• BlanketsWithSmallpox
    link
    English
    10
    edit-2
    3 months ago

    … how is a linear scaled graph on both the X and Y axis misleading when the point of the graph is to show change over time and percentage within a small range of about 25% - 55%? The extra 70% is useless.

    A Y axis graph going from 0 to 100 results in a squashed graph that’s hard to read.

    Do you just read graphs without looking at the scale or something? It literally has the data points listed on each fucking dot FFS lol.

    There is literally, not even metaphorically, nothing misleading about this graph.

    Both of you quit using that word. It does not mean what you think it means.

    • @Thrashy
      link
      English
      1
      edit-2
      3 months ago

      This seems an… overly-vitriolic response.

      Also you’re wrong. :P

      Look at it this way: in the context of the data being shown here, the relevant reference points are 0% and (arguably) 100%, or at least a point somewhere equidistant from the top of the line as the ~30% low point of the line is to zero. Casually glancing at the chart, a viewer who doesn’t take time to look at the scale and the labeled points would take away:

      A large majority of college-age men used to binge drink, and now almost none do!

      Instead of what the data is actually showing, which is

      Half of college-age men used to binge drink and now only three in ten do, while about a third of college-age women have consistently binged.

      I don’t think the chart designers are being intentionally misleading, but cutting out half of the 0%-100% range means that the graphics are telling a different story than the labels are, and outside the context of a scientific paper not everybody is going to take the time to scrutinize the labels. Omitting the high and low ends of the range also exaggerates the difference between the two lines, since the graph coincidentally cuts off just below the relatively flat line for female binge drinking right after the line for male binge drinking crosses it on the right.

      Besides which, for the purposes of the story showing at least the range from 0%-60% wouldn’t obscure the overall trend – there’s not a lot of noise in the data, and barring the odd spike in female binge-drinking between '14-'15 – that critically, doesn’t appear to be the subject the of article this comes from – there aren’t any smaller-scale trends or oddities in the data that demand scrutiny. Squashing the Y-axis a bit to tell a truer story about the absolute values of the data wouldn’t obscure the message of the graph in any meaningful way.

      • BlanketsWithSmallpox
        link
        English
        1
        edit-2
        3 months ago

        the relevant reference points are 0% and (arguably) 100%

        … The relevant reference points are the ones in the graph which absolutely should be tailored to the dataset. Not ones you arbitrarily prefer.

        but cutting out half of the 0%-100% range means that the graphics are telling a different story than the labels are

        … Again, you deciding that the graph should be 0-100% is just… Pedantically dumb lol.

        It’s literally not misleading.

        You jumping to conclusions by really putting the ass in assumptions without reading is just bad comprehension skills.