Note on Control of Variances and Control of Extraneous Variance.

W3Schools
  • Note
  • Things to remember

A man observes that in days when he hears the cry of crows, he often get letters. So he concludes that crying of crows brings the letters.Is this conclusion based on the scientific method? Give your own view.

my opinion if a man gets letters he hears the cry of crows, he concludes that crying of a crow brings a letter.It is not a scientific method. But it is in the method of intuition because his natural inclinations and intuitive propositions and own personal experience and feeling make him conclude that.

To become a scientific method, it has to fill some characteristics of scientific methods related to basic criteria such as verifiability, generality, predictability, objectivity and system. But above condition does not match the text of such criteria of scientific criteria and therefore it is not a scientific method.

Analysis and synthesis.

After the information is collected from the two different source I,e primary as well as a secondary source they are processed and analysed to draw a proper inference.The first step in this process is to analysis is to prepare raw tablet.These tables are prepared by classifying the information into different groups as well as cells. This table required for transforming qualitative information to quantitative information called data. The data constancy is then tested and checked.

If data is found consistent, different statistical indicators are estimated from the data and presented in a table- the heading of which shows the analysed part of the information obtained from these indicators.

Control of Variance.

This is based on the principle of separating the variance due to different independent variables from the total variance of dependent variable I,e to split the total sum of a square into the different sum of squares, one for each independent variable and rest for errors.

The control of variance consisting of maximisations of a variance of concerned independent variables by controlling the variance of extraneous variable and maximising error variance.

Maximisation of Experimental variance.

The experimental variance I,e the variance of the dependent variable is maximised by the taking the level or types of independent variables as different as possible. If the independent level does not vary substantially, there is little chance of separating its effect from as the total variance of the dependent variable.

The maximum principle, thus, stresses to design, plan, and conduct research so that the experimental condition are as different possible.

Control of Extraneous Variance.

The control of the extraneous variance can be done by the following methods.

  1. Eliminate the extraneous variance from the experiment.
  2. Use randomization.Theoretically , randomization is the only method of controlling all possible extraneous variables.
  3. Use the extraneous variable as one of independent in the experiment.

Minimization of errors can be done by

  1. Reducing of errors by through controlled condition.
  2. Increasing the reliability of measures.

Sector of Research design.

[I] Sample design.

  1. defining the population and sample units.
  2. Fixing the desired errors.
  3. Choosing the stages of the sampling.
  4. Choosing the nature of sampling for each stage of sampling.
  5. Determination of overall sample size.
  6. Allocation of sample size at different stages of sampling and to different strata or segments of the population.

The sample design is mostly used in survey method.

[II] Design of an experiment.

  1. Defining the experimental units, treatments and their levels.
  2. Determination of replications to be used for each treatment.
  3. Choosing the strategies of allocating the treatment to experimental units.

Strategies of allocating the treatment to experiments units are called experimental designs. Some of the well-known designs are completely randomised block design. Randomised, Latin square design, Factorial designs (based on different levels of a treatment). Of these Factorial design controls most of the undesirable variations and gives the pinpointed elaborated results. The design of experiments is very mostly in experimental researches.

An example of Various Variable /Variances.

When intentions of the study is to investigate the cause and effect of the relationship between two variable X and Y, where Y is dependent variable (effect variable) and X=main independent variable (cause have some effects on the variable, Y, then the dependent of Y which might have some effects on the variable, Y, then the dependent of Y on independent variables is written as Y=X+K+Z+e, where Y is dependent variables, X is the principal independent variable, K, Z are extraneous variable I,e not of concern for the study and the error. The above Variance of X, the variance of K, Z etc and variance of E.

In above example the synthesis is the variance of Y and Variance of X, an Extraneous variance is the variance of K, Z etc and variance of error is the variance of E. However, our interest in research design is the variance of dependent variable which is greatly influenced by the variance of dependent variable X.

Example as

We know that a yield of a crop depends on upon types of fertiliser fertility of land under which it is grown, the quality of seeds of the crop, Water facilities, Labor used and many other things.

Then e we write, Y(yield)=Mean yield+yield due to Fertilizer +Yield due to fertility+ Yield due to the quality of the crop+ Yield due to watering + Yield due to labour+ Yield due to other factors. Here yield is the dependent variable. Quantity or quality of fertiliser, land labour, water etc are independent variables and other factors are the random or error variable.

Suppose now we are interested in studying the association between type of fertiliser and types of seeds in giving a different quantity of yield. Then fertiliser and seeds are the independent variables of interest and other variables such as water, labour and extraneous variable.

Reference.

Kerlinger, F.N. Foundation of Behavioural Research. New Delhi: Surjeet Publication, 2000.

Kothari, C.R. Research Methodology. India: Vishwa Prakashan, 1990.

Singh, M.L. and J.M Singh. Understanding Research Methodology. 1998.

Singh, Mrigendra Lal. Understanding Research Methodology. Nepal: National Book centre, 2013.

  1. To become a scientific method, it has to fill some characteristics of scientific methods related to basic criteria such as verifiability, generality, predictability, objectivity and system. 
  2. After the information is collected from the two different source I,e primary as well as a secondary source they are processed and analysed to draw a proper inference.
  3.  However, our interest in research design is the variance of dependent variable which is greatly influenced by the variance of dependent variable X.

.

Very Short Questions

0%

DISCUSSIONS ABOUT THIS NOTE

No discussion on this note yet. Be first to comment on this note