+ All Categories
Home > Documents > 1 Let’s Have a Cup of Tea! Minjuan Wang ED 690 2 T test Video 21. Inference for One Mean the t...

1 Let’s Have a Cup of Tea! Minjuan Wang ED 690 2 T test Video 21. Inference for One Mean the t...

Date post: 16-Jan-2016
Category:
Upload: brenda-stephanie-carson
View: 216 times
Download: 0 times
Share this document with a friend
Popular Tags:
12
1 Let’s Have a Cup of Tea! Minjuan Wang ED 690
Transcript
Page 1: 1 Let’s Have a Cup of Tea! Minjuan Wang ED 690 2 T test Video  21. Inference for One Mean the t statistic for use when σ is not known. Emphasis is on.

1

Let’s Have a Cup of Tea!

Minjuan WangED 690

Page 2: 1 Let’s Have a Cup of Tea! Minjuan Wang ED 690 2 T test Video  21. Inference for One Mean the t statistic for use when σ is not known. Emphasis is on.

2

T test Video

21. Inference for One Mean the t statistic for use when σ is not

known. Emphasis is on paired samples and the t confidence test and interval..

22. Comparing Two Means How to recognize a two-sample

problem and how to distinguish such problems from one- and paired-sample situations are the subject of this program.

Page 3: 1 Let’s Have a Cup of Tea! Minjuan Wang ED 690 2 T test Video  21. Inference for One Mean the t statistic for use when σ is not known. Emphasis is on.

3

Paired Samples T Test (T for one) Inference for One Mean

http://learner.org/resources/series65.html?pop=yes&vodid=328482&pid=159#

Watch minutes: 12:30 to 20:58 Paired t test on neutralsweet Taste evaluation

One right after it is made One after 4 months H0= no differences in sweetness Ha: one-tailed test (there are differences) Before and after- sweetness values from each

panelists Calculate p (probability of chance)

Page 4: 1 Let’s Have a Cup of Tea! Minjuan Wang ED 690 2 T test Video  21. Inference for One Mean the t statistic for use when σ is not known. Emphasis is on.

4

Independent Samples (Tea for Two)

Comparing Two Means http://learner.org/resources/series65.html?pop=

yes&vodid=328482&pid=159# Do women make more money after

participating in the Options program in Baltimore? (a 3-year study)

Watch the beginning 10-11 minutes N=1398 The dataset will be too large for us to practice

in class. Pay attention to the t calculation on video

Page 5: 1 Let’s Have a Cup of Tea! Minjuan Wang ED 690 2 T test Video  21. Inference for One Mean the t statistic for use when σ is not known. Emphasis is on.

5

• T(ea) for One (flow chart p. 208 of Salkind)– T test for dependent samples (paired t test)– To test hypothesis– Examine changes between one group of participant

(or two matched pairs) on one or more variables– To compare if there is significant change in the

means of the variables studied

– Example: • (Not significant) attitude change of 21 youth before

and after their participation in Expeditions• Significant change in self-confidence and

competence• 10 patients going through a 2-hour psychotherapy

Page 6: 1 Let’s Have a Cup of Tea! Minjuan Wang ED 690 2 T test Video  21. Inference for One Mean the t statistic for use when σ is not known. Emphasis is on.

6

Types of T(ea) for One

• Repeated measures• Matched pairs

– Two groups matched on critical variables

• Example– E-classroom on fire

• Compare the occurring frequency of flaming and buffoonery in traditional and E classrooms

• Content analysis of discussions (counting)• Which test to use?

Page 7: 1 Let’s Have a Cup of Tea! Minjuan Wang ED 690 2 T test Video  21. Inference for One Mean the t statistic for use when σ is not known. Emphasis is on.

7

Interpretation (Results)• Level of significance

– Indicates how much the differences found are due to chance rather than intervention

– Usually set at 5% (a = 0.05)– Shown as Confidence interval = 95%

• Attitudinal gains (change)?– P > 0.05

• Learning Gains?– P <0.05

Page 8: 1 Let’s Have a Cup of Tea! Minjuan Wang ED 690 2 T test Video  21. Inference for One Mean the t statistic for use when σ is not known. Emphasis is on.

8

T(ea) for Two• Flow chart p. 194 of Salkind)

– T test for independent samples (unpaired t)– Test for a difference

• Examine differences between two groups of participants on one or more variables

– To compare if there is significant difference between the means of the variables studied

– Examples: • Mid-term scores of class A and class B• User engagement in two types of multimedia training

systems • The article on user engagement

Page 9: 1 Let’s Have a Cup of Tea! Minjuan Wang ED 690 2 T test Video  21. Inference for One Mean the t statistic for use when σ is not known. Emphasis is on.

9

Type of T(ea) for Two• Archetypal experiment

– Randomly selected and assigned to 2 groups

– Anti-depression drug versus placebo– T to compare mean differences on a

depression measurement scale • In Situ design

– Pre-assigned to 2 groups by nature or God

– Insomnia on work efficiency– Gender and …

Page 10: 1 Let’s Have a Cup of Tea! Minjuan Wang ED 690 2 T test Video  21. Inference for One Mean the t statistic for use when σ is not known. Emphasis is on.

10

• Is there significant difference in the intensity of eating disorder across different cultures– 297 Australian->249 Indian University

students

– Eating Attitudes Test and Goldfarb Fear of Fat Scale

– To measure Intensity of eating disorder– Run t test for independent samples (unpaired t-test)

– Any difference on the mean of eating attitude scores– Any difference on the mean of Fear of Fat scores

T(ea) for Two

Page 11: 1 Let’s Have a Cup of Tea! Minjuan Wang ED 690 2 T test Video  21. Inference for One Mean the t statistic for use when σ is not known. Emphasis is on.

11

• Results:

– Descriptive• Indian students scored higher on both of the

tests (higher intensity)

– Is the mean difference statistically significant?

• Take it to unpaired t test for independent samples

• T (eating attitude)= -4.19, p < .0001• T (fear of fat)= -7.64, P < 0.0001

Interpretation

Page 12: 1 Let’s Have a Cup of Tea! Minjuan Wang ED 690 2 T test Video  21. Inference for One Mean the t statistic for use when σ is not known. Emphasis is on.

12

– Judge by P value•The probability that the difference is due to chance

– P < 0.0001 (very small chance that the differences are due to things other than group membership)

– There are significant differences between Australian and Indian students in their intensity of eating disorder.

Conclusion


Recommended