decided to normalise the values instead.
If there was also an issue with the estradiol quantification, I would again
question their assay of choice rather than normalise the values to find
results. Together, these results make me slightly worried about the care that
the authors put into choosing assays, and whether or not they were fishing for
results and changing the way they analyse data.
Yes, they did exclude quite a few subjects, and their group sizes were pretty different from each other...good job noticing that
ReplyDelete