SPSSX Discussion

Classes for stafive

Classic

List

Threaded

7 messages Options

Salbod

Classes for stafive

How would I create classes for stafive. Solutions for sten and stanine are
available online, but I cannot find a discussion of the algorithm. Can
someone point me in the direction of a book chapter or article that
discusses the creation of standard scores (stafive, stasix, staseve,
staeight, stanine and sten).

Thank you.

Stephen Salbod, Pace University, NYC

--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================
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

MLIves

Re: Classes for stafive

OK, I'll bite...and may get bitten back.

This is the first I heard of these measures, so forgive me if I'm way off base.

I found a table for Sten and it seems to me that following that logic, stafive would be a simple collapse of that (e.g., Sten 1+2 becomes Stafive 1).

Standard/Z scores, percentages, percentiles, and sten scores
Z-scores	< −2.0	−2.0 … −1.5	−1.5 … −1.0	−1.0 … −0.5	−0.5 … −0.0	+0.0 … +0.5	+0.5 … +1.0	+1.0 … +1.5	+1.5 … +2.0	> +2.0
Percent	2.28%	4.41%	9.18%	14.99%	19.15%	19.15%	14.99%	9.18%	4.41%	2.28%
Percentile	1.14	4.48	11.27	23.36	40.43	59.57	76.64	88.73	95.52	98.86
Sten	1	2	3	4	5	6	7	8	9	10

Stafive		1		2		3		4		5
Z-scores		< -1.5		-1.5 to -0.5		-0.5 to +0.5		+0.5 to +1.5		>+1.5
Percent		6.69%		24.17%		38.30%		24.17%		6.69%
Percentile		4.48		23.36		59.57		88.73		98.86

Then if the tails are static (are they?) other iterations would divide 97.76 (100-(2*1.14)) for the percentiles, 4 for the Z-scores, and 95.44 for the percentiles by the value of interest.

Am I missing something?

I am interested in this answer.

Melissa

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Salbod
Sent: Tuesday, November 21, 2017 8:30 AM
To: [hidden email]
Subject: [SPSSX-L] Classes for stafive

Thank you.

Stephen Salbod, Pace University, NYC

Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================

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

This correspondence contains proprietary information some or all of which may be legally privileged; it is for the intended recipient only. If you are not the intended recipient you must not use, disclose, distribute, copy, print, or rely on this correspondence and completely dispose of the correspondence immediately. Please notify the sender if you have received this email in error. NOTE: Messages to or from the State of Connecticut domain may be subject to the Freedom of Information statutes and regulations.

===================== 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

Salbod

Re: Classes for stafive

Thank you, Melissa. Your solution sounds right: sten à stafive, but I don’t know if it is a standard solution. --Steve

From: Ives, Melissa L [mailto:[hidden email]]
Sent: Wednesday, November 22, 2017 10:21 AM
To: Salbod, Mr. Stephen <[hidden email]>; [hidden email]
Subject: RE: [SPSSX-L] Classes for stafive

OK, I'll bite...and may get bitten back.

This is the first I heard of these measures, so forgive me if I'm way off base.

I found a table for Sten and it seems to me that following that logic, stafive would be a simple collapse of that (e.g., Sten 1+2 becomes Stafive 1).

Standard/Z scores, percentages, percentiles, and sten scores
Z-scores	< −2.0	−2.0 … −1.5	−1.5 … −1.0	−1.0 … −0.5	−0.5 … −0.0	+0.0 … +0.5	+0.5 … +1.0	+1.0 … +1.5	+1.5 … +2.0	> +2.0
Percent	2.28%	4.41%	9.18%	14.99%	19.15%	19.15%	14.99%	9.18%	4.41%	2.28%
Percentile	1.14	4.48	11.27	23.36	40.43	59.57	76.64	88.73	95.52	98.86
Sten	1	2	3	4	5	6	7	8	9	10

Stafive		1		2		3		4		5
Z-scores		< -1.5		-1.5 to -0.5		-0.5 to +0.5		+0.5 to +1.5		>+1.5
Percent		6.69%		24.17%		38.30%		24.17%		6.69%
Percentile		4.48		23.36		59.57		88.73		98.86

Then if the tails are static (are they?) other iterations would divide 97.76 (100-(2*1.14)) for the percentiles, 4 for the Z-scores, and 95.44 for the percentiles by the value of interest.

Am I missing something?

I am interested in this answer.

Melissa

-----Original Message-----
From: SPSSX(r) Discussion [[hidden email]] On Behalf Of Salbod
Sent: Tuesday, November 21, 2017 8:30 AM
To: [hidden email]
Subject: [SPSSX-L] Classes for stafive

Thank you.

Stephen Salbod, Pace University, NYC

Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================

===================== 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

Mike

Re: Classes for stafive

In reply to this post by MLIves

I did a little searching on the internet and in some journal databases

and though stanines are pretty common in psychometric textbooks

such as the following but the other terms do not:

Guilford, J. P., & Fruchter, B. (1973). Fundamentals statistics in

psychology and education (5e); see Chapter 19 Test Scales and

Scores.

(other editions should contain a chapter with similar info)

and

Anastasi, A. (1997). Psychological testing. (7e or earlier editions)

One source that I did find for the term Sten is the following:

Canfield, A. A. (1951). The "Sten" Scale-A Modified C-Scale.

Educational and Psychological Measurement, 11(2), 295-297.

http://journals.sagepub.com/doi/abs/10.1177/001316445101100213?journalCode=epma

One source that shows the relationship between stanines and stafives is:

Mink, O. G. (1964). Using Stanines to Study IQ-Achievement Relationships.

The School Counselor, 12(1), 43-45.

Mink's Figure 1 shows how the frequency distribution of raw scores

Expressed as stanines and on the stafive scale which consists of the

ordered categories Low, Below Average, Average, Above Average, and

High. I have copied Figure 1 and reproduce it below.

Figure 1. Percentage of Cases at Each Stanine Level for a Normal Distribution

Stanine Categories With Percentage of Values									Stafive
								9 (4%)	High (4%).
							8 (7%)		Above
						7 (12%)			Average (19%)
					6 (17%)
				5 (20%)					Average (54%)
			4 (17%)
		3 (12%)							Below
	2 (7%)								Average (19%)
1 (4%)									Low (4%)

It seems to me that all of these "st-terms" are ways of converting

a sample of raw score values into a simplified ordered categorical

scheme that assumes an underlying normal population distribution

of scores.

However, if the distribution of raw scores are skewed or

obviously non-normal in some other way, they have to be

"normalized" (if skewed, the skew is removed by matching the

percentile rank of score in the sample to that in the standard

normal distribution and converting the orignal score into one

that) before being converted to an st-term scale..

I believe that it is assumed that raw scores are based on

classic test theory of Raw Score = True Score + error.

IRT (not the subway ;-) models apparently can also be used

but I believe would have to use a 2 or 3 parameter model

for the analysis.

But if one can do an IRT analysis of the data, it is not clear what

converting scores to stanines or any other st-term categorization

really buys one outside of perhaps oversimplifying where an

individual raw score falls. It would seem to me that a raw

score's percentile rank would be enough but obviously there

is some other purpose or goal in mind. Which raises the

question why would anyone convert from raw score to one

of these "standardized" scores today?

-Mike Palij

New York University

[hidden email]

----- Original Message -----

From: [hidden email]

To: [hidden email]

Sent: Wednesday, November 22, 2017 10:21 AM

Subject: Re: Classes for stafive

OK, I'll bite...and may get bitten back.

This is the first I heard of these measures, so forgive me if I'm way off base.

I found a table for Sten and it seems to me that following that logic, stafive would be a simple collapse of that (e.g., Sten 1+2 becomes Stafive 1).

Standard/Z scores, percentages, percentiles, and sten scores

Z-scores

< −2.0

−2.0 … −1.5

−1.5 … −1.0

−1.0 … −0.5

−0.5 … −0.0

+0.0 … +0.5

+0.5 … +1.0

+1.0 … +1.5

+1.5 … +2.0

> +2.0

Percent

2.28%

4.41%

9.18%

14.99%

19.15%

19.15%

14.99%

9.18%

4.41%

2.28%

Percentile

1.14

4.48

11.27

23.36

40.43

59.57

76.64

88.73

95.52

98.86

Sten

1

2

3

4

5

6

7

8

9

10

Stafive

1

2

3

4

5

Z-scores

< -1.5

-1.5 to -0.5

-0.5 to +0.5

+0.5 to +1.5

>+1.5

Percent

6.69%

24.17%

38.30%

24.17%

6.69%

Percentile

4.48

23.36

59.57

88.73

98.86

Then if the tails are static (are they?) other iterations would divide 97.76 (100-(2*1.14)) for the percentiles, 4 for the Z-scores, and 95.44 for the percentiles by the value of interest.

Am I missing something?

I am interested in this answer.

Melissa

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Salbod
Sent: Tuesday, November 21, 2017 8:30 AM
To: [hidden email]
Subject: [SPSSX-L] Classes for stafive

How would I create classes for stafive. Solutions for sten and stanine are available online, but I cannot find a discussion of the algorithm. Can someone point me in the direction of a book chapter or article that discusses the creation of standard scores (stafive, stasix, staseve, staeight, stanine and sten).

Thank you.

Stephen Salbod, Pace University, NYC

--

Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================

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

This correspondence contains proprietary information some or all of which may be legally privileged; it is for the intended recipient only. If you are not the intended recipient you must not use, disclose, distribute, copy, print, or rely on this correspondence and completely dispose of the correspondence immediately. Please notify the sender if you have received this email in error. NOTE: Messages to or from the State of Connecticut domain may be subject to the Freedom of Information statutes and regulations.

===================== 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

Salbod

Re: Classes for stafive

Hi Mike,

Thank for the leads. You reminded, that I have Guildford’s early psychometric book. I’m going check out the Canfield and Mink articles.

Your last sentence, probably explains why I’m having difficulty nailing this down: “Which raises the question why would anyone convert from raw score to one of these "standardized" scores today?”.

Have a Happy Thanksgiving, Steve

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Mike Palij
Sent: Wednesday, November 22, 2017 2:40 PM
To: [hidden email]
Subject: Re: Classes for stafive

I did a little searching on the internet and in some journal databases

and though stanines are pretty common in psychometric textbooks

such as the following but the other terms do not:

Guilford, J. P., & Fruchter, B. (1973). Fundamentals statistics in

psychology and education (5e); see Chapter 19 Test Scales and

Scores.

(other editions should contain a chapter with similar info)

and

Anastasi, A. (1997). Psychological testing. (7e or earlier editions)

One source that I did find for the term Sten is the following:

Canfield, A. A. (1951). The "Sten" Scale-A Modified C-Scale.

Educational and Psychological Measurement, 11(2), 295-297.

http://journals.sagepub.com/doi/abs/10.1177/001316445101100213?journalCode=epma

One source that shows the relationship between stanines and stafives is:

Mink, O. G. (1964). Using Stanines to Study IQ-Achievement Relationships.

The School Counselor, 12(1), 43-45.

Mink's Figure 1 shows how the frequency distribution of raw scores

Expressed as stanines and on the stafive scale which consists of the

ordered categories Low, Below Average, Average, Above Average, and

High. I have copied Figure 1 and reproduce it below.

Figure 1. Percentage of Cases at Each Stanine Level for a Normal Distribution

Stanine Categories With Percentage of Values									Stafive
								9 (4%)	High (4%).
							8 (7%)		Above
						7 (12%)			Average (19%)
					6 (17%)
				5 (20%)					Average (54%)
			4 (17%)
		3 (12%)							Below
	2 (7%)								Average (19%)
1 (4%)									Low (4%)

It seems to me that all of these "st-terms" are ways of converting

a sample of raw score values into a simplified ordered categorical

scheme that assumes an underlying normal population distribution

of scores.

However, if the distribution of raw scores are skewed or

obviously non-normal in some other way, they have to be

"normalized" (if skewed, the skew is removed by matching the

percentile rank of score in the sample to that in the standard

normal distribution and converting the orignal score into one

that) before being converted to an st-term scale..

I believe that it is assumed that raw scores are based on

classic test theory of Raw Score = True Score + error.

IRT (not the subway ;-) models apparently can also be used

but I believe would have to use a 2 or 3 parameter model

for the analysis.

But if one can do an IRT analysis of the data, it is not clear what

converting scores to stanines or any other st-term categorization

really buys one outside of perhaps oversimplifying where an

individual raw score falls. It would seem to me that a raw

score's percentile rank would be enough but obviously there

is some other purpose or goal in mind. Which raises the

question why would anyone convert from raw score to one

of these "standardized" scores today?

-Mike Palij

New York University

[hidden email]

----- Original Message -----

From: [hidden email]

To: [hidden email]

Sent: Wednesday, November 22, 2017 10:21 AM

Subject: Re: Classes for stafive

OK, I'll bite...and may get bitten back.

This is the first I heard of these measures, so forgive me if I'm way off base.

I found a table for Sten and it seems to me that following that logic, stafive would be a simple collapse of that (e.g., Sten 1+2 becomes Stafive 1).

Standard/Z scores, percentages, percentiles, and sten scores

Z-scores

< −2.0

−2.0 … −1.5

−1.5 … −1.0

−1.0 … −0.5

−0.5 … −0.0

+0.0 … +0.5

+0.5 … +1.0

+1.0 … +1.5

+1.5 … +2.0

> +2.0

Percent

2.28%

4.41%

9.18%

14.99%

19.15%

19.15%

14.99%

9.18%

4.41%

2.28%

Percentile

1.14

4.48

11.27

23.36

40.43

59.57

76.64

88.73

95.52

98.86

Sten

1

2

3

4

5

6

7

8

9

10

Stafive

1

2

3

4

5

Z-scores

< -1.5

-1.5 to -0.5

-0.5 to +0.5

+0.5 to +1.5

>+1.5

Percent

6.69%

24.17%

38.30%

24.17%

6.69%

Percentile

4.48

23.36

59.57

88.73

98.86

Then if the tails are static (are they?) other iterations would divide 97.76 (100-(2*1.14)) for the percentiles, 4 for the Z-scores, and 95.44 for the percentiles by the value of interest.

Am I missing something?

I am interested in this answer.

Melissa

-----Original Message-----
From: SPSSX(r) Discussion [[hidden email]] On Behalf Of Salbod
Sent: Tuesday, November 21, 2017 8:30 AM
To: [hidden email]
Subject: [SPSSX-L] Classes for stafive

How would I create classes for stafive. Solutions for sten and stanine are available online, but I cannot find a discussion of the algorithm. Can someone point me in the direction of a book chapter or article that discusses the creation of standard scores (stafive, stasix, staseve, staeight, stanine and sten).

Thank you.

Stephen Salbod, Pace University, NYC

--

Sent from: http://spssx-discussion.1045642.n5.nabble.com/

=====================

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

This correspondence contains proprietary information some or all of which may be legally privileged; it is for the intended recipient only. If you are not the intended recipient you must not use, disclose, distribute, copy, print, or rely on this correspondence and completely dispose of the correspondence immediately. Please notify the sender if you have received this email in error. NOTE: Messages to or from the State of Connecticut domain may be subject to the Freedom of Information statutes and regulations.

===================== 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

Rich Ulrich

Re: Classes for stafive

In reply to this post by Salbod

Wiki says that stanines were used by the military in WW II. Stanines gave

a one-digit representation; and discouraged people from getting hung up

on trivial differences. Having huge samples, they used the percentiles to

define groups.

You can either cite the reference for what you use for any of the others, or

justify what you find most useful for the fractions left in the extremes. - Using

equal-normal-distances seemed to me to be the underlying standard for dividing

what lies between the extremes, but Mike Palij shows an example that converts

stanine to stafive by using (1, 2, 3, 2, 1) of the stanine groups.

If I did not use the percentiles, I would try to start with a composite score

that used transformations (before of after getting a raw composite) to

avoid bad skew.

Rich Ulrich

From: SPSSX(r) Discussion <[hidden email]> on behalf of Salbod <[hidden email]>
Sent: Tuesday, November 21, 2017 8:30:09 AM
To: [hidden email]
Subject: Classes for stafive

Salbod

Re: Classes for stafive

Thanks, Rich, for the mini history lesson: sounds like it was used with the Army General Classification Test. I’m going to look in the references that Mike Palij mentioned. -Steve

From: Rich Ulrich [mailto:[hidden email]]
Sent: Wednesday, November 22, 2017 6:34 PM
To: [hidden email]; Salbod, Mr. Stephen <[hidden email]>
Subject: Re: Classes for stafive

Wiki says that stanines were used by the military in WW II. Stanines gave

a one-digit representation; and discouraged people from getting hung up

on trivial differences. Having huge samples, they used the percentiles to

define groups.

You can either cite the reference for what you use for any of the others, or

justify what you find most useful for the fractions left in the extremes. - Using

equal-normal-distances seemed to me to be the underlying standard for dividing

what lies between the extremes, but Mike Palij shows an example that converts

stanine to stafive by using (1, 2, 3, 2, 1) of the stanine groups.

If I did not use the percentiles, I would try to start with a composite score

that used transformations (before of after getting a raw composite) to

avoid bad skew.

Rich Ulrich

From: SPSSX(r) Discussion <[hidden email]> on behalf of Salbod <[hidden email]>
Sent: Tuesday, November 21, 2017 8:30:09 AM
To: [hidden email]
Subject: Classes for stafive

Classes for stafive

Classes for stafive

Re: Classes for stafive

Re: Classes for stafive

Re: Classes for stafive

Stanine Categories With Percentage of Values

Stafive

Re: Classes for stafive

Stanine Categories With Percentage of Values

Stafive

Re: Classes for stafive

Re: Classes for stafive