07 Independent Samples t-Tests in SPSS – SPSS for Beginners - YouTube

Welcome to the seventh video in SPSS for Beginners from RStats Institute at

Missouri State University. In the last video, we learned how to compare

means when the same sample was measured twice. In this test, we are going to

measure one group (males) and then measure a second group (females) and then see if

the mean for males is statistically significantly different than the mean

for females. Two groups measured one time. Because each of the groups are

independent of each other, we will use an independent samples t-test. Using the

same data set that we created in the first video, let's look at the variables.

We need two groups. The variable "Gender" will work nicely for these groups. We

have males and females. We need to measure the groups on something that can

vary. We measured height and weight, so we can answer a question like: "is there a

significant difference in height between males and females." So let's do it. Go to

Analyze -> Compare Means -> Independent Samples t-Test. the first thing that we

need is the variable that we measured. So, we measured both height and weight. So

let's get crazy and examine both variables at the same time. We will move

both height and weight into the variables box. And, we want to compare

between groups, so we move gender into the grouping variables box. But notice

that okay is still not available? And we have these question marks? What else do

we need to do? SPSS is telling us that it does not know what

our coding scheme is. We need to tell SPSS what are the two groups. Now, you may

ask: "WHY do we need to tell SPSS about the groups? We only have two groups."

Yes, WE have two groups, but there may be times that you are using a categorical

variable coded for multiple groups, like freshman, sophomore, junior, senior. But we

only want to compare freshmen to seniors. So, ee have to be able to specify which

of the groups should be compared. Click on Define Groups. For Group 1, we

will assign "1", which is males; and for Group 2, we will assign "2" for females.

Now we have our two groups. Click Continue and you can now see that OK

is available. So we are ready to run this test. Click OK.

Just as before, we get one table with descriptive statistics. It tells the

sample size, the mean, the standard deviation, and the standard error of the

mean for each of our groups: males and females. Below that, we see the table

containing the inferential statistics. We can mostly focus on this middle part of

the table. We have the t-score, the degrees of freedom for our two groups,

and the p-value that corresponds to that t score at those degrees of freedom,

which we use to determine if there is a significant difference, or not. So I want

to mention that for each t-test there are two rows, each with different results

for the test. The top row is for equal variances assumed, and the bottom row is

for equal variances NOT assumed. You will learn all about what that means later.

For now it's okay to just stick with the top row.

Remember that there are

three ways to determine statistical significance. First,

we can compare our t-value to a critical value that we look up in a table called

Students t Table. So I did that, and for 8 degrees of freedom, the critical

value is 2.306. This will be the same critical value for both

of these t tests. For height, the t-value is 2.214, which is

smaller than our critical value of 2.306. But, the t-value for

weight is 3.413, which is larger than our critical value

of 2.306. When the t-value

exceeds the critical value, then

the means are different. We can also look at the probability value to see if it is

smaller than .05. The p value for height is .058, slightly

larger than .05, but the .009 for weight is

smaller than .05. And we can look at the confidence interval to see

if it crosses zero. For height, the lower value is negative and the upper value is

positive, so it crosses zero. Not different. For weight, both values are

positive, so it does not cross zero.

So what do you think? Were there any

significant differences for height or weight?

Height was NOT different, but

weight was different. We can check the actual means in the Descriptive

Statistics, for more clarity. Males weighed 142.8 pounds and

females weighed 123.2 on the average. So males weighed significantly

more than females, but their Heights were not statistically significantly

different. You know, I've been talking about "statistical significance" as if it

tells the whole story, but that is not actually correct. There is something else

that we should calculate in addition to the

p-value and the confidence interval. We should calculate the effect size. I have

another video in which I go into a great deal of explanation about effect sizes

and why they matter. For now, just know that an effect size is exactly what it

says on the tin: it is the size of the effect the difference between the means.

Check out the RStats Effect Size Calculator for t Tests, for a spreadsheet

that you can use to compute effect sizes directly from SPSS output. When you're

ready to do an independent samples t-test for real, check out these other

videos from RStats Institute that will teach you more about statistical

theory, setting up your t-test, interpreting the results, and writing up

your findings in APA style. Thanks for watching these videos about SPSS for

Beginners from RStats Institute at Missouri State University. We have one

more short video to wrap up this series. If you've liked these videos so far, be

sure to subscribe and have all of our videos available for your viewing when

you take your statistics course.