0:00:00.000,0:00:02.193
(female narrator)[br]This is just a couple of tips

0:00:02.193,0:00:04.894
on how to use the [br]significance test flow chart.

0:00:04.894,0:00:11.738
As you go through each problem, this[br]flow chart can help you to determine

0:00:12.298,0:00:15.908
what type of significance test you [br]should perform on that problem.

0:00:15.908,0:00:18.858
The first thing you want to do is you[br]want to read the problem.

0:00:18.858,0:00:21.871
When you're done reading the [br]problem, you need to ask yourself,

0:00:21.871,0:00:24.778
"What was the data collected from [br]each member of the sample?"

0:00:24.778,0:00:27.605
For instance, did you collect [br]a weight, or a height,

0:00:27.605,0:00:29.606
or did you take their temperature,

0:00:29.606,0:00:32.100
or did you count the number [br]of dogs that they had?

0:00:32.100,0:00:35.355
That would be numerical data. [br]When you read the problem,

0:00:35.355,0:00:40.223
did you see some key words like mean,[br]or average, or standard deviation.

0:00:40.223,0:00:44.920
All of those things indicate that you[br]have a quantitative data problem.

0:00:45.510,0:00:52.023
On the flip side of that, were there some[br]keywords like proportion, percent, rates,

0:00:52.023,0:00:55.752
or was each member of the [br]sample asked a yes or no question,

0:00:55.752,0:00:59.108
such as, "Have you had [br]a heart attack? Yes or no?"

0:01:00.118,0:01:02.769
Maybe they were asked, "Are you [br]overweight? Yes or no?"

0:01:02.769,0:01:04.968
"Do you support the [br]president? Yes or no?"

0:01:04.968,0:01:07.967
So that type of a thing. That [br]would be qualitative data.

0:01:07.967,0:01:10.364
Once you determine the type of data,

0:01:10.364,0:01:13.056
then you follow the decision [br]tree down that side.

0:01:13.056,0:01:17.867
For instance, if you have quantitative[br]data, then you would ask yourself,

0:01:17.867,0:01:24.415
"Do I have two populations or [br]do I have one population?"

0:01:24.415,0:01:28.632
Now remember, an easy way to tell[br]whether you have one or two populations

0:01:28.632,0:01:31.464
is to look for the samples [br]that are given in the problem.

0:01:31.464,0:01:34.854
If you're only given one sample,[br]then you only have one population.

0:01:34.854,0:01:37.880
If you're given two samples, [br]then you have two populations.

0:01:37.880,0:01:41.996
You have to be given the complete[br]information about both of those samples,

0:01:41.996,0:01:47.296
like the sample mean and sample standard[br]deviation, for two separate samples

0:01:47.296,0:01:49.695
in order to determine [br]you have two populations.

0:01:51.945,0:01:55.412
Once you determine how [br]many populations that you have,

0:01:55.412,0:01:59.969
if you have one population, you're going[br]to do the t-stat one sample.

0:01:59.969,0:02:02.278
If you have two independent populations,

0:02:02.278,0:02:05.627
then you're going to do the [br]t-stat two samples.

0:02:05.627,0:02:11.293
Keep in mind that you need to determine[br]whether you're going to pull the variances

0:02:11.293,0:02:13.177
or not pull the variances.

0:02:13.177,0:02:16.242
Down here, we have the little [br]tip on how to decide that.

0:02:16.242,0:02:19.509
It's just the ratio of the larger [br]sample standard deviation

0:02:19.509,0:02:22.129
divided by the smaller sample[br]standard deviation,

0:02:22.129,0:02:25.417
and looking to see if it's greater[br]than two, then you remove

0:02:25.417,0:02:28.247
the pulled variances check mark, [br]and if it's less than two,

0:02:28.247,0:02:31.567
then you keep the pulled [br]variances check mark.

0:02:31.567,0:02:33.917
Now remember if you have [br]two dependent samples,

0:02:33.917,0:02:39.780
the two samples are related to each other,[br]such as a before and an after,

0:02:39.780,0:02:45.058
a pre and a post test. The two samples are[br]related because it's the same person

0:02:45.058,0:02:47.972
taking the pre-test and the [br]same person taking the post-test,

0:02:47.972,0:02:50.965
or maybe you're testing [br]two different kinds of tires,

0:02:50.965,0:02:53.848
so you take a car and you [br]drive it with the first tire,

0:02:53.848,0:02:57.101
and then you take the same car and you[br]drive it with the second tire.

0:02:57.101,0:03:00.671
Those two samples are related because[br]it's the same car driving both times.

0:03:00.671,0:03:03.497
If they're dependent samples, [br]then you have a choice

0:03:03.497,0:03:06.629
of either finding the differences [br]between the two samples

0:03:06.629,0:03:13.678
and doing a one sample t-test, or under[br]t-stats you can do the paired t-test.

0:03:15.098,0:03:19.676
If you have more than two [br]populations, we only have one test

0:03:19.676,0:03:24.242
where we had more than two populations[br]and that was from chapter 14,

0:03:24.242,0:03:28.966
and that was where we had three, four,[br]five, six, as many populations

0:03:28.966,0:03:34.192
as were needed, and that was to perform[br]the one-way ANOVA test.

0:03:34.922,0:03:39.975
Now, if you have two quantitative[br]variables, like you were collecting

0:03:39.975,0:03:49.089
the height of a person and [br]the weight of a person,

0:03:49.089,0:03:56.511
and you were asked to decide if the [br]variables were dependent,

0:03:56.511,0:03:59.548
or associated, or independent, if [br]you saw those words in there,

0:03:59.548,0:04:03.531
but you have two quantitative [br]variables, a height and the weight,

0:04:03.531,0:04:06.914
then that would be a [br]regression type problem,

0:04:06.914,0:04:13.099
and you would perform that significance [br]test using the test on beta, the slope,

0:04:13.099,0:04:17.847
using the regression simple linear[br]command in StatCrunch,

0:04:17.847,0:04:20.276
so again that's for two [br]quantitative variables,

0:04:20.276,0:04:23.514
and you're going to look for phrases[br]like "Determine if the variables

0:04:23.514,0:04:25.879
are dependent, or [br]associated, or independent"

0:04:26.919,0:04:30.745
Back over here on the qualitative [br]side, it's a little bit shorter

0:04:30.745,0:04:36.205
on the qualitative side, but starting out[br]the same way we did on the quantitative

0:04:36.205,0:04:39.775
side, we need to determine if we [br]have one or two populations.

0:04:39.775,0:04:43.957
Again, if you have two populations then[br]you're going to have information

0:04:43.957,0:04:47.757
for two samples. You're going to be [br]given the number of successes

0:04:47.757,0:04:51.224
and the total sample size [br]for two different samples.

0:04:51.874,0:04:55.654
If you do determine that [br]you have one population,

0:04:55.654,0:04:58.955
then you would do a [br]proportion stat one sample.

0:04:59.185,0:05:02.315
If you do determine that you [br]have two populations,

0:05:02.315,0:05:04.606
then again you're going [br]to do a proportion stat,

0:05:04.606,0:05:06.389
but you're going to do two samples.

0:05:06.389,0:05:11.654
Our last test for qualitative data,[br]is if we have two variables.

0:05:11.654,0:05:19.487
For examples, we might be looking at[br]gender and happiness.

0:05:19.487,0:05:25.020
That would be two qualitative variables.[br]For gender, you would be male or female.

0:05:25.020,0:05:30.936
For happiness, they might indicate very[br]happy, or happy, and not happy,

0:05:30.936,0:05:34.168
so those are word answers to those.[br]"How happy are you?"

0:05:34.168,0:05:37.834
Very happy, happy, not happy.

0:05:37.834,0:05:42.153
Those are word answers so those are[br]definitely qualitative variables,

0:05:42.153,0:05:44.577
and it was two variables. [br]We were collecting gender

0:05:44.577,0:05:48.466
and we were collecting level of happiness[br]so we have two qualitative variables.

0:05:48.466,0:05:52.688
In the problem you can also look to see[br]if it says the variables are dependent,

0:05:52.688,0:05:54.648
associated, or independent.

0:05:54.648,0:05:58.957
Also note, a little hint here, that when[br]you do have two qualitative variables,

0:05:58.957,0:06:02.863
it's typically the data will be [br]shown in a contingency table,

0:06:02.863,0:06:05.681
so all of these clues here [br]help you determine

0:06:05.681,0:06:08.915
that you're going to perform a [br]chi-square independence test.

0:06:10.325,0:06:13.763
So this is a little bit about how [br]to use this decision tree,

0:06:15.663,0:06:19.547
or significance test flow chart. [br]Sometimes I call it a decision tree

0:06:19.547,0:06:21.412
because it branches off,

0:06:21.412,0:06:24.594
or sometimes I just call it a [br]significance test flow chart.

0:06:24.594,0:06:28.861
But anyway, this is a little bit about[br]how to use this document.