CoGAT: raw to scaled score?

[Anonymous]

Anonymous wrote:49/50 on Quant -> 150
September birthday.

September as in youngest or oldest?

[Anonymous]

I don't think the raw to scaled score is a simple conversion of "For this birth month, it's this many points off per question."

It's more like "What is the likelihood of a child with this birth month getting this many questions right?" It's stats, not a simple 160 - number wrong * x.

[Anonymous]

Anonymous wrote:I don't think the raw to scaled score is a simple conversion of "For this birth month, it's this many points off per question."

It's more like "What is the likelihood of a child with this birth month getting this many questions right?" It's stats, not a simple 160 - number wrong * x.

It is this. They have several different norming groups for each age bracket. If your kid is 7 years + 5 months, they'll be compared against the norming group of kids who are 7 years + 3 months through 7 years + 5.999 months, assuming that they use 4 norming groups for each year.

[Anonymous]

Anonymous wrote:

Anonymous wrote:I don't think the raw to scaled score is a simple conversion of "For this birth month, it's this many points off per question."

It's more like "What is the likelihood of a child with this birth month getting this many questions right?" It's stats, not a simple 160 - number wrong * x.

It is this. They have several different norming groups for each age bracket. If your kid is 7 years + 5 months, they'll be compared against the norming group of kids who are 7 years + 3 months through 7 years + 5.999 months, assuming that they use 4 norming groups for each year.

It's not a formula though - probably a chart like:

Number | Age | Age
Correct | 7.0-7.25 |7.25-7.5
50 | 150 | 146
49 | 145 | 140
48 | 142 | 139
etc...

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:I don't think the raw to scaled score is a simple conversion of "For this birth month, it's this many points off per question."

It's more like "What is the likelihood of a child with this birth month getting this many questions right?" It's stats, not a simple 160 - number wrong * x.

It is this. They have several different norming groups for each age bracket. If your kid is 7 years + 5 months, they'll be compared against the norming group of kids who are 7 years + 3 months through 7 years + 5.999 months, assuming that they use 4 norming groups for each year.

It's not a formula though - probably a chart like:

Number | Age | Age
Correct | 7.0-7.25 |7.25-7.5
50 | 150 | 146
49 | 145 | 140
48 | 142 | 139
etc...

Well yes. It's not a formula. It's a norming group, as I said in the quoted post. Like, if they are norming against 10000 kids to find the rarity of each score, they'll have 10000 kids of age 7 years 0 months - 7 years 2.99 months in the norming group.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:I don't think the raw to scaled score is a simple conversion of "For this birth month, it's this many points off per question."

It's more like "What is the likelihood of a child with this birth month getting this many questions right?" It's stats, not a simple 160 - number wrong * x.

It is this. They have several different norming groups for each age bracket. If your kid is 7 years + 5 months, they'll be compared against the norming group of kids who are 7 years + 3 months through 7 years + 5.999 months, assuming that they use 4 norming groups for each year.

It's not a formula though - probably a chart like:

Number | Age | Age
Correct | 7.0-7.25 |7.25-7.5
50 | 150 | 146
49 | 145 | 140
48 | 142 | 139
etc...

Well yes. It's not a formula. It's a norming group, as I said in the quoted post. Like, if they are norming against 10000 kids to find the rarity of each score, they'll have 10000 kids of age 7 years 0 months - 7 years 2.99 months in the norming group.

DP (poster of original quoted post here): some people in this thread are assuming it's a direct formula though. The point was that it's not. It's more complex math than that. You can't extrapolate it by getting a handful of DCUM posters to put birth month, raw score, and scaled score in a thread. And the norming depends on the year used for norms, too (found in the upper right of your report).

[Anonymous]

My understanding was they just compared kids scores with the same birth month.

[Anonymous]

Anonymous wrote:My understanding was they just compared kids scores with the same birth month.

Not the same month, but there are 3 month brackets. Kids are compared to kids scores within the same 3 month bracket.