Welcome to DU! The truly grassroots left-of-center political community where regular people, not algorithms, drive the discussions and set the standards. Join the community: Create a free account Support DU (and get rid of ads!): Become a Star Member Latest Breaking News General Discussion The DU Lounge All Forums Issue Forums Culture Forums Alliance Forums Region Forums Support Forums Help & Search

ck4829

(35,096 posts)
Mon Jan 17, 2022, 03:37 PM Jan 2022

Statistics 101 - The larger a sample size, the more legitimacy your data has

Having a poll or survey that asks a question to 14 people has less credibility than a poll or survey that asks that question to 100 people and that in turn has less credibility than a poll or survey that asks that same question to 10,000 people.

Just a little something to remember... legitimacy, credibility... when Republicans want a "red wave" this year and in 2024 when everyone is not able to vote. Do we have an appropriate sample size?

5 replies = new reply since forum marked as read
Highlight: NoneDon't highlight anything 5 newestHighlight 5 most recent replies
Statistics 101 - The larger a sample size, the more legitimacy your data has (Original Post) ck4829 Jan 2022 OP
Generally true as long as the sample is truly RANDOM. Sample size determines the POWER hlthe2b Jan 2022 #1
Provided the sample is random. nt Xipe Totec Jan 2022 #2
I took Statistics for Social Science. multigraincracker Jan 2022 #3
All other things being equal, and when is that the case? L. Coyote Jan 2022 #4
Statistics was a hoot, I think I got a C grade pwb Jan 2022 #5

hlthe2b

(102,509 posts)
1. Generally true as long as the sample is truly RANDOM. Sample size determines the POWER
Mon Jan 17, 2022, 03:42 PM
Jan 2022

of the study--i.e., the ability of the study to identify statistically-significant differences and correlations where they exist. However., a non-random sample--one which introduces sampling bias--is not more valid no matter the size.

L. Coyote

(51,129 posts)
4. All other things being equal, and when is that the case?
Mon Jan 17, 2022, 04:49 PM
Jan 2022

Statistics 102 - Statistics is complex. Sampling, inferential statistics, always has error compared to descriptive statistics, polling everyone instead of sampling. So, given sampling, you don't know if what you said is true or not, just that is is statistically more probable, in other words using "credibility" is correct and certainty would not be.

In political polling, objectivity, independence, lack of bias, and ability to correctly sample are huge factors, and there are a lot of bad actors producing skewed polling to sway public opinion by purposefully designing polls to reflect desired results, a very easy thing to do. That, and bad actors such as #FailedCoupGuy (double entendre intended) just pay people to lie about polls.

pwb

(11,308 posts)
5. Statistics was a hoot, I think I got a C grade
Mon Jan 17, 2022, 05:00 PM
Jan 2022

and I put in some effort. Your right, ask 3 people how they like Joe and if 2 don't like him = 33& approval rating. Ask 100 people and he might be 60% approval. I am not seeing a red wave again ever. The Blue builds.

Latest Discussions»General Discussion»Statistics 101 - The larg...