Scribd
Upload
87 views5 pages

Analysis of Covariance Techniques

Analysis of covariance is a statistical technique used to eliminate the effect of concomitant variables on the response variable in an experiment. It involves fitting a regression line of the response variable on the concomitant variables. The response variables are then adjusted using the regression line and treatments are compared using the adjusted values. It assumes the relationship between the covariate and response variable is linear and same across treatments, and the covariate is not affected by treatments. The technique is explained for a completely randomized design, with calculations shown for sums of squares, regression coefficient, adjusted means, and analysis of covariance table to test treatment effects.

Uploaded by

Aby Mathew
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
87 views5 pages

Analysis of Covariance Techniques

Analysis of covariance is a statistical technique used to eliminate the effect of concomitant variables on the response variable in an experiment. It involves fitting a regression line of the response variable on the concomitant variables. The response variables are then adjusted using the regression line and treatments are compared using the adjusted values. It assumes the relationship between the covariate and response variable is linear and same across treatments, and the covariate is not affected by treatments. The technique is explained for a completely randomized design, with calculations shown for sums of squares, regression coefficient, adjusted means, and analysis of covariance table to test treatment effects.

Uploaded by

Aby Mathew
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

Analysis of covariance

In many experiments for each experimental unit we have observation on one more
supplementary variable in addition to the response variable. They are usually concomitant
observations or concomitant variables. If the concomitant variables are unrelated to the
treatments and influence the response variable, the variation in response variable caused by them
should be eliminated before comparing the treatments. For example, in animal feeding
experiments designed to compare the effect of different diets on growth, the initial weight of
animal is expected to affect the increase in weight recorded at the end of the experimental period
and therefore comparison of diets should be made after the variation in weight increase resulting
from the difference in initial weights of animal has been eliminated.

The technique of analysis used to eliminate the variation resulting from the influence of
concomitant variable on the response variable is called Analysis of covariance. It is an adaptation
of methods of regression analysis to experimental designs and consists essentially of fitting a
regression of response variable on the concomitant variables. The response variables are then
‘adjusted’ by regression and comparison of treatments are carried out by using the adjusted
response variables. At the same time, variation due to non-uniformity of the experimental
material is eliminated by the use of suitable design.

Assumptions of Analysis of covariance


1. The relationship between covariate X and response variable Y is linear.
2. The relationship is same for each treatment.
3. The covariates are not affected by treatments.
4. The observations are from a normal population.

Analysis of covariance for CRD

For the CRD with t treatments, r i replication for ith treatment,i=1,2,3,…..t where =n , the
model is

where yij is the jth observation on the response variable taken under ith treatment,

- is the general effect, ti- ith treatment effect, is linear regression coefficient indicating the
dependency of yij on xij , xij is the observation on the concomitant variable corresponding to y ij,
-is the mean of the xij values and eij is the random error component which is independently and

normally distributed with mean zero variance .


The data arrangement is as given below
Treatments
1 2 …. t
y x y x y x
y11 x11 y21 x21 yt1 xt1
y12 x12 y22 x22 yt2 xt2
.
.
.
. Grand Total
Total Ty1 Tx1 Ty2 Tx2 Tyt Txt GTy GTx

The first step in the analysis of covariance is to compute the sum of squares for the variable (Y)
and the covariate (X) as well as the sum of products for Y and X. The sums of squares for both Y
and X are computed in the usual manner for a CRD. The sum of products (SP) of Y and X is
computed as follows:

Total SP,

Treatment SP = Tyx =

Error SP Eyx = Gyx-Tyx

Analysis on Y :

CFy=
Gyy=TSSy=

SST=Tyy=

SSE=Eyy=Gyy-Tyy

Analysis on X : CFx=

Gxx=TSS=

Txx=SSTx=

Exx = SSEx= TSSx-SSTx=Gxx-Txx

The next step is to verify whether covariate is affected by treatments. If X is not affected by
treatments there should not be significant difference between treatments with reference to X.

Then the regression coefficient within treatments is computed as . The significance of

is tested using the F-test. The test statistic F is given by

F=

=
The F follows a F distribution with 1 and (n-t-1) degrees of freedom. If the regression coefficient
is significant we proceed to make adjustments for the variate. If it is not significant, it is not
worth while to make the adjustments.

The adjusted values for the variable Y are then computed as follows:

One degree of freedom is lost in error due to fitting regression line.

Analysis of covariance table is given below:

Adjusted values for y F-ratio


Source df Sum of
Products df Sum of Mean sum of
YY XX XY squares squares
Total n-1 Gyy Gxx Gyx n-2 G1yy

Treatment t-1 Tyy Txx Tyx t-1 T1yy

Error n-t Eyy Exx Eyx n-t-1 E1yy

If F is not significant for treatments, we can conclude that the treatment effect do not differ
significantly. If F is significant, we calculate CD= t x SE(d) where t denote the table value of t
for a specified level of significance and error degrees of freedom.

Then the adjusted treatment means are obtained from the formula
and the standard error for the difference between two adjusted means is given by SE(d)=

.
When the number of replications is same for all treatments and when averaged over all values of

SE(d) = .

Two treatments are significantly different if the difference between adjusted treatment means is
greater than the calculated CD value, other wise they are not significantly different.

Assignment: Analysis of covariance for RBD and LSD

Common questions

Powered by AI

Adjusted treatment means in ANCOVA are calculated by adjusting the observed treatment means for the effects of a covariate, reflecting what the means would be if all experimental subjects had the same value of the covariate. This adjustment is essential as it isolates the treatment effects by removing variation due to the covariate, which can confound the comparison. This provides a clearer, unbiased estimation of the treatment effects, ensuring differences are attributed to treatments and not to initial covariate differences .

In a CRD, analysis of covariance adjusts for concomitant variables by first computing the sums of squares for the response variable (Y) and the concomitant variable (X), as well as the sum of products of Y and X. It then tests whether the covariate (X) is affected by treatments. Assuming X is not significantly affected by treatments, it proceeds to compute the regression coefficient within treatments. If this coefficient is significant, adjusted values of Y are computed by fitting a regression line. The adjusted values of Y are then used for comparing treatments .

Fitting a regression line in ANCOVA introduces complexities as it involves estimating additional parameters, specifically, the regression coefficients. This estimation impacts the statistical model by reducing the available degrees of freedom in the error term, as one degree of freedom is lost per regression coefficient fitted. This reduction in degrees of freedom can affect the precision of the estimation and the power of statistical tests, potentially complicating the process of detecting significant treatment effects .

The valid application of ANCOVA requires the following assumptions: 1) The relationship between the covariate X and the response variable Y should be linear. 2) This relationship should be the same for each treatment. 3) The covariates should not be affected by the treatments. 4) The observations should come from a normal population. These assumptions ensure that the adjustment of the response variable is accurate and that treatment comparisons are not biased .

In the context of ANCOVA analysis, the F-test is used to determine the significance of the regression coefficient between the response variable and the covariate. If the F-test is significant, it indicates that the covariate accounts for a significant portion of the variability in the response variable. This supports the need to adjust the response using the covariate. Additionally, if the F-test for treatment effects is significant after adjustment, it suggests that there is a significant difference between treatments .

Failing to meet the assumption of linearity in ANCOVA can lead to inaccurate adjustments of the response variable and, consequently, erroneous treatment comparisons. If the relationship between the covariate and the response variable is non-linear, the linear model used in ANCOVA will not accurately capture the effect of the covariate, possibly over- or under-adjusting the response. This mis-specification can inflate the error term or introduce bias, compromising the validity of the F-tests and potentially leading to incorrect conclusions about the significance of the treatment effects .

The purpose of using analysis of covariance (ANCOVA) in experimental designs is to eliminate the variation in the response variable caused by concomitant variables that are unrelated to the treatments but influence the response variable. This allows for a fair comparison of treatment effects by adjusting the response variable based on the linear relationship with the concomitant variable, like initial weight in animal feeding experiments. This method ensures that the comparison of treatments is not biased by the initial differences between subjects .

In ANCOVA, two treatments are determined to be significantly different if the difference between their adjusted treatment means is greater than the calculated critical difference (CD) value. The CD is computed as t multiplied by the standard error (SE) of the difference between two adjusted means. This is evaluated at a specific significance level with the error degrees of freedom. If the observed difference exceeds the CD, the treatments are significantly different; otherwise, they are not .

The significance of the regression coefficient in ANCOVA is tested using an F-test. If this coefficient is significant, it indicates that the covariate is contributing to explaining the variance in the response variable, justifying the need for adjustment of the response variable for valid treatment comparisons. If the coefficient is not significant, it suggests that adjusting the response variable for the covariate would not provide meaningful results, and thus, ANCOVA would not proceed to adjust Y .

It is crucial to check whether the covariate is affected by treatments in ANCOVA because if the covariate is significantly influenced by the treatments, the validity of adjusting the response variable using that covariate is compromised. The covariate should remain independent of the treatment effects; otherwise, adjusting for it might remove some of the treatment effects themselves, leading to inaccurate conclusions about the significance of treatment differences .

You might also like