<p>1930/2050
March/June</p>
<p>not enough data yet to start crunching, please keep posting.</p>
<p>1860/2010</p>
<p>...</p>
<p>After seeing this stuff, Im definitely retaking</p>
<p>2150/2230
March/June</p>
<p>Composite score 2270</p>
<p>In March, my CR score was a 670 because I do not have a good enough vocabulary for the sentence completions, and I'm not a perfect reader. I worked on vocab and practice tests. I managed to bring my CR score up 120 points to a 790! So, I am really happy and proud of myself!</p>
<p>42 data pairs so far......any more before I tally them up?</p>
<p>D's: 2180/2270</p>
<p>March- 1860 590 cr 650 m 620 w
May- 2010 680 cr 700 m 630 w</p>
<p>March:2070
May:2140</p>
<p>March 1820
June 1820</p>
<p>Good Lord! :(</p>
<p>March: 2110 (610CR, 700M, 800W)
June: 2170 (730CR, 680M, 760W)</p>
<p>Meh...</p>
<p>3 more makes 50, a nice round number</p>
<p>ps.....I'll be doing a simpler analysis of what I did for PSAT-SAT correlation on this thread:</p>
<p>1830 / 1930</p>
<p>2140/2200</p>
<p>first: m 670, cr 690, w 780
second: m 670, cr 740, w 790 haha the only thing i didn't go up in is the one thing that i wanted to go up in (math) oh well, i did awesome on my acts so who cares? lol</p>
<p>2000/2210</p>
<p>March/June</p>
<p>First Test: 760 M/650 V/ 590 W</p>
<p>I spent time studying for the writing exam in between the first and the second. (Writing multiple choice, not essay)</p>
<p>Second Test: 800 M/ 730 V/ 680 W</p>
<p>DATA PAIRS SORTED BY 1ST-TRY SAT I</p>
<p>1st SAT / 2nd SAT / DIFFERENCE
1770 / 1870 / 100
1780 / 2000 / 220
1820 / 1820 / 0
1830 / 1930 / 100
1840 / 2040 / 200
1860 / 1970 / 110
1860 / 2010 / 150
1860 / 2010 / 150
1910 / 2090 / 180
1920 / 2000 / 80
1930 / 2050 / 120
1960 / 2150 / 190
1970 / 2010 / 40
1990 / 2030 / 40
1990 / 2160 / 170
1990 / 2180 / 190
2000 / 2110 / 110
2000 / 2210 / 210
2040 / 2000 / -40
2050 / 2270 / 220
2070 / 2140 / 70
2070 / 2200 / 130
2070 / 2260 / 190
2090 / 2260 / 170
2100 / 2160 / 60
2110 / 2170 / 60
2110 / 2170 / 60
2110 / 2180 / 70
2120 / 2160 / 40
2130 / 2220 / 90
2130 / 2230 / 100
2140 / 2200 / 60
2140 / 2220 / 80
2150 / 2230 / 80
2150 / 2240 / 90
2160 / 2190 / 30
2160 / 2190 / 30
2170 / 2220 / 50
2170 / 2400 / 230
2180 / 2270 / 90
2180 / 2290 / 110
2190 / 2310 / 120
2190 / 2310 / 120
2200 / 2280 / 80
2200 / 2310 / 110
2200 / 2330 / 130
2200 / 2340 / 140
2210 / 2220 / 10
2230 / 2290 / 60
2250 / 2160 / -90
2300 / 2280 / -20</p>
<p>MATCHED PAIR ANALYSIS</p>
<p>Matched Pairs
Difference: 2nd SAT I-1st SAT I</p>
<p>2nd SAT I 2163.53<br>
1st SAT I 2063.73<br>
Mean Difference 99.8039<br>
Std Error 9.72762<br>
Upper95% 119.342<br>
Lower95% 80.2654<br>
N 51<br>
Correlation 0.8629 </p>
<p>The DIFFERENCE mean is ~100 SAT points, with upper and lower 95% confidence intervals at 119 & 80, respectively. Because this confidence interval does NOT span 0, the significance between the two groups (1st try vs 2nd try) is real.</p>
<p>DISTRIBUTIONS</p>
<p>1st SAT I
Quantiles</p>
<p>100.0% maximum 2300.0
99.5% 2300.0
97.5% 2285.0
90.0% 2200.0
75.0% quartile 2170.0
50.0% median 2110.0
25.0% quartile 1970.0
10.0% 1844.0
2.5% 1773.0
0.5% 1770.0
0.0% minimum 1770.0</p>
<p>Moments</p>
<p>Mean 2063.7255
Std Dev 135.88172
Std Err Mean 19.027247
upper 95% Mean 2101.9428
lower 95% Mean 2025.5081
N 51</p>
<p>2nd SAT I
Quantiles</p>
<p>100.0% maximum 2400.0
99.5% 2400.0
97.5% 2382.0
90.0% 2310.0
75.0% quartile 2260.0
50.0% median 2190.0
25.0% quartile 2050.0
10.0% 2000.0
2.5% 1835.0
0.5% 1820.0
0.0% minimum 1820.0</p>
<p>Moments</p>
<p>Mean 2163.5294
Std Dev 127.74699
Std Err Mean 17.888157
upper 95% Mean 2199.4588
lower 95% Mean 2127.6
N 51</p>
<p>DIFFERENCE
Quantiles</p>
<p>100.0% maximum 230.0
99.5% 230.0
97.5% 227.0
90.0% 198.0
75.0% quartile 150.0
50.0% median 100.0
25.0% quartile 60.0
10.0% 14.0
2.5% -75.0
0.5% -90.0
0.0% minimum -90.0</p>
<p>Moments</p>
<p>Mean 99.803922
Std Dev 69.469135
Std Err Mean 9.7276247
upper 95% Mean 119.34243
lower 95% Mean 80.265412
N 51</p>
<p>In the group of 51, 3 data pairs showed a loss, one that held steady, and the remaining 47 showed improvement. Interesting to note also that 2 of the 3 losing data pairs were the 2 highest initial scoring SAT Is.</p>
<p>DISCUSSION</p>
<p>47 out of 51 reported data pairs showed improvement, or 92%. Average improvement was 100 points. I find these as startling statistics. Assuming the tests were equivalent in difficulty, these results support the adage that test-takers learn quite a bit from one test to the next. This compares to about 75% improving from PSAT to SAT, with an 89 point average improvement & much greater scattering of the DIFFERENCE data as represented by standard deviation .69 for SAT>>SAT DIFFERENCE versus 132 for PSAT>>SAT DIFFERENCE. (Reference for PSAT>>SAT data: <a href="http://talk.collegeconfidential.com/showthread.php?t=52299&page=12%5B/url%5D">http://talk.collegeconfidential.com/showthread.php?t=52299&page=12</a>)</p>
<p>The fact that 2 out of the 3 declines in 2nd try SAT I scores occurred with the highest scoring 1st try SAT Is also shows that it is easier to lose when starting high, much like the DIFFERENCE pattern with PSAT>>SAT scoring. Although, there does not appear to be drastically better 1st try SAT I to 2nd try DIFFERENCE improvements at the lower end of the reported scores.</p>
<p>im guessing you were at the top of your statistics class =P</p>
<p>I wonder if taking it a third time will bring the same changes.. if it does than it is definately worth it to take it again</p>
<p>51 data sets doesn't exactly carry much external validity...the collegeboard usually employs a few thousand data sets for smaller studies, and often uses a few hundred thousand students when they modify the test. </p>
<p>You can read some of the studies done here: <a href="http://www.collegeboard.com/research/home/%5B/url%5D">http://www.collegeboard.com/research/home/</a></p>
<p>Among the more interesting studies completed were those of fatigue effects, the harder math, and the new writing section.</p>
<p>The collegeboard has this to say about retakes:</p>
<p>Overall, 55% of juniors taking the test improved their scores as seniors, 35% had their scores decrease, and 10% had no change.
On average, juniors repeating the SAT as seniors improved their verbal scores by about 12 points and their math scores by about 13 points. About 1 in 25 gained 100 or more points on verbal or math, and about 1 in 90 lost 100 or more points</p>
<p>points well taken, eeepikk. The survey above is not only tiny relative to what the College Board can do, it is also NOT a random sample, and the self selection effects are probably substantial. Nevertheless, I still find it interesting how many test takers improved in this group of 51...a much higher percentage than those reporting improvements on the PSAT thread. I don't know what to make of this, & assuming the self-selection bias is similar between the 2 threads, possible explanations include either (1) the training provided by going from SAT to SAT is better than PSAT to SAT, and/or (2) that the March SAT was more harshly graded than subsequent tests. I'd suspect the former. Fun to speculate, but we'll never know from this meager, self-selected sample. The most important conclusions, I believe, are that (1) substantail improvement is most definitely possible, & (2) don't get cocky, because 2nd-try scores can on the other hand worsen.</p>