I made a video on how to detect and correct mistakes in your Zotero items

edited October 27, 2020
Hope it's useful to some!

  • edited October 27, 2020
    Allow me to recommend your method to detect and correct muddy abstracts from publisher metadata.

    In short, @MavropaliasG presents a video demonstrating flaws in the title and abstract text imported from certain publishers. (By and large, these are things that Zotero's translators cannot correct [see link to thread below].) MavropaliasG demonstrates how a Zotero search can identify your records that have these problems and he then shows that it is easy to use Zotero to quickly link to the article on the publishers' websites to copy the title or abstract from the website and paste the text into a Zotero record. I've long used a less elegant way to identify these problems in Zotero records.

    My own process had been to do the abstract cut/paste from the publisher's website to Zotero based on my knowledge of the publisher. (With some publishers the abstracts they provide in their metadata are always so cluttered as to make them almost useless.) You mentioned MDPI but there are several others that place web-formatting html tags in their metadata. MDPI and some other publishers encode math symbols, Greek characters, and accented extended alphabetical characters. Zotero accepts the pasted material and keeps the human readable form. I don't know if behind-the-scenes Zotero is storing these characters with the formatting hidden; but I really appreciate that Zotero allows me to keep the text attractive and readable after a paste. Some publishers use tiny image files for math symbols but these can be easily recognized and properly edited.

    edit: for some reason my auto-correct was insisting that the word "paste" should instead be "past".

    edtt2: example of the messy text from the article webpage header:

    60°·s<sup>−1</sup> using a Kin Com isokinetic dynamometer. Large correlations were revealed between COD performance (time to complete 5 m approach, 180° turn, 5 m return) and ECC-EXT (<span class="html-italic">R</span> = −0.674) and ECC-FLEX (<span class="html-italic">R</span> = −0.603). Moderate to large correlations were observed between approach velocity (AV) and COD performance (<span class="html-italic">R</span> = −0.484) and ECC-EXT (<span class="html-italic">R</span> = 0.724). Stronger participants (<span class="html-italic">n</span> = 9) recorded significantly (<span class="html-italic">p</span> < 0.05) faster AV (4.01 ± 0.18 vs. 3.74 ± 0.24 m·s<sup>−1</sup>, <span class="html-italic">d</span> = 1.27) and a greater reduction in velocity (−1.55 ± 0.17 vs. −1.37 ± 0.21 m·s<sup>−1</sup>, <span class="html-italic">d</span> = −0.94) during penultimate contact than weaker (<span class="html-italic">n</span> = 9) subjects. Greater ECC-STR is associated with faster COD performance in female

    Actually, it is even worse than that because (although I enclosed my pasted header text within code tags) the various span tags within open span tags fooled this bbs software and truncated the worst parts of my example above. [I didn't think that html formatting tags were "allowed" in page headers and that modern browsers should reject such mess. Maybe my show-source is converting the lt gt to angle brackets.] Still not as bad as the example shown in the linked thread below:

    See also: https://forums.zotero.org/discussion/85250/import-of-journal-articles-with-lot-of-html-formatting-tags-in-abstract#latest
  • <sup>-1</sup> can also be done as ⁻¹ (\u207B\u00B9) so that can be done without HTML markup.
Sign In or Register to comment.