Group 2 Summary ANOVA

Chapter Summary ANOVA- Analysis of Variance
Group 2
C MANJUNATH(P16011), AMIT DHALIA(P16015), NITESH REDDY(P16031), M V VAMSI(P16032),
VIKRAM KHANNA(P16027), SANDEEP S KUMAR(P16029)
ANOVA is a statistical technique used as a test of means for two or more populations. The null
hypothesis, is that all means are equal. Analysis of variance must have one dependent variable
(metric) and one or more independent variable. Moreover, at least one of the independent
variables must be categorical, and the categorical variable must have more than one categories.
Categorical independent variables are also called factors. If the set of variables consists of both
categorical and metric variable, this technique is called analysis of covariance and metric
variable are called covariates, categorical variables still referred to as factors.
One-way ANOVA: It involves one categorical variable. Statistics associated with one-way
analysis of variance:
- eta2 (2): It measures strength of the effect of independent variable on dependent

variable.
2 = SSx / SSy
- F statistic: It is the ratio of mean square related to X and mean square related to
error. It is used to test the null hypothesis that the category means are equal in the
population.
Conducting One-way Analysis of Variance:
1. Identify the dependent and independent variable and their scale.

2. Decompose the total Variance: Decompose the total variance into a sum of square
corrected for the mean. So total variance in Y denoted by SSy, can be written as:
SSy = SSbetween + SSwithin
SSbetween is the variation in Y related to mean of the categories of X. It represents
variation between the categories of X. SSwithin is variation in Y related to the variation
within each category of X.
SSy = (Yi-)2
SSx =(j-)2
SSerror = (Yij-)2
3. Measure the effect by eta square statistic.
4. Test the significance: If the null hypothesis is true, then both SSx and SSerror comes from
the same source of variance. So estimation of the population variance of Y,
Sy2 = SSx /(c-1)
= mean square due to X
= MSx
Or,
Sy2 = SSerror /(n-c)
= mean square due to error
= MSerror
c = no. of categories
The Null hypothesis may be tested by the F-statistic based on the ratio between those two
estimates:
F = MSx / MSerror
F Statistic follows F-distribution with (c-1)*(n-1) degree of freedom.
5. Interpret the result based on a comparison between F statistic and critical value which
level of significance.
Assumptions in Analysis of Variance:
Categories of the independent variable are assumed to be constant which is referred to as a fixed-
effect model. Other models are also available like:
- Random-effect model: categories of independent variables are random sample from

universe of treatment
- Mixed effect model: some treatments are considered as fixed, and some are random.
1) Error term is normally distributed, with zero mean and constant variance.
2) Error terms are uncorrelated.
N-Way Analysis of Variance: It involves the simultaneous examination of two or more

categorical independent variables. It is same as one-way ANOVA with a major advantage that
interaction between the independent variables can be examined. The significance of the overall
effect, interaction terms, and main effects of individual factors are examined by appropriate F-
test. Interaction occurs when the effect of the independent variable on the dependent variable is
different for different category or level of another independent variable. If the interaction is
significant, it may be ordinal or disordinal.
In repeated measures analysis of variance, respondents are exposed to more than one treatment
condition and repeated measurements are obtained.
Nonmetric analysis of variance examines the difference in the central tendencies of more than
two groups when the dependent variable is measured on an ordinal scale.
Multivariate analysis of variance is similar to ANOVA, except that instead of one metric
independent variable, two or more variables can be used. MANOVA can examine group
difference across multiple dependent variables simultaneously. In MANOVA, the null
hypothesis is that the vectors of mean on multiple dependent variables are equal across groups.

Group 2 Summary ANOVA

Uploaded by

Copyright:

Available Formats

Group 2 Summary ANOVA

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Group 2 Summary ANOVA

Uploaded by

Copyright:

Available Formats

Chapter Summary ANOVA- Analysis of Variance

- eta2 (2): It measures strength of the effect of independent variable on dependent

Conducting One-way Analysis of Variance:

1. Identify the dependent and independent variable and their scale.

F Statistic follows F-distribution with (c-1)*(n-1) degree of freedom.

Assumptions in Analysis of Variance:

- Random-effect model: categories of independent variables are random sample from

N-Way Analysis of Variance: It involves the simultaneous examination of two or more

You might also like