Community Chat

[sam96]

---

@Lila Belle wrote:
400 is usually considered an appropriate and reliable # for a political poll. that makes 1200 significant and probably more reliable.

---

Not the same type of sampling of individuals...not even close.

Political polls have a varied sample space, calls are random, then the actual number of registered voters surveyed, and the margin of error computed.  

This was one single sample space, chosen by the researcher, and not random.  It was statistically different in so many ways, that comparing it to a political poll is comparing apples to oranges.

[Lila Belle]

That journal is legit. The criteria is as reliable as with a political poll.

[Bestdressed]

Polls and studies are very different.

[Maudelynn]

---

@sam96 wrote:

---

@Maudelynn wrote:

It is a statistically significant number. Studies are not done without using subjects in statistically significant quantities.

---

It is not a statistically significant number, and studies are done all the time using insignificant quantities.  The researcher determines the pool of subjects.

How do I know this is not a statistically significant number (besides common sense in comparing 1200 to the total number of children in this age group to the total number on the planet?). I have 2 degrees in mathematics, and have studied statistics. 

---

Are you saying the researchers used flawed numbers to come to their conclusion?  What would be a statistically significant number for the study they were conducting?  And how do you arrive at your number?

[Lila Belle]

---

@Bestdressed wrote:

Polls and studies are very different.

---

````````````````````````````````````````````````

But the results of a questionaire would be the same.

[Maudelynn]

---

@Lila Belle wrote:
400 is usually considered an appropriate and reliable # for a political poll. that makes 1200 significant and probably more reliable.

---

I've worked with smaller numbers which were statistically significant based on what we were looking to study. 

I'm interested in what our math major can come up with when testing the hyposthesis that religion does have/does not have an impact on the empathy level, generousness and acceptance level of children 5-12.

[Bestdressed]

---

@Maudelynn wrote:

---

@Lila Belle wrote:
400 is usually considered an appropriate and reliable # for a political poll. that makes 1200 significant and probably more reliable.

---

I've worked with smaller numbers which were statistically significant based on what we were looking to study. 

I'm interested in what our math major can come up with when testing the hyposthesis that religion does have/does not have an impact on the empathy level, generousness and acceptance level of children 5-12.

---

I would love to see the entire study.  There has to be more to it then:  to play a game in which they were asked to make decisions about how many stickers to share with an anonymous person from the same school and a similar ethnic group.

There would be no way they can conclude the below results, from children sharing stickers: 

The data also show that religious children judged interpersonal harm as being meaner and deserving of harsher punishment than did children from non-religious households. Those findings are consistent with past research in adults showing that religiousness is directly related to increased intolerance for and punitive attitudes toward interpersonal offenses, including the probability of supporting harsh penalties.

[Lila Belle]

---

@Maudelynn wrote:

---

@Lila Belle wrote:
400 is usually considered an appropriate and reliable # for a political poll. that makes 1200 significant and probably more reliable.

---

I've worked with smaller numbers which were statistically significant based on what we were looking to study. 

I'm interested in what our math major can come up with when testing the hyposthesis that religion does have/does not have an impact on the empathy level, generousness and acceptance level of children 5-12.

---

```````````````````````````````````````````````````````````````

^ 5 this !

[sam96]

---

@Maudelynn wrote:

---

@sam96 wrote:

---

@Maudelynn wrote:

It is a statistically significant number. Studies are not done without using subjects in statistically significant quantities.

---

It is not a statistically significant number, and studies are done all the time using insignificant quantities.  The researcher determines the pool of subjects.

How do I know this is not a statistically significant number (besides common sense in comparing 1200 to the total number of children in this age group to the total number on the planet?). I have 2 degrees in mathematics, and have studied statistics. 

---

Are you saying the researchers used flawed numbers to come to their conclusion?  What would be a statistically significant number for the study they were conducting?  And how do you arrive at your number?

---

The beauty of statistics is that a researcher, or anyone for that matter, can manipulate the data to their personal advantage, and call it "statistically significant ".  That's why statisticians make so much money!

The real questions are: against what population did s/he norm the data, if at all? If so, within how many sd's was it? Is the subject group a subgroup or a single population that was hand picked?  What variables were regarded and rejected that affected the outcome?

Without these answers, any discussions as to the relevance of statistical significance is meaningless as it is conjecture.

[Bestdressed]

---

@sam96 wrote:

---

@Maudelynn wrote:

---

@sam96 wrote:

---

@Maudelynn wrote:

It is a statistically significant number. Studies are not done without using subjects in statistically significant quantities.

---

It is not a statistically significant number, and studies are done all the time using insignificant quantities.  The researcher determines the pool of subjects.

How do I know this is not a statistically significant number (besides common sense in comparing 1200 to the total number of children in this age group to the total number on the planet?). I have 2 degrees in mathematics, and have studied statistics. 

---

Are you saying the researchers used flawed numbers to come to their conclusion?  What would be a statistically significant number for the study they were conducting?  And how do you arrive at your number?

---

The beauty of statistics is that a researcher, or anyone for that matter, can manipulate the data to their personal advantage, and call it "statistically significant ".  That's why statisticians make so much money!

The real questions are: against what population did s/he norm the data, if at all? If so, within how many sd's was it? Is the subject group a subgroup or a single population that was hand picked?  What variables were regarded and rejected that affected the outcome?

Without these answers, any discussions as to the relevance of statistical significance is meaningless as it is conjecture.

---

Exactly!  I can devise the same study, with the exact opposite results.