| 
  • If you are citizen of an European Union member nation, you may not use this service unless you are at least 16 years old.

  • You already know Dokkio is an AI-powered assistant to organize & manage your digital files & messages. Very soon, Dokkio will support Outlook as well as One Drive. Check it out today!

View
 

Paired t-test

This version was saved 9 years, 3 months ago View current version     Page history
Saved by Alec Fefferman
on December 3, 2014 at 7:45:13 pm
 

A Paired T-Test (no need to capitalize) takes the difference of the means from two samples and produces a t-value from this information. So basically a t-test will measure the difference between two samples. Not quite correct. One has to make a distinction between an independent samples t-test and a paired samples t-test. Also, the test is not really "measuring" anything. A hypothesis test gives us a way to make decisions based on understanding whether an observation is statistically significant or could have arisen by chance.

 

A t-test is variation  type of a hypothesis test. The first step in this test would be to state the null and alternate hypothesis. As this is a t-test the equations will be different. Different from what?

null hypothesis:  .  The null is stated as the average of the means or μ. But what is μ measuring here?

alternative hypothesis     The alternate hypothesis is the average of the means of what we are trying to prove. The alternative hypothesis is a research question, a full sentence and not just a single number.

The next step would be to check the conditions as we would in a normal hypothesis test however there is one different condition instead of success/failure we check the nearly normal condition, otherwise the other two conditions are the same. You are confused about the terminology. The test for proportions we learned in class is just one of many hypothesis tests. It is not "the" quintessential hypothesis test. So it makes no sense to compare the paired t-test to it. Just state what is true for the paired t-test. The next step you would take would be to calculate your t-value if you are doing it by hand for that you would use this equation. Otherwise you would compute your t-value using SPSS, the steps for that are listed below.  So why not just say, "Compute the test statistic." At this point in the page, it doesn't matter that some people might use SPSS and some might do it by hand.

 

Depending on results from your test you would either choose to reject the null value or if there is insufficient evidence then you would not reject the null value.  Be more specific about what this means for a paired t-test.

 

Overall, you just need to follow the rubric.

 

  • Where is your fully worked out example? That is a critical part of the page!

 

The formatting of the section below is not correct. Please go back and read the directions and look at the sample page I've provided (Clustered bar graph).

 

 To conduct a Paired Sample T-Test in SPSS follow these simple steps:

 

  •  Go to "Analyze", "Compare Means", and select "Paired Samples T-Test". 
  •  Drag the variables you would like to compare into the "Paired Variables" window, Ideally you want to put your variable with larger numbers  under "Variable 1" and the smaller number as "Variable 2".
  • You can select "Options" in the top right corner if you want to change your "Confidence Interval Percentage". 
  • Click "Okay" and SPSS will give an you an Output of three tables, you will find the information you need in the third table.

 

Unable to display content. Adobe Flash is required.

 

 

You are forgetting your audience. This video will be publicly available for anyone in the world to see. So what good does it do to point people to some data files folder in a Canvas course that won't even exist in a few weeks?

 

I can see that you don't have a "blank slate" because when you open up the paired samples t-test window, the variables are already in place from the last time you tried it. This doesn't work for someone who is watching this video and following along. Open up a fresh copy of SPSS that doesn't have default selections already set from before.

 

Do you really want Jan and then July? It's not wrong, but then the difference in temperatures will all be negative, and that leads to a less intuitive conclusion. (You want to be able to say it's warmer in July, right?)

 

It's not really three different outputs. It's one output, even though the output is organized into three separate tables.

 

The means in the first box are misleading. We don't care about the means for the individual months. We only care about the differences. Also, you correctly mention that you can ignore the second box, but you take the time to name it and talk about it for a few seconds to say that. Just move past it more quickly.

 

You don't need to be as deliberate in pointing out that the degrees of freedom are labeled as "df" and the t-score is labeled as "t". 

 

By the way, you have not really conducted a paired samples t-test in the video. You have shown one step. You should at least make a passing mention of the fact that you have hypotheses (what are they here?) and that you need to check conditions. Ideally, since this sample size is nowhere near 40, you should compute the differences and get a histogram and QQ-plot. The steps in your video are the mechanics, so just finish the video with the actual conclusion.

 

Comments (0)

You don't have permission to comment on this page.