crim 483 chapter 9: testing differences between the means of different groups
DESCRIPTION
The T-Test Statistic The t-test can be used to examine the following: –The mean difference between two groups –The mean difference within related groups (pre/post-test) This class only focuses on the first type of test –See formula on p. 163 Deciding to use the t-test: –Must be examining the difference between groups on one factor –The same participants are not being tested more than once –There are only two groups (Note: A t-test cannot be used to test the mean differences between more than two groups)TRANSCRIPT
CRIM 483
Chapter 9: Testing differences between the means of different
groups
Putting It Into Perspective
• Two essential components in using statistics appropriate include:– Having a clear understanding of what you are
testing– Having a clear understanding of which
statistical test is appropriate• T-test statistic is method that expands the
types of questions you can examine
The T-Test Statistic• The t-test can be used to examine the following:
– The mean difference between two groups– The mean difference within related groups (pre/post-test)
• This class only focuses on the first type of test– See formula on p. 163
• Deciding to use the t-test:– Must be examining the difference between groups on one factor– The same participants are not being tested more than once– There are only two groups (Note: A t-test cannot be used to test
the mean differences between more than two groups)
When to Use--Examples• A t-test of this type is appropriate to use in the
following examples:– To determine whether perceptions of equality differ
across males and females– To determine age differs between property offenders
and violent offenders• In each of the examples above, only one point in
time is considered (i.e., data were not collected at multiple points in time)
• Do not use a t-test for nominal variables
Book Example
1. State the null and research hypothesisNull: The teaching methods for Alzheimer patients will
not result in different levels of memory recall.Research: Alzheimer patients who received visual aid
teaching methods will differ in their ability to remember the order of daily tasks from the patients who received intense verbal rehearsal.
Example, Continued
2. Set the level of risk: .05—There is only a 5% chance that the relationship is due to chance alone.
3. Select the appropriate statistic—Since the hypotheses are testing whether the level of memory recall differed across two different teaching styles, a t-test would be appropriate.
Example, Continued4. Compute the value—Using the data on p. 164, the
resulting t-value is -.145. Identify the critical value—The point at which the
probability is .05 or less• Determine the degrees of freedom ((n1+n2)-
2))=df; 30+30-2=58• Using the t-test table, determine that the critical t-
value is 2.0016. Compare obtained value to critical value
• -.14 is less than 2.0017. Make a decision…does memory recall differ by
teaching method?
Effect Size• Significance v. meaningfulness• An effect may be significant but not meaningful—to
assess meaningfulness, you can calculate effect size: – ES=(X1-X2)/SD (NOTE: Assumes the standard deviations
between two groups is equal• Evaluating effect size
– 0=very similar and overlap entirely– Small effect=0-.20– Medium effect=.20-.50– Large effect=.50 and above – 1.00=overlap between variables=45%
• In the current example, the effect size is .37 ((7.46-6.90)/1.53) or medium
Book example using SPSS: Chapter 9 Data Set 1
Homework
• Page 176: 1-3
• Chapter 9 Data Sets 2 & 3 are posted on the class website