Suuppuppose yo

Dear Dr. David,

Thanks for the reply but I did not get you. Please elaborate it.

Regards

A V V Subba Rao

The null hypothesis of the F statistic is adequacy of the model and alternative hypothesis is inadequacy of your model. The adequacy of your model means that the model is good.

Suppose you are fitting the following model Y = b0 + b1*X1 + b2*X2 + e

The null hypothesis in F statistic is equal to H0: b1 = b2 = 0

If this hypothesis is significant, then the model has a good fit to your data. If the hypothesis is not significant it means that considered linear model is not good for fitting to your data. Thus, you should change your model.

Dear Dr. Seyad,

Thank you very much for the reply, still I have some clarification on it and kindly address them. Thanks in advance.

1) What does it mean by H0: b1=b2=0

2) Why doesn't it be like H0: b0=b1=b2=0

3) When we can say the hypothesis is significant and when not?

Regards

Adhikari V V Subba Rao

H0 means there is no significant difference between variances of Independent variables means b1=b2=0 or b1-b2=0

In Multiple regression analysis, F-value tests the null hypothesis that

H0: The Proposed study model has no good fit

Ha: The proposed study model has a good fit

significant F-value (Sig. value less than 0.05, in case hypothesis is tested at 5% significance level) will lead to conclude that H0 is rejected, means proposed study model has a good fit.

This is not about fit! Ho: The researcher is rather stupid because non of his variables is statistically significant (b1=b2=b3=...=0), H1: At least one (!!!) variable is significant. So if you have one significant variable and five random, not connected with Y, the F test will show significance (Ho is rejected). So, do not use F test! Test regression coefficients individually, but mind the multiple testing problem.

In that case, what is use of t-value in coefficient table of the regression output

t-value is used to test Ho:Beta(j)=0 against H1=Beta(j) not equal 0 - just one coefficient (for one variable) at a time. In fact observed t-value should be used only to calculate p-value, which is later compared to significance level (0.05) in order to make a decision to reject or not to reject Ho.

Dear All,

Thank you very much for the valuable information. Kindly shed complete light on what is t value, p value and where to find them in the table.

Please elaborate with an example.

Regards

Rao

Please go through the attachment (page no. 2)and help me also to understand it better

Dear Dr. Gautam,

Thanks and I will.

Regards

Rao

p-value is an area under the density function of the probability distribution of test statistic ( Student t in this example) calculated from the empirical value of the test statistics toward the direction (>,

Dear Prof. Sokolowski,

Thank you very much for the clear explanation, I am expecting from you a numerical example to illustrate this concept of all about F value, t value, H0, Ha, significant and non significant. I would like to appreciate your valuable time for me enormously.

Regards

Rao

What is your field of application? Medicine, economics, psychology ,... any other?

Dear Prof. Sokolowski,

My field of application is Labour Studies. My topic for the M.Phil. dissertation is "Social Security in Indian Manufacturing Industry". I want to do research on it but unable to frame research problem, research objectives, research questions and H0, H1. Coming from mathematics and statistics background I am interested in QUANTITATIVE methodology. Your support in terms of your views/ideas/suggestions/inputs/frame work is of paramount importance to me. Thanks in advance,

Regards

Adhikari V V Subba Rao

8247328165

Suppose you have

A model of

CP=β0+ β1CO2 +β2PD, where

Cp= cereal production

Co2= carbon dioxide

PD=population density

Dependent Variable: CP

Method: Least Squares

Date: 02/04/18 Time: 06:10

Sample: 1971 2014

Included observations: 44

Variable

Coefficient

Std. Error

t-Statistic

Prob.

CO2

396.3660

86.40423

4.587345

0.0000

PD

-0.012641

0.013361

-0.946141

0.3496

C

126029.5

62061.00

2.030735

0.0488

R-squared

0.341237

Mean dependent var

335954.8

Adjusted R-squared

0.309102

S.D. dependent var

129070.8

S.E. of regression

107284.0

Akaike info criterion

26.07009

Sum squared resid

4.72E+11

Schwarz criterion

26.19174

Log likelihood

-570.5420

Hannan-Quinn criter.

26.11521

F-statistic

10.61893

Durbin-Watson stat

1.409384

Prob(F-statistic)

0.000192

The f –statistics tell us if all predictor variables are jointly explain to response variable or not

If p.value of F-statistics is less than 5% , then we say that all explanatory (predictor) variables can statistical significant relationship to response variable.

The F –test is global test vs which is individual test.

In our example ,the response variable is CP while co2 and PD are predictor variables . F-statistics (41,2)= 10.61893 , p.v

Dear Friend Liibaan Cali ,

It is of worth explanation you did and thanks a lot to you for your time to illustrate the concept with a very lucid manner. Keep on sharing such information related to regression, correlation and any type of statistical concept to me either on R.G. or on personal message.

Regards

Adhikari V V Subba Rao

8247328165

Liban Ali Mohamud

Thank you