Calling all Stat hounds. . . - DPChallenge Forums

Canon EF-S 18-200mm f/3.5-5.6 IS

05/01/2008 10:18:11 PM · #1

I recently completed testing the 3rd, 6th, 9th, and 11th graders at a very small local private school. The test and software that I use gives me all the cool stats for each individual student showing me their standard scores, percentile rankings, etc.

What I am wondering is, do I have enough RAW data to be able to determine some "trends" to help the school pinpoint where their curriculum and instruction may be weak or strong so that they can identify the issues and correct them.

Of the very simplest computations, I can figure out the average scale score for each subtest, and can tell which subject averages highest and lowest, but, that comes out pretty close to average (100), which makes sense to me -- that's why it is called average.

What do I need to determine --

a. which subject is the strongest at this school?
b. which subject is the weakest?
c. if their students (the ones that are tested) are basically at grade level or above/below as a collective?

CEJ

Canon EOS-40D

05/01/2008 10:24:59 PM · #2

One data set is not enough to show you trends. You would need more than one set of scores from multiple years. That, IMO, would be the answer to all of the questions you asked. Even though you have 4 separate grades, you only have one year of data. To see trends either within or across those same grades you would need many years of consistent data.

Is there historical data from others that have used the same criteria to test in the past that you can access and incorporate into your data?

Message edited by author 2008-05-01 22:29:03.

SDW

Canon EOS-60D

Tamron AF 28-300mm f/3.5-6.3 XR Di LD Aspherical IF for Canon

05/01/2008 10:27:45 PM · #3

I agree with CEJ, you must have several test over time to get a trend. With that said you have enough data for a comparison between grade levels. If I'm reading your OP correctly I believe a comparison is what your looking for.

Tamron AF 28-300mm f/3.5-6.3 XR Di LD Aspherical IF for Canon

05/01/2008 10:34:38 PM · #4

yea, "trend" probably isn't the right word. or maybe it is, and like you said, I need a few more years worth of data.

Would it be possible to identify (other than the average) which areas are strong and which areas are weak?

SDW

Canon EOS-60D

05/01/2008 10:44:24 PM · #5

You should be able to do the following

1. Show how each grade level performs in each subject. You can even break it down by gender/subject.
Example. 3rd grade girls have a lower average in science than boys but the data show by 6th grade, girls are stronger in science than boys even though the average score is "x" points lower/higher than the 3rd grade average. Just one example.

dsterner

Canon EOS-40D

Tamron AF 18-200mm f/3.5-6.3 XR Di II for Canon

05/01/2008 11:33:35 PM · #6

are you comparing students to each other or to a standard for achievement?

GeneralE

Canon PowerShot S3 IS

Canon 12x IS USM (S3 IS built-in Lens) -- 36-432mm (35mm EQ)

05/01/2008 11:40:02 PM · #7

If the software can do it, I'd look for which individual questions were missed most often -- those should indicate either a poor question or an undertaught subject.

Secondly, I'd look for any sections where all of the grades scored consistently higher or lower than their overall average ... though this would really only have meaning if most of the kids have been attending for a while and went through the earlier grades there as well.

Canon RF 24-70mm f/2.8 L IS

05/02/2008 09:09:41 AM · #8

@dsterner -- standard achievement. It is the Woodcock Johnson Test of Achievement III

@GeneralE -- That is kinda what I'm wanting to do. Just glancing at the standard scores, there are definitely areas where all the students scored well. For example, in one test (Story Recall, measures short term memory and listening comprehension), 10 of the students scored in the "Average Range (90 - 110)," 2 scored in the "High Average Range (111-120)," and 2 in the "Very Superior Range (131+)."

On another test (Academic Knowledge - measure general knowledge across the fields of science, SS, and humanities. can be very "curriculum dependent" in some ways), 9 students scored in the average range, 3 in the low average, and 3 in the very low average.

To me, that would indicate that the first area is generally higher than the second. Is that a correct assumption/conclusion to make?

(I should know this. I have had stats classes, but it was several years and two babies ago. :) )

kirbic

Canon EOS R5

05/02/2008 09:58:28 AM · #9

Do you have access to comparative data? I'd compare the performance of your students to averages for like communities. You *do* have more than one data set. Since you have students at more than one level of study, you can track whether your students consistently score higher or lower than the comparison group(s) in certain areas. If, for instance, your first graders score high in reading skills, and that advantage is maintained and strengthened as they progress to 2nd and 3rd grade, then that is a very strong indicator that your reading program is performing well. Conversely, if your students fall behind in an area and do not catch up as they progress, that area almost certainly needs attention.

Canon EF 24-105mm f/4.0L IS

05/02/2008 11:16:31 AM · #10

This year, I do not -- this was my first year testing them. However, we've already scheduled next year's test, so I will next year.

DJWoodward

Canon EOS-5D Mark III

05/02/2008 12:12:45 PM · #11

I may have missed this already being mentioned with a quick read of the responses but another key factor is the sample size. You say it's a "very small" school. The power of any stats you create are very dependent on how many students are in your sample.

LoudDog

Canon EOS-350D Rebel XT

Canon EF 17-40mm f/4.0L USM

05/02/2008 12:30:28 PM · #12

You probably won't have any actionable data, but you should at least be able to pull out enough data to indentify areas to watch or maybe investigate.

Unless something stands out like a sore thumb (i.e. all grade levels have low scores in the same subject)