Objectives  With the increasing application of tag maps, there is an urgent need for research on the evaluation of tag weight expression strategies. Tag weight differences in tag maps are usually reflected by different font sizes. One of the common strategies is evaluated, where the sequence of font sizes is an arithmetic progression.

  Methods  The five countries with different, representative shapes are selected as study cases. Five corresponding tag maps are produced by randomly generated tags in alphabetical order from left to right and top to bottom. On this basis, a controlled experiment is conducted. Two application scenarios of unpurposed free browsing and purposeful reading analysis are set, and tag selection, recognition/search, recalling, and subjective evaluation tasks are assigned. The subjects' eye movement data and some other derived data are collected and statistically analyzed by the Kruskal-Wallis test and MannWhitney U test methods.

  Results  The results show that: (1) Tags of different sizes do not show significant differences in terms of visual salience, visual attractiveness, weight recalling and search efficiency, reading efficiency, and cognitive load in recognition/search tasks. (2) Tags with font size at the upper level are easier to be recognized/searched and of higher interest to subjects than those with font size at the lower level, but it does not mean that tags with larger size are more likely to be recognized/searched and of higher interest to subjects. (3) The overall evaluation of tag maps using font size strategy of arithmetic progression is at a good level.

  Conclusions  The paper is helpful for map designers to further understand the characteristics of font size strategy. The font size is not a panacea in the tag weight expression of tag maps. In addition to font size, tag weight differences in tag maps can be shown by combining other visual attributes such as brightness, color, typeface, etc., or accompanying auxiliary charts according to the actual application. The interaction of the shape, area, the density of tags, and other visual attributes with the font size and some other font size strategies need to be studied for different users.