Second Thought
April 22, 2025
At a time when the general public is distrustful of science, mainstream media and politicians, it was heartening to read the timely article by Anne McIntosh in the Second Thought column. Her analysis and interpretation of inferential statistics and polling helps diligent readers sort out fact from fiction within the “noise” of the surrounding political hogswallow.
I would also add that reputable pollsters use valid standardized, stratified random sampling; namely, if 20 per cent of the population are females between the ages of 20 and 30, then the poll randomly selects 20 per cent of the sample strata as females between 20 and 30 years of age and so forth. Similarly, if our population distribution is 10 per cent of males over the age of 65, then the random sample would consist of 10 per cent male over 65 years.
Stratified random sampling ensures that the poll results can then be extrapolated to the population in a valid manner. The standard confidence interval is 95 per cent, which means that one can be certain that the sample results are correct within a narrow margin 95 per cent of the time…excellent odds for decision making.
The Alberta UCP has a penchant for polls that rely upon a method of people selecting themselves by phone in, write in, phone in town halls or online polls. This methodology provides meaningless results that are a waste of money, totally misleading and, a cynic might conclude, to support or legitimize a political agenda.
Don’t be fooled when uninformed (statistically naïve) editors, politicians and opinion writers make reference to the results of such polls.
Lynn Clark,
Camrose
At a time when the general public is distrustful of science, mainstream media and politicians, it was heartening to read the timely article by Anne McIntosh in the Second Thought column. Her analysis and interpretation of inferential statistics and polling helps diligent readers sort out fact from fiction within the “noise” of the surrounding political hogswallow.
I would also add that reputable pollsters use valid standardized, stratified random sampling; namely, if 20 per cent of the population are females between the ages of 20 and 30, then the poll randomly selects 20 per cent of the sample strata as females between 20 and 30 years of age and so forth. Similarly, if our population distribution is 10 per cent of males over the age of 65, then the random sample would consist of 10 per cent male over 65 years.
Stratified random sampling ensures that the poll results can then be extrapolated to the population in a valid manner. The standard confidence interval is 95 per cent, which means that one can be certain that the sample results are correct within a narrow margin 95 per cent of the time…excellent odds for decision making.
The Alberta UCP has a penchant for polls that rely upon a method of people selecting themselves by phone in, write in, phone in town halls or online polls. This methodology provides meaningless results that are a waste of money, totally misleading and, a cynic might conclude, to support or legitimize a political agenda.
Don’t be fooled when uninformed (statistically naïve) editors, politicians and opinion writers make reference to the results of such polls.
Lynn Clark,
Camrose