The familiar average IQ of 100 is a feature of a scoring scale. It is not the percentage of questions someone answered correctly, and it does not require every group that takes a test to average exactly 100.
A mean belongs to a group
The mean is the total of a group's scores divided by the number of scores. If five illustrative results are 80, 90, 100, 110 and 120, their sum is 500 and their mean is 100.
Now imagine that only the two highest scorers return for a second activity. Their original results average 115. The reference scale has not changed; the selected group has. Website visitors, volunteers and members of a particular course can differ from the group used to establish a scale.
Center and spread are separate
Two groups can average 100 while having very different distributions. The scores 98, 100 and 102 have the same mean as 60, 100 and 140. The second group is much more spread out.
This is why an average does not establish a standard deviation, a bell-shaped distribution or a test's accuracy. Each needs its own calculation or evidence. Moving a scale's center is a numerical operation; showing what its scores mean for people is a broader task.
Why starp's calibration average is 100
starp used 12,700 simulated completed assessments to tune a fixed score mapping. Their calibrated score total is 1,270,000. Dividing by 12,700 gives exactly 100. The mapping retains a minimum of 42 and a maximum of 144.
This figure describes that archived simulation. A later group can average higher or lower. The system does not change an individual's result simply to force today's visitors to average 100.
What to check when you see an average
Ask which group was counted, when it was measured and which scoring version was used. Those details turn an attractive headline number into an interpretable statistic.