Q. What’s statistical reliability ? Suggest some instruments to establish it. 10 Marks.

Skip to content
## Blog

# Research Method

##
6 thoughts on “Research Method”

### Leave a Reply

Q. What’s statistical reliability ? Suggest some instruments to establish it. 10 Marks.

Copyright © 2017 Beautiful Mind

The research findings through all scientific means are observed under the central tendency of the findings. These central tendency can be mean, mode or median.

Statistical reliability is a confirmation to these findings tgat it’s not because of just chance factor but it is significant. It tests the data that by how much it is significantly reliable.

The psychological findings may be because of the effect of that particular day events, subject biases or any thing that may effect internal or external consistency.

Statistical reliability ensures that the test data is checked on several such issues and have provided it’s consistency throughout and hence it adds further trust on reliability of the finding.

There can be two type of statistics – Parametric and non parametric, where parametric statistics is tested using T-test and anova tests.

T-tests finds the significance between two means and it reflects the amount by which it can because if chance factor. Similarly Anova tests is find statistical reliability by providing level of significance between more than two means.

Hence these two can be used to establish statistical reliability.

Example- The mean for two samples, one for smokers with cancer and other for smokers without cancer finds some significant difference. We can apply T-test on these two means to realise the mathematical significance.

Similarly of another sample of non-smokers with cancer exists then we go for Anova test.

Similarly there are tools for non parametric statistics where the population is infinite and no normal curve can be reached.

Thus statistical reliability is good test to actual reliability proven by mathematics and better scientific tools.

Statistical reliability in essence means consistency of a measure. The scores obtained by a research are statistically reliable if they are reproducible and consistent.

Statistical reliability is proved when for a hypothesis the scores from various research methods and different subjects are correlated with statistical proof. It ensures that the test results are free from biases or any conformities.

Establishing statistical reliability:

1. Testing and retesting and correlating the scores can help prove reliability. For e.g.- to see if a drug produces desired reaction in mice, it is given again and again to mice (after well planned intervals) and the results are correlated.

2. Parametric statistic measures like t-Test and ANOVA. In t-Test 2 means are compared and if there is no significant difference then the data is reliable. Similarly, if there are more than 2 means, analysis of variance (ANOVA) is used. Both these tests are used when there is a finite population and the research produces a normal distribution curve.

3. In case of skewed distribution, non parametric measures can be used.

4. Item response theory can also be used to prove statistical reliability.

Reliability refers to the consistency or repeatability of anything that we are considering and statistical reliability refers to the consistency with which we obtain analyse data using mathematical techniques to obtain the desired results. For example, a students consistent high grades ascertain his performance. Health of a person is judged based on the data obtained on a consistent basis from the tests performed.

1. t – tests are used as a tool in establishing statistical reliability as this test helps us to assess to what extent is the variance that we have obtained between two means is due to chance factor.

2. ANOVA is another test that does analysis of variance which helps us in comparing more than two means simultaneously.

3. We can also use Non-parametric Statistical Tests when the distribution isn’t normal and the sample size is small.

4. Test – retest reliability helps us in finding the consistency of our results.

For example, if we are two compare the results obtained from two experimental groups, we can go for t-test. And if there is a control group along with the two experimental groups, we can use ANOVA. Repeatability of the results from the tests conducted ensures the reliability of any statistical data available.

Statistics offers us the probability of interlinking variables and deriving the relationship between them, when the links werenâ€™t certain. Statistical reliability is when we are able to establish the relationship as many times upon the repetition of experiment. Reliability ascertains us that the relation was correct and the hypothesis is certified.

For example, we test the effectiveness of a drug helpful in pain relief. The time taken to relieve from the pain is noted. If every repeated test offers equal or approximately same values, we can state that the test is reliable.

Instruments to establish reliability:

1. t-test and ANOVA test can be used to find correlation and mean results.

2. Repetition of the test can also help us establish reliable results.

Various other methods of correlation can be repeated to establish conformity of results and reliability can be assured.

Stability reliability refers to the replication of the same results number of times in order to ensure that the result obtained is valid I.e. the statistical reliability is about consistency of a result.

Eg. If we want statistical findings on the increase in air pollution in Delhi in different areas then we can find air quality index of those areas. Then by comparing the data ultimately increase in pollution can be find out in different areas.

Method to check statistical reliability is-

1. Parameteric method: This method assumes the shape of distribution in underlying population and about the form of the assumed distribution. This method includes following test:

1. t-test: A method of testing hypothesis about the mean of a small sample drawn from a normally distributed population when the population sample is unknown.

Eg. Consider the education of girls in Haryana nd UttarPradesh and to carry out research on this we can take sample of girls from both state and can carry out the t test. After calculating the mean for individual states data we can get the ultimate results.

2. Anova test: It is also a type of parametric method. Analysis of variance is a statistical method used to test difference between two or more means.

Eg. In a company if a manufacturer has two different method to manufacture a bulb then if he wants to find which one is better then he can go for Anova test of the two method and the method which gives the most reliable, quick and valid results can be brought into practise.

Statistical Reliability refers to the consistency the tests/ research produces which are not coincidental in nature and are repeatable to be relied upon for a greater validity.

Statistical tests can be further divided into 2 parts:

Parametric and Non Parametric.

Tests which refers to a finite population, when population is normally distributed and measurement is on interval scale, it is Parametric Stats. Reverse of it can be termed as Non-Parametric Stats.

Few Tests:

1.) T-test: When difference in the mean is significant and may not be because of chance factor.

2.) ANOVA: To compare more than 2 means simultaneously and analyse the variance.