How should I write this. There are many results.
we remove them them from our results' statistics
we remove them from our results statistics
we remove them from our result statistics
The first sentence is valid: results' is a possessive plural, so the sentence is equivalent to:
we remove them them from the statistics of our results
In the second and third sentences, result(s) and statistics are combined to form a compound noun. statistics is the most significant noun, so all other nouns have to be singular. result must therefore be singular, so only the third sentence is correct.
This answer takes into account the actual use case of collating results and preparing statistics of those results, not solely the English usage.
There is some implied information here. You are collecting results (from observations) and from the results you are making statistics to understand those observations.
If that statement is true, then the statistics are the summary, and each of the results that you have is a datapoint within those statistics.
If you then want to exclude certain results (where the result appears to be an anomaly, or the observations appears not to be valid) and have already explained which results are to be excluded then you should say:
we remove those results from our statistics
From a language perspective, the statics are a summary of the results, they are not owned by the results so writing "results' statistics" is incorrect.
Using either "results statistics" or "result statistics" as a compound nound to mean the statistics which summarises the set of results would both be understood, but may need additional context to be clear. You should also take into account the use of plurals in a compound noun, as the dominant noun is statistics you should use result (singular) instead of results.
to remove x from statistical results.
statistical results is the proper term here. There is no need for a possessive at all.
As shown in this text:
Changes in data are manifested "immediately" throughout the spreadsheet as a result of the cell formulas. From a pedagogical perspective, this immediacy can be a great advantage. We can examine the effect of outliers or misrecorded numbers on our statistical results, including graphs. Unfortunately, an important exception may be "analysis boxes," which do not get updated and are a potential source of misleading output. Immediate updating is not, however, the preserve of spreadsheets. Notable among statistical packages are Data Desk and Lisp-Stat.
So, for example, one might write: