Skip to main content
. 2020 May 26;11:884. doi: 10.3389/fpsyg.2020.00884

TABLE 6.

Trend estimation for mathematics in Germany.

Method 2003 2006 2009 2012 2015 Trend 2012 → 2015
Est SEtot SEp SEi
Original 503 504 513 514 506 −8 5.4 4.1 3.5
Without field test (all items) C1 503 513 515 522 510 −12 5.8 3.0 5.0
C2 503 515 517 524 511 −13 5.8 3.0 5.0
H1 503 512 514 521 505 −16 6.0 4.2 4.2
H2 503 517 521 528 515 −13 6.2 4.2 4.5
S1 503 512 514 518 503 −15 6.1 4.2 4.4
Without field test (invariant items) C1I 503 513 515 522 521 −1 5.8 3.0 5.0
C2I 503 515 517 524 522 −2 5.8 3.0 5.0
H1I 503 512 514 521 512 −9 6.4 4.2 4.8
H2I 503 516 521 528 524 −4 7.0 4.2 5.6
With field test C1F 503 512 515 516 518 2 5.8 3.0 5.0
H1F 503 512 514 514 515 1 7.5 4.2 6.2
S1F 503 512 514 518 517 −1 5.8 3.0 5.0

SEtot = standard error due to sampling of persons and items; SEp = standard error due to sampling of persons; SEi = standard error due to sampling of items (link error).