how to find out the correlation and multiple regression by using MEGASTAT in Excel for the data in frequency format.

[lalit.dbe09]

Hello everyone,

please suggest me that how to find out the correlation and multiple regression by using MEGASTAT in Excel  for the data in frequency format..
i have no idea regarding  the type of data one need to perform correlation and regression. Do i need to tranform the data into Mean???
If yes, then please also suggest that how to convert the following data into Mean by using EXCEl....

i want to test the relationship between HRD practices and employee's development climate
by using Pearson correlation coefficient. Further,the independent
variables would  be tested in the second level analysis with multivariate
regression models to explain the process through which HRD practices
influences on employee’s development. The regression models will all
rigorously tested to avoid multi-co-linearity and auto correlation.


i have taken five HRD practices. Here, all these five HRD practices are independent variables and employee's development climate is dependent variable.....

Given below is the table having data in frequency of the four selected districts from Delhi and Haryana( two from each).

Table 2: Hypothesis Two
To examine the relationship between HRD practices and employees’ development climate.
recruitment		Rohtak	Jind	Central	West	Total
	Yes	77	63	114	116	370
	No	8	25	6	4	43
	Uncertain	35	32	0	0	67
	Total	120	120	120	120	480
training 1		Rohtak	Jind	Central	West	Total
	Yes	103	93	100	86	382
	No	10	18	6	20	54
	Uncertain	7	9	14	14	44
	Total	120	120	120	120	480
training 2		Rohtak	Jind	Central	West	Total
	Beneficial	94	72	66	5	237
	Very Good	23	31	18	0	72
	Good	0	14	18	40	72
	Average	3	3	12	74	92
	Below Average	0	0	6	0	6
	Very Poor	0	0	0	1	1
	Total	120	120	120	120	480
performance appraisal 1		Rohtak	Jind	Central	West	Total
	Yes	97	106	108	89	400
	No	20	10	6	17	53
	Uncertain	3	4	6	14	27
	Total	120	120	120	120	480
performance appraisal 2		Rohtak	Jind	Central	West	Total
	Extra..	37	35	28	15	115
	Changes…	3	7	29	60	99
	Support…	26	35	29	14	104
	Helps…	19	13	22	0	54
	Opportunities..	19	21	23	9	72
	All…	75	65	96	40	276
	Total	179	176	227	138	720
promotion 1		Rohtak	Jind	Central	West	Total
	Yes	83	80	97	66	326
	No	26	25	23	37	111
	Uncertain	11	15	0	17	43
	Total	120	120	120	120	480
promotion2		Rohtak	Jind	Central	West	Total
	Yes…	44	50	41	36	171
	Yes…	70	47	73	68	258
	No	6	23	6	16	51
	Total	120	120	120	120	480
Organizational development 1		Rohtak	Jind	Central	West	Total
	Yes	107	98	108	100	413
	No	4	16	8	16	44
	Uncertain	9	6	4	4	23
	Total	120	120	120	120	480
Organizational development 2		Rohtak	Jind	Central	West	Total
	Yes…	91	78	104	103	376
	No…	29	42	16	17	104
	Total	120	120	120	120	480

Thanks,
Lalit.

[Neeraj Kaushik]

You need not calculate mean as Chi sq test works on Frequency & not on means.

Plz help yourself by watching this video: https://www.youtube.com/watch?v=8j5H4NHd8Hw

Watch from timeline of 1:56 onwards

Also u need to click on Expected Frequency 

These values shd not be less than 5 for any cell. Check it manually

--
You received this message because you are subscribed to the Google Groups "DataAnalysis" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataanalysistrai...@googlegroups.com.
To post to this group, send email to dataanalys...@googlegroups.com.
Visit this group at http://groups.google.com/group/dataanalysistraining.
For more options, visit https://groups.google.com/d/optout.

[lalit.dbe09]

Dear Sir,

i dont want to use chi-square bcoz it does not suggest the relationship between different independent and dependent variables. Instead, i want to use correlation and regression for the data mentioned in the table above....So, please suggest me how to find out the correlation and regression for the hypothesis mentioned above.....





Thanks & Regards,
Lalit.

...

[Neeraj Kaushik]

On given data Correlation & Regression can't be applied.

Chi sq test is the best test here and remember Chi sq test is also a gud test to check for the association between the variables.

Actually choice of tests does no depend on us coz it depends on the way we've measured the data.

--