Topic 5

Topic 5 DQ 1

When hypothesis testing, when might you use a related sample versus an independent sample? Provide examples of both population to illustrate the differences.

Lindsey

Week 5 CAT

In your own words explain the process of estimation. Provide an example in your explanation.

Topic 5 DQ 2

Mary, a colleague of yours, wants to test the effectiveness of a new lesson plan on two different groups of students. She wants to perform a t-test, but she is not sure whether to use a paired or unpaired test. Base on the information provided, how would you advise your colleague? If she asked you to perform the calculation for her, how would you use SPSS to do so?

Mary

Re: Topic 5 DQ 1

In hypothesis testing, there are different samples that are observed. The one that we learned about last week is independent samples. Independent samples are when participants are unrelated and observed only once (Privitera, 2015). An example of an independent sample is observing two different groups while eating at a buffet; one group is instructed to eat slow, and the other is instructed to eat fast. The number of calories consumed is recorded. This is done to test whether or not the rate of eating has an influence on the amount consumed (Privitera, 2015).

On the other hand, there is related samples, which is samples that are not independent (Privitera, 2015). According to our book, “in a related sample, participants are related. Participants can be related in one of two ways: They are observed in more than one group (a repeated-measures design), or they are matched, experimentally or naturally, based on common characteristics or traits (a matched-pairs design).” (Privitera 2015, page 297). An example of a related sample is observing athlete’s performances in a sample both before and after training camp (Privitera, 2015).

 

Patricia

Re: Topic 5 DQ 1

The text explains that independent samples consist of different participants that are observed one time in each sample or group. Furthermore, these participants are selected from two or more populations or selected from a single population and randomly assigned to different groups (Privitera, 2015). Unlike independent samples, participants in related samples are related. These type of study is often used in testing medical conditions or medications, where the first reading is before the subject has started the medication and the second reading is taken after the subject has used the medication for a specified amount of time (Grand Canyon University). Since related samples are more practical it makes it easier to observe the behavior of the same participants before and after treatment which means in cases like these related samples are a better option.

Grand Canyon University (2015) PSY-380 Topic 5 Essential Reading: Introduction to Statistics

Privitera, G. (2015) Statistics for the behavioral sciences (2nd ed.). Los Angeles, CA: SAGE. ISBN-13: 9781483381169

Christy Person 

Re: Topic 5 DQ 1

The text defines related samples, also called dependent samples, as those that participants are observed in more than one treatment called a repeat-measure design or based on common factors called matched-pair design. Related samples can be used for pre-post design such as the effect medicine has on participants. Matched-pair design might be used for intelligence or personality testing being they can match like levels and characteristics Privitera, 2015). Independent samples are where participants are unrelated and observed only one time. A sample can be selected from one population and then divided or chosen from two different populations. These types of study can evaluated the difference between subjects such as men and women or use a control and experimental group in the comparison ( GCU, Topic 5, 2015)  . An example would be one group drinks water with dinner one drinks nothing with dinner compare the amount consumed.

Grand Canyon University (2015) PSY-380 Topic 5 Essential Reading: Introduction to Statistics

Privitera, G. (2015) Statistics for the behavioral sciences (2nd ed.). Los Angeles, CA: SAGE. ISBN-13: 9781483381169 

Mary

1 posts

Re: Topic 5 DQ 2

If I were to advise my colleague, Mary, I would recommend a paired test. This can be done by using the matched-pairs design. In this design, participants are chosen and the matched-experimentally or naturally-based on mutual characteristics or qualities. Our book says, “different, yet matched, participants are observed in each treatment, and scores from each matched pair of participants are compared.” (Privitera 2015, page 298). SPSS can be used to perform the calculation by entering the variables and the data. After this is done, you would click analyze and then compare means and paired-sample t-test, move the correct items over to the right box, click OK, and then run the command (Privitera, 2015).

 

Terry Falkenberg 

1 posts

Re: Topic 5 DQ 2

I would definitely advise Mary to use the paired test. When you use the information from each pair instead of each student you eliminate the source of error. It is important when doing T-tests that you eliminate as much error as you can to prove your hypothesis. When you eliminate the error between persons your total value of error is smaller; which will reduce the standard error. I would go into SPSS to the variable view tab and enter the column names; making sure I use whole numbers. Then I would enter the data under the data view tab, from there I would click on analyze and compare the means and paired-samples t-test. This will give me a box where I can move the data so it is side by side and then hit OK or the run command. (Privitera, 2015)

 Reference

Privitera, G. (2015) Statistics for the behavioral sciences (2nd ed.). Los Angeles, CA: SAGE. ISBN-13: 9781483381169

Patricia

1 posts

Re: Topic 5 DQ 2

Since Mary is testing the effectiveness of the new lesson plan with two different groups of students, I would advise her to use a paired test. Paired tests are normally used when data or collected values are related in some way , paired test are used to compare two means where subjects have been measured twice which is exactly what Mary is doing. The text explain how SPSS can be used to compute related- samples t-test using the Analyze, Compare Means, and Paired-Samples T Test options in the menu bar, the text goes on by telling is how these actions will display a dialog box that allows you to identify the groups and run the test (for more details, see Section 10.4) (Privitera, 2015).

Privitera, G. (2015) Statistics for the behavioral sciences (2nd ed.). Los Angeles, CA: SAGE. ISBN-13: 9781483381169