Hypothesis Testing and Results Interpretation

1 / 10

# Hypothesis Testing and Results Interpretation - PowerPoint PPT Presentation

Hypothesis Testing and Results Interpretation. By Minjuan Wang ED 690 Educational Technology. Types of Hypothesis. Null hypothesis There is no change, difference or relationship between A and B A starting point or a benchmark

I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.

## PowerPoint Slideshow about 'Hypothesis Testing and Results Interpretation' - oya

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.

- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript

### Hypothesis Testing and Results Interpretation

By Minjuan Wang

ED 690

Educational Technology

Types of Hypothesis
• Null hypothesis
• There is no change, difference or relationship between A and B
• A starting point or a benchmark
• Many articles have implied null hypothesis, but may not clearly stated.
• Few studies are designed to verify the nonexistence of a relationship
• Coffee maker is broken
• no relationship between its broken and the humming birds on the tree outside my balcony.
Alternative/Research Hypothesis
• Hypothesis that is implicitly accepted if the null hypothesis is rejected
• Directional (one-tailed test)
• Non-directional (two-tailed test)
• Hypothesis: not to be proven but to be supported
• Does your study fail if your hypothesis is not supported by the data?
Hypothesis testing
• Core of Inferential Statistics
• contributes to the science of education primarily by expanding, refining, or revising its knowledge base.
Hypothesis Testing Procedure
• Come up with a hypothesis
• Set a: level of risk you are willing to take; or the cut-off point for a test result to be significant
• Select the test
• Compute the obtained value: t, f, r, etc.
• Find the critical value in the respective table
• To reject a null hypothesis->Obtained value must be > critical value
• Otherwise, fail to reject null hypothesis (H0)
• But, never “accept” null hypothesis
• Many tests are needed to confirm that A is not different from or associated with B.
• When rejecting null-P, the alternative 2-tailed P is implicitly accepted.
P from Fancy Schmancy Software
• Inferential statistics
• T, ANOVA, Correlation, Regression
• P is the probability of chance (indicator of significance)
• Free us from the test tables
• Results vary
• P<.05
• P<.001
• P=.013 (the exact probability of the outcome/effect due to chance—SPSS)
• Outcome: difference, change, or association
• P>.05 or p=ns (nonsignificant)
• The probability of rejecting a null-P exceeds 0.05 (the cut-off point) (Salkind)
• So reject it
Examples & Exercise
• Scenario:
• two groups of patients: anti-depression drug group; and placebo group
• Run t for two (null-HP)
• t(58)=2.45, p<.05
• t: the test that was used
• 58: degree of freedom
• 2.45: the obtained value
• P: the probability of chance is within the cut-off point (level of significance/or risks allowed)
• Significant mood difference exists between the two groups
• So the treatment (drug) did work
Examples & Exercise
• Scenario:
• two groups of patients: anti-depression drug group; and placebo group
• Run t for two
• t(58)=0.14, p>.05
• t: the test that was used
• 58: degree of freedom
• 0.14: the obtained value
• The probability of rejecting the null-p exceeds the cut-off point, so reject it!
• Also means, the probability of chance exceeds the cut-off point (level of significance/or risks allowed)
• No significant mood difference exists between the two groups
• So the treatment (drug) did not work
Examples & Exercise
• Scenario:
• two groups of patients: anti-depression drug group; and placebo group
• Run t for two
• t(58)=0.14, p=.891
• t: the test that was used
• 58: degree of freedom
• 0.14: the obtained value < critical value 2.001
• Not extreme enough for us to conclude the difference is due to treatment
• P: the exact probability that the outcome (difference) is due to chance
• Also means, the probability of chance exceeds the cut-off point (level of significance/or risks allowed)
• No significant mood difference exists between the two groups
• So the treatment (drug) did not work
Shaprio-W Normality Test
• Seems to have a different interpretation
• But only used in Analyse-it
• The higher the p, the more normal the distribution is.