Download Lecture 25

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Foundations of statistics wikipedia , lookup

Psychometrics wikipedia , lookup

Resampling (statistics) wikipedia , lookup

Omnibus test wikipedia , lookup

Misuse of statistics wikipedia , lookup

Student's t-test wikipedia , lookup

Transcript
Research Methodology
Lecture No :25
(Hypothesis Testing – Difference in Groups)
Recap
• Goodness of data is measured by reliability and
validity.
• Three measures of central tendency: mean, median
and mode.
• Dispersion is the variability.
• Three measures of dispersion are: range, variance
and standard deviation.
• Correlation
• SPSS Cronbach Alpha (Reliability) Factor Analysis
(Validity)
Hypotheses Testing
• Difference between groups
• Relationship between variables
Types of Hypotheses
Null
•that no statistically significant difference exists between
the groups
•No Statistically significant relationship exists between
variables
Alternative
•logical opposite of the null hypothesis
•that a statistically significant difference does exist
between groups
•That statistically significant relationship exists
Choose Appropriate Tests
• Based on the number of variables
i.e. two variables relationship (Univariate)and
many variables (Multivariate) statistical techniques.
• The type of scales Nominal, Ordinal(Non Metric) ,
Interval and Ratio(Metric) used choose appropriate
tests
• See page 338 of the text book.
Computer Outputs
• See the output results of the computer generated
outputs indicating the significance level.
Testing for Statistical Significance
•State the null hypothesis
•Choose the statistical test
•Select the desired level of significance
•Compute the calculated difference value
•Obtain the critical value
•Usually the software now provides the standard
significance values and the f or t values. Based on the
significance level value one can interpret the test
•Interpret the test
Selected Group Difference Cases
• Group difference
– Testing single mean
– Testing two related means (ratio)
– Testing two related samples when data is in
ordinal / nominal
– Testing two in unrelated means
– Testing when more than two groups on their
mean scores
Testing a hypothesis about a
single mean
• One sample t test
• Mean of the population from which a sample
is drawn is equal to comparison standard.
• i.e. we known that the in general the students
on an average study for 32 hours.
• Now you want to test that the students at VCIIT which are part of the student population
study less.
• So the sample of V-CIIT differ from the rest of the
population needs to be tested.
• Hypothesis generated would be
• Ho: The number of study hours of students V_CIIT is
equal to the number of hours studied in
general.(same)(no difference)
• Ha: The number of hours students of V_CIIT is less
then the number of hours studied in general (< )
•
•
•
•
SPSS
Analysis Compare means  One sample T Test.
Say you set the significance level to 0.05 then
See the output results of generated from the
software. See if the differences are significant or the
relationship significant. (lecture 6-7)
• If the differences are not significant then we accept
the null hypotheses other wise accept the alternate
• Out Put (T value and significance level)
Testing hypotheses about two related means
• Paired samples t-test
• Examine the difference in the same group before and
after the treatment
• Performance before training and after training
• Two observation each employee
• Null hypothesis
– There is no difference between the performance
of before and after the training
• SPSS
• use pair t test and see the value of t and it’s
significance level
• If the differences are not significant then we accept
the null hypotheses other wise accept the alternate
• Meaning the before and after training there was no
change i.e. Null hypothesis is accepted
– There is no difference between the performance
of before and after the training
Non Parametric Test for paired sampled
• When population cannot be assumed to be normally
be assumed distributed
• Use Wilcoxon singed –rank test ,
• Use McNemar’s test for non parametric and nominal
data
Testing about two unrelated means
• Group difference when groups are not related
and variable of interest data is in interval and
ratio scales.
• E.g: Groups MBA and Non MBA compared on
sales achieved.
• SPSS Analyze  Compare means
Independent samples T Test
• If more than two groups use ANOVA ( sales by
different level of education(Metric, FA,
BA/BS,Masters )
• SPSS excercises