Name:   MAJ USA RET - Subject:   If the difference between n and n-1 matters to you Date:   5/18/2022 5:14:30 PM

Any question on a reliable survey must meet these two tests:

1 – Is it a valid question?

EXAMPLE:

"Do you support clinical abortion?"

          [ ] No, not for any reason

          [ ] Yes, but for a limited time after conception and only for the mother’s health and in cases of rape or incest

          [ ]   Yes, without exception

          [ ]   I am undecided

EXAMPLE of a bad survey question:

"Democrats openly support abortion. Are you in favor of the murder of unborn children?"

[ ] Yes, it is okay to kill clumps of cells in the womb.

[ ] No, it is immoral to suck out the life of an unborn child.

2 – Is it a reliable question? Do the results accurately measure the opinion of the total population?

In the first example above, there is no political, religious, or moral stipulation. There are no unnecessary adjectives or adverbs. There are enough choices that everyone can answer the survey question.

In the second example, a political statement is made to support a question burdened with hyperbole. The participant is faced with only two choices. The two choices are neither factual nor an accurate measure or the participant’s true position.

A reliable survey is based on valid and reliable questions, delivered to a representative sample of the population, in a distinct period of cultural climate.  A survey on abortion is 1 - not a valid or relable measure on the question of Roe v Wade and 2 - a survey on abortion which started before the SCOTUS leak and ended a week after is not a valid or relable measure on the percieved legality of abortion.

1 – representative sample: “A sample in which the characteristics of the individuals closely match the characteristics of the overall population.”

To measure the sentiment of voters, you have to insure that your sample takes into consideration the makeup of the voting population - e.g. equally male or female, citizen of the United States, equally liberal or conservative (not Democrat or Republican), broadly representative of racial makeup, broadly representative of financial status, broadly representative of educational background, etc.

I used the terms “equally” and “broadly.” If you noodle the sample too much you can unintentionally skew the results.

Simple RANDOM samples are usually representative of the population of interest because every member has an equal chance of being included in the sample.

If you do not understand why a survey of register voters is NOT a representative survey of the general population, then stop reading here.

2 – sample size: This is, for most folks, an arduous project. Let’s just say that there are a plethora of sample size calculators out there on the internet.

Another (simplistic) example of the relative order of magnitude for sample size is:

To accurately represent 1 million people, with a confidence level of 95% the number of returned RANDOM surveys is 384.

OH! And by the way be sure to include your voluntary contribution of

[ ] $50   [ ] $100   [ ] 250   [ ] other____ to the Republican National Convention