SPSSX Discussion

change the 95% confidence interval for linear regression?

Classic

List

Threaded

5 messages Options

Joanne Tsai

change the 95% confidence interval for linear regression?

Hi, SPSS users;

I am wondering if there is any way I can change the 95% confidence interval for B to 90% or 85%?

This is the syntax I currently I have, if someone could please tell me how to modify it, I’d greatly appreciate it.

Thank you

*************************************************************************************************************

Sort cases by Prod.

Split file by prod.

REGRESSION

/MISSING LISTWISE

/STATISTICS COEFF OUTS CI R ANOVA

/CRITERIA=PIN(.05) POUT(.10)

/NOORIGIN

/DEPENDENT score

/METHOD=ENTER piece.

Coefficients(a)

PROD	Model		Unstandardized Coefficients		Standardized Coefficients	t	Sig.	95% Confidence Interval for B
PROD	Model		B	Std. Error	Beta	t	Sig.	Lower Bound	Upper Bound
A	1	(Constant)	50.252	1.580		31.809	.000	47.092	53.412
A	1	piece	5.292	3.934	.171	1.345	.184	-2.576	13.161
B	1	(Constant)	49.183	1.903		25.845	.000	45.376	52.989
B	1	piece	4.778	5.664	.108	.844	.402	-6.551	16.107

a Dependent Variable: score

Reutter, Alex

Re: change the 95% confidence interval for linear regression?

Prior to Statistics 17, you can’t change the confidence level. In v17, the command was updated to allow, for example:

/STATISTICS CI(90)

Cheers,

Alex

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Joanne Tsai
Sent: Friday, October 23, 2009 1:47 PM
To: [hidden email]
Subject: change the 95% confidence interval for linear regression?

Hi, SPSS users;

I am wondering if there is any way I can change the 95% confidence interval for B to 90% or 85%?

This is the syntax I currently I have, if someone could please tell me how to modify it, I’d greatly appreciate it.

Thank you

*************************************************************************************************************

Sort cases by Prod.

Split file by prod.

REGRESSION

/MISSING LISTWISE

/STATISTICS COEFF OUTS CI R ANOVA

/CRITERIA=PIN(.05) POUT(.10)

/NOORIGIN

/DEPENDENT score

/METHOD=ENTER piece.

Coefficients(a)

PROD	Model		Unstandardized Coefficients		Standardized Coefficients	t	Sig.	95% Confidence Interval for B
PROD	Model		B	Std. Error	Beta	t	Sig.	Lower Bound	Upper Bound
A	1	(Constant)	50.252	1.580		31.809	.000	47.092	53.412
A	1	piece	5.292	3.934	.171	1.345	.184	-2.576	13.161
B	1	(Constant)	49.183	1.903		25.845	.000	45.376	52.989
B	1	piece	4.778	5.664	.108	.844	.402	-6.551	16.107

a Dependent Variable: score

mpirritano

Re: change the 95% confidence interval for linear regression?

In reply to this post by Joanne Tsai

To calculate by hand is simple enough.

(critical value for t of .10 or .15 from t table based on your degrees of freedom) * (Std.Error.) plus and minus the unstandardized beta.

Matthew Pirritano, Ph.D.

Research Analyst IV

Medical Services Initiative (MSI)

Orange County Health Care Agency

(714) 568-5648

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Joanne Tsai
Sent: Friday, October 23, 2009 11:47 AM
To: [hidden email]
Subject: change the 95% confidence interval for linear regression?

Hi, SPSS users;

I am wondering if there is any way I can change the 95% confidence interval for B to 90% or 85%?

This is the syntax I currently I have, if someone could please tell me how to modify it, I’d greatly appreciate it.

Thank you

*************************************************************************************************************

Sort cases by Prod.

Split file by prod.

REGRESSION

/MISSING LISTWISE

/STATISTICS COEFF OUTS CI R ANOVA

/CRITERIA=PIN(.05) POUT(.10)

/NOORIGIN

/DEPENDENT score

/METHOD=ENTER piece.

Coefficients(a)

PROD	Model		Unstandardized Coefficients		Standardized Coefficients	t	Sig.	95% Confidence Interval for B
PROD	Model		B	Std. Error	Beta	t	Sig.	Lower Bound	Upper Bound
A	1	(Constant)	50.252	1.580		31.809	.000	47.092	53.412
A	1	piece	5.292	3.934	.171	1.345	.184	-2.576	13.161
B	1	(Constant)	49.183	1.903		25.845	.000	45.376	52.989
B	1	piece	4.778	5.664	.108	.844	.402	-6.551	16.107

a Dependent Variable: score

Art Kendall

Re: change the 95% confidence interval for linear regression?

In reply to this post by Joanne Tsai

Why would you want to change it?
In my experience few if any decision makers (including statisticians) can deal with comparing results at different confidence levels.
For cognitive psychological reasons but not necessarily mathematical reasons I often recommend that we stick with 95%. The same way that chemists usually stick to "at standard temperature and pressure".
In addition, in many environments people with just a smattering of exposure to stat see using unconventional confidence levels as a red flag.

Art Kendall
Social Research Consultants

Joanne Tsai wrote:

Hi, SPSS users;

I am wondering if there is any way I can change the 95% confidence interval for B to 90% or 85%?

This is the syntax I currently I have, if someone could please tell me how to modify it, I’d greatly appreciate it.

Thank you

*************************************************************************************************************

Sort cases by Prod.

Split file by prod.

REGRESSION

/MISSING LISTWISE

/STATISTICS COEFF OUTS CI R ANOVA

/CRITERIA=PIN(.05) POUT(.10)

/NOORIGIN

/DEPENDENT score

/METHOD=ENTER piece.

            Coefficients(a)






PROD

Model

Unstandardized Coefficients

Standardized Coefficients

t

Sig.

95% Confidence Interval for B

B
Std. Error

Beta
Lower Bound
Upper Bound

A

1

(Constant)

50.252

1.580

31.809

.000

47.092

53.412

piece
5.292

3.934

.171

1.345

.184

-2.576

13.161

B

1

(Constant)

49.183

1.903

25.845

.000

45.376

52.989

piece
4.778

5.664

.108

.844

.402

-6.551

16.107

a Dependent Variable: score

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

Art Kendall
Social Research Consultants

Richard Ristow

Re: change the 95% confidence interval for linear regression?

At 03:41 PM 10/23/2009, Art Kendall wrote:

>Why would you want to change [the confidence interval]? Few decision
>makers (including statisticians) can deal with comparing results at
>different confidence levels. For cognitive psychological reasons but
>not necessarily mathematical reasons I often recommend that we stick with 95%.

Granted. On the other hand, especially in these days of many tests, I
often use, and recommend, 99%. It's a useful rough-and-ready screen
for false precision in estimating the confidence interval.

As you write, "Few can deal with comparing results at different
confidence levels." Quite true. So I'd be sure to use 99% throughout
the study.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD