Appendix A3. Table A3.1. General linear model results for RMSE under the unconditional model. Source DF SS Mean Square

Σχετικά έγγραφα
Supplementary Appendix

Statistics 104: Quantitative Methods for Economics Formula and Theorem Review

Biostatistics for Health Sciences Review Sheet

Λογαριθμικά Γραμμικά Μοντέλα Poisson Παλινδρόμηση Παράδειγμα στο SPSS

Πτυχιακή Εργασι α «Εκτι μήσή τής ποιο τήτας εικο νων με τήν χρή σή τεχνήτων νευρωνικων δικτυ ων»

FORMULAS FOR STATISTICS 1

Αν οι προϋποθέσεις αυτές δεν ισχύουν, τότε ανατρέχουµε σε µη παραµετρικό τεστ.


5.4 The Poisson Distribution.

Άσκηση 10, σελ Για τη μεταβλητή x (άτυπος όγκος) έχουμε: x censored_x 1 F 3 F 3 F 4 F 10 F 13 F 13 F 16 F 16 F 24 F 26 F 27 F 28 F

2. ΧΡΗΣΗ ΣΤΑΤΙΣΤΙΚΩΝ ΠΑΚΕΤΩΝ ΣΤΗ ΓΡΑΜΜΙΚΗ ΠΑΛΙΝΔΡΟΜΗΣΗ

p n r

Si + Al Mg Fe + Mn +Ni Ca rim Ca p.f.u

Εισαγωγή στην Ανάλυση Συνδιακύμανσης (Analysis of Covariance, ANCOVA)

Dietary success of a new key fish in an overfished ecosystem: evidence from fatty acid and stable isotope signatures

!"!"!!#" $ "# % #" & #" '##' #!( #")*(+&#!', & - #% '##' #( &2(!%#(345#" 6##7

Παράδειγμα: Γούργουλης Βασίλειος, Επίκουρος Καθηγητής Τ.Ε.Φ.Α.Α.-Δ.Π.Θ.

ΤΜΗΜΑΕΠΙΧΕΙΡΗΜΑΤΙΚΟΥΣΧΕΔΙΑΣΜΟΥ & ΠΛΗΡΟΦΟΡΙΑΚΩΝΣΥΣΤΗΜΑΤΩΝ. ΤΕΧΝΙΚΕΣ ΠΡΟΒΛΕΨΕΩΝ& ΕΛΕΓΧΟΥ ΜΑΘΗΜΑ ΘΕΩΡΙΑΣ-ΣΤΑΣΙΜΕΣΔΙΑΔΙΚΑΣΙΕΣ-ΥΠΟΔΕΙΓΜΑΤΑ ARIMA (p,d,q)

Bayesian SEM: A more flexible representation of substantive theory Web tables

Επιστηµονική Επιµέλεια ρ. Γεώργιος Μενεξές. Εργαστήριο Γεωργίας. Viola adorata

[2] T.S.G. Peiris and R.O. Thattil, An Alternative Model to Estimate Solar Radiation


Παράδειγμα: Γούργουλης Βασίλειος, Επίκουρος Καθηγητής Τ.Ε.Φ.Α.Α.-Δ.Π.Θ.

ΔΙΠΛΩΜΑΤΙΚΗ ΕΡΓΑΣΙΑ: Mετακύλιση τιμών βασικών προϊόντων και τροφίμων στην περίπτωση του Νομού Αιτωλοακαρνανίας

Shenzhen Lys Technology Co., Ltd

Δθαξκνζκέλα καζεκαηηθά δίθηπα: ε πεξίπησζε ηνπ ζπζηεκηθνύ θηλδύλνπ ζε κηθξνεπίπεδν.

Suppose Mr. Bump observes the selling price and sales volume of milk gallons for 10 randomly selected weeks as follows

Δοκιμές προτίμησης και αποδοχής

MATHACHij = γ00 + u0j + rij

Supplementary Materials: A Preliminary Link between Hydroxylated Metabolites of Polychlorinated Biphenyls and Free Thyroxin in Humans

Aluminum Electrolytic Capacitors (Large Can Type)

SOLUTIONS TO MATH38181 EXTREME VALUES AND FINANCIAL RISK EXAM

Χαρακτηριστικά της ανάλυσης διασποράς. ΑΝΑΛΥΣΗ ΙΑΣΠΟΡΑΣ (One-way analysis of variance)

Table IB. Gross over-enumeration of establishments in the Census (based on the. Κατ' έκτίμησιν ποσοστόν διαφυγής. overenumerated. proportion

Ceramic PTC Thermistor Overload Protection

NMBTC.COM /

Supplementary Information 1.


Daewoo Technopark A-403, Dodang-dong, Wonmi-gu, Bucheon-city, Gyeonggido, Korea LM-80 Test Report


Chapter 1 Introduction to Observational Studies Part 2 Cross-Sectional Selection Bias Adjustment

Data sheet Thick Film Chip Resistor 5% - RS Series 0201/0402/0603/0805/1206

NKT NTC Thermistor. Negative Temperature Coefficient Thermistor FEATURES

LAMPIRAN. Fixed-effects (within) regression Number of obs = 364 Group variable (i): kode Number of groups = 26

Supplementary figures

Homework 3 Solutions

Έλεγχος υποθέσεων ΙI ANOVA

HW 3 Solutions 1. a) I use the auto.arima R function to search over models using AIC and decide on an ARMA(3,1)

ΔΕΙΚΤΗΣ ΜΑΖΑΣ ΣΩΜΑΤΟΣ ΚΑΙ ΟΙ ΔΙΑΤΡΟΦΙΚΕΣ ΣΥΝΗΘΕΙΕΣ ΦΟΙΤΗΤΩΝ

Aluminum Electrolytic Capacitors

PENGARUHKEPEMIMPINANINSTRUKSIONAL KEPALASEKOLAHDAN MOTIVASI BERPRESTASI GURU TERHADAP KINERJA MENGAJAR GURU SD NEGERI DI KOTA SUKABUMI

Description of the PX-HC algorithm

Αξιολόγηση των Φασματικού Διαχωρισμού στην Διάκριση Διαφορετικών Τύπων Εδάφους ΔΙΠΛΩΜΑΤΙΚΗ ΕΡΓΑΣΙΑ. Σπίγγος Γεώργιος

Table 2 Suggested prediction methods to use for a given set of available input parameters per each examined soil hydraulic property) a.


RC series Thick Film Chip Resistor

519.22(07.07) 78 : ( ) /.. ; c (07.07) , , 2008

Μηχανική Μάθηση Hypothesis Testing

Does anemia contribute to end-organ dysfunction in ICU patients Statistical Analysis

Ταυτοποίηση ασθενούς μέσω ραδιοσυχνικής αναγνώρισης (RFID) με σκοπό τον έλεγχο της χορήγησης αναισθησίας κατά τη διάρκεια χειρουργικής επέμβασης

APPENDICES APPENDIX A. STATISTICAL TABLES AND CHARTS 651 APPENDIX B. BIBLIOGRAPHY 677 APPENDIX C. ANSWERS TO SELECTED EXERCISES 679

Εισαγωγή στην Ανάλυση Διακύμανσης

Other Test Constructions: Likelihood Ratio & Bayes Tests

Szabolcs Sofalvi, M.S., D-ABFT-FT Cleveland, Ohio

Homework for 1/27 Due 2/5

7KH VHQVLWLYLW\ RI UDGLRWKHUDS\ WR WLVVXH FRPSRVLWLRQ DQG LWV HVWLPDWLRQ XVLQJ QRYHO GXDO HQHUJ\ &7 PHWKRGV Guillaume Landry

ΣΤΑΤΙΣΤΙΚΗ ΕΠΙΧΕΙΡΗΣΕΩΝ ΕΙΔΙΚΑ ΘΕΜΑΤΑ. Κεφάλαιο 13. Συμπεράσματα για τη σύγκριση δύο πληθυσμών

Na/K (mole) A/CNK

SOLUTIONS TO MATH38181 EXTREME VALUES AND FINANCIAL RISK EXAM

for fracture orientation and fracture density on physical model data

RECIPROCATING COMPRESSOR CALCULATION SHEET ISOTHERMAL COMPRESSION Gas properties, flowrate and conditions. Compressor Calculation Sheet

DATA SHEET Surface mount NTC thermistors. BCcomponents

Summary of the model specified

SECTION II: PROBABILITY MODELS

ΔΙΑΚΡΙΤΟΣ ΜΕΤΑΣΧΗΜΑΤΙΣΜΟΣ FOURIER - Discrete Fourier Transform - DFT -

EE512: Error Control Coding

RECIPROCATING COMPRESSOR CALCULATION SHEET

Π.Μ.Σ. ΒΙΟΣΤΑΤΙΣΤΙΚΗ ΑΝΑΛΥΣΗ ΔΙΑΣΠΟΡΑΣ ΚΑΙ ΠΑΛΙΝΔΡΟΜΗΣΗΣ ΤΕΛΙΚΟ ΔΙΑΓΩΝΙΣΜΑ 27/6/2016

Wan Nor Arifin under the Creative Commons Attribution-ShareAlike 4.0 International License. 1 Introduction 1

± 20% (rated cap. [µf] ) 1000 Leakage Current: For capacitance values > 33000µF, add the value of:

Fused Bis-Benzothiadiazoles as Electron Acceptors

; +302 ; +313; +320,.

Υπολογιστικές Μέθοδοι στις Κατασκευές

A life-table metamodel to support management of data deficient species, exemplified in sturgeons and shads. Electronic Supplementary Material

SAMSUNG ELECTRONICS CO., LTD TEST REPORT SAMSUNG ELECTRONICS CO., LTD. 1, Samsung-Ro, Giheung-Gu, Yongin-Si, Gyeonggi-Do 17113, Korea

TAMIL NADU PUBLIC SERVICE COMMISSION REVISED SCHEMES

Lampiran 1 Output SPSS MODEL I

Αθήνα: 28/8/14 Αριθμ. Πρωτ.:29062

the total number of electrons passing through the lamp.

NPI Unshielded Power Inductors

2 4 5

5.1 logistic regresssion Chris Parrish July 3, 2016

Ενότητα 3: Ανάλυση Διακύμανσης κατά ένα παράγοντα One-Way ANOVA

Wan Nor Arifin under the Creative Commons Attribution-ShareAlike 4.0 International License. 1 Introduction 1

! " # $ &,-" " (.* & -" " ( /* 0 (1 1* 0 - (* 0 #! - (#* 2 3( 4* 2 (* 2 5!! 3 ( * (7 4* 2 #8 (# * 9 : (* 9

255 (log-normal distribution) 83, 106, 239 (malus) 26 - (Belgian BMS, Markovian presentation) 32 (median premium calculation principle) 186 À / Á (goo

Μαντζούνη, Πιπερίγκου, Χατζή. ΒΙΟΣΤΑΤΙΣΤΙΚΗ Εργαστήριο 5 ο

ΠΟΛΥΤΕΧΝΕΙΟ ΚΡΗΤΗΣ ΣΧΟΛΗ ΜΗΧΑΝΙΚΩΝ ΠΕΡΙΒΑΛΛΟΝΤΟΣ

ΠΣΤΥΙΑΚΗ ΔΡΓΑΙΑ. Μειέηε Υξόλνπ Απνζηείξσζεο Κνλζέξβαο κε Τπνινγηζηηθή Ρεπζηνδπλακηθή. Αζαλαζηάδνπ Βαξβάξα

Μενύχτα, Πιπερίγκου, Σαββάτης. ΒΙΟΣΤΑΤΙΣΤΙΚΗ Εργαστήριο 5 ο

Transcript:

Appendix A3 Table A3.1. General linear model results for RMSE under the unconditional model. Source DF SS F Value Pr > F Model 107 374.68 3.50 8573.07 <.0001 Error 53786 21.97 0.00 Corrected Total 53893 396.65 R- Coeff Var Root MSE 0.944614 3.84876 0.02 0.53 Proportion of Variation Accounted for Eta-d 0.94 ω2 0.94 95% CI (0.94,0.95) Source DF Type III SS F Value Pr > F Semipartial Eta-d SS 2 0.05 0.03 66.23 <.0001 0.00 NI 2 324.96 162.48 397799.00 <.0001 0.82 IM 2 41.83 20.92 51211.10 <.0001 0.11 DF 1 1.74 1.74 4264.15 <.0001 0.00 PD 1 2.73 2.73 6695.68 <.0001 0.01 SS*NI 4 0.00 0.00 0.42 0.80 0.00 SS*IM 4 0.00 0.00 0.75 0.56 0.00 SS*DF 2 0.00 0.00 0.64 0.53 0.00 SS*PD 2 0.00 0.00 1.48 0.23 0.00 NI*IM 4 0.47 0.12 286.17 <.0001 0.00 NI*DF 2 0.49 0.25 602.99 <.0001 0.00 NI*PD 2 0.92 0.46 1126.22 <.0001 0.00 IM*DF 2 0.10 0.05 117.13 <.0001 0.00 IM*PD 2 0.06 0.03 69.85 <.0001 0.00 DF*PD 1 1.14 1.14 2799.16 <.0001 0.00 SS*NI*IM 8 0.01 0.00 1.63 0.11 0.00 SS*NI*DF 4 0.00 0.00 0.90 0.46 0.00 SS*NI*PD 4 0.00 0.00 0.19 0.94 0.00 SS*IM*DF 4 0.00 0.00 1.73 0.14 0.00 SS*IM*PD 4 0.00 0.00 2.32 0.05 0.00 SS*DF*PD 2 0.00 0.00 1.11 0.33 0.00 NI*IM*DF 4 0.01 0.00 7.96 <.0001 0.00 NI*IM*PD 4 0.03 0.01 15.88 <.0001 0.00 NI*DF*PD 2 0.20 0.10 248.13 <.0001 0.00 IM*DF*PD 2 0.00 0.00 4.92 0.01 0.00 SS*NI*IM*DF 8 0.00 0.00 0.52 0.84 0.00 SS*NI*IM*PD 8 0.00 0.00 1.04 0.41 0.00 SS*NI*DF*PD 4 0.00 0.00 1.51 0.20 0.00 SS*IM*DF*PD 4 0.00 0.00 0.70 0.59 0.00 NI*IM*DF*PD 4 0.00 0.00 1.18 0.32 0.00 SS*NI*IM*DF*PD 8 0.00 0.00 0.53 0.83 0.00 Note: Under Source, the following abbreviations are used: SS =sample size; NI = number of items; IM = impact condition; DF = DIF magnitude; PD = percent of items with DIF.

Table A3.2. General linear model results for RMSE under the impact-only model. Source DF SS F Value Pr > F Model 107 407.27 3.81 5419.33 <.0001 Error 53786 37.78 0.00 Corrected Total 53893 445.05 R- Coeff Var Root MSE 0.915118 5.197 0.03 0.51 Proportion of Variation Accounted for Eta-d 0.92 ω2 0.91 95% CI (0.91,0.92) Source DF Type III SS F Value Pr > F Semipartial Eta-d SS 2 2.25 1.12 1600.43 <.0001 0.01 NI 2 284.29 142.14 202384.00 <.0001 0.64 IM 2 4.19 2.10 2984.11 <.0001 0.01 DF 1 40.75 40.75 58023.30 <.0001 0.09 PD 1 54.05 54.05 76951.70 <.0001 0.12 SS*NI 4 0.40 0.10 141.69 <.0001 0.00 SS*IM 4 0.09 0.02 30.48 <.0001 0.00 SS*DF 2 0.08 0.04 56.16 <.0001 0.00 SS*PD 2 0.06 0.03 44.97 <.0001 0.00 NI*IM 4 0.56 0.14 199.17 <.0001 0.00 NI*DF 2 0.04 0.02 27.84 <.0001 0.00 NI*PD 2 0.67 0.34 478.49 <.0001 0.00 IM*DF 2 0.16 0.08 113.59 <.0001 0.00 IM*PD 2 0.06 0.03 45.26 <.0001 0.00 DF*PD 1 20.43 20.43 29090.10 <.0001 0.05 SS*NI*IM 8 0.07 0.01 12.30 <.0001 0.00 SS*NI*DF 4 0.02 0.01 8.32 <.0001 0.00 SS*NI*PD 4 0.02 0.00 5.99 <.0001 0.00 SS*IM*DF 4 0.01 0.00 2.76 0.03 0.00 SS*IM*PD 4 0.00 0.00 1.51 0.20 0.00 SS*DF*PD 2 0.03 0.02 22.99 <.0001 0.00 NI*IM*DF 4 0.02 0.00 5.75 0.00 0.00 NI*IM*PD 4 0.00 0.00 1.03 0.39 0.00 NI*DF*PD 2 0.48 0.24 343.87 <.0001 0.00 IM*DF*PD 2 0.02 0.01 10.97 <.0001 0.00 SS*NI*IM*DF 8 0.00 0.00 0.66 0.73 0.00 SS*NI*IM*PD 8 0.01 0.00 0.94 0.48 0.00 SS*NI*DF*PD 4 0.01 0.00 3.49 0.01 0.00 SS*IM*DF*PD 4 0.00 0.00 0.61 0.66 0.00 NI*IM*DF*PD 4 0.00 0.00 0.70 0.59 0.00 SS*NI*IM*DF*PD 8 0.00 0.00 0.58 0.80 0.00 Note: Under Source, the following abbreviations are used: SS =sample size; NI = number of items; IM = impact condition; DF = DIF magnitude; PD = percent of items with DIF.

Table A3.3. General linear model results for RMSE under the impact-and-dif model. Source DF SS F Value Pr > F Model 107 313.33 2.93 6926.33 <.0001 Error 53786 22.74 0.00 Corrected Total 53893 336.07 R- Coeff Var Root MSE 0.932336 4.39733 0.02 0.47 Proportion of Variation Accounted for Eta-d 0.93 ω2 0.93 95% CI (0.93,0.93) Source DF Type III SS F Value Pr > F Semipartial Eta-d SS 2 2.87 1.44 3398.87 <.0001 0.01 NI 2 303.33 151.67 358736.00 <.0001 0.90 IM 2 5.59 2.79 6605.16 <.0001 0.02 DF 1 0.02 0.02 49.83 <.0001 0.00 PD 1 0.66 0.66 1552.12 <.0001 0.00 SS*NI 4 0.37 0.09 218.47 <.0001 0.00 SS*IM 4 0.13 0.03 76.42 <.0001 0.00 SS*DF 2 0.00 0.00 0.79 0.45 0.00 SS*PD 2 0.15 0.08 179.73 <.0001 0.00 NI*IM 4 0.39 0.10 231.88 <.0001 0.00 NI*DF 2 0.00 0.00 0.66 0.51 0.00 NI*PD 2 0.12 0.06 141.49 <.0001 0.00 IM*DF 2 0.05 0.03 64.33 <.0001 0.00 IM*PD 2 0.00 0.00 2.18 0.11 0.00 DF*PD 1 0.00 0.00 9.22 0.00 0.00 SS*NI*IM 8 0.06 0.01 18.04 <.0001 0.00 SS*NI*DF 4 0.00 0.00 1.10 0.35 0.00 SS*NI*PD 4 0.05 0.01 27.58 <.0001 0.00 SS*IM*DF 4 0.00 0.00 1.18 0.32 0.00 SS*IM*PD 4 0.02 0.00 10.40 <.0001 0.00 SS*DF*PD 2 0.00 0.00 2.03 0.13 0.00 NI*IM*DF 4 0.00 0.00 2.34 0.05 0.00 NI*IM*PD 4 0.01 0.00 7.12 <.0001 0.00 NI*DF*PD 2 0.00 0.00 1.59 0.20 0.00 IM*DF*PD 2 0.00 0.00 2.35 0.10 0.00 SS*NI*IM*DF 8 0.00 0.00 0.52 0.84 0.00 SS*NI*IM*PD 8 0.01 0.00 2.94 0.00 0.00 SS*NI*DF*PD 4 0.00 0.00 2.32 0.05 0.00 SS*IM*DF*PD 4 0.00 0.00 0.62 0.65 0.00 NI*IM*DF*PD 4 0.00 0.00 0.72 0.58 0.00 SS*NI*IM*DF*PD 8 0.00 0.00 0.88 0.53 0.00 Note: Under Source, the following abbreviations are used: SS =sample size; NI = number of items; IM = impact condition; DF = DIF magnitude; PD = percent of items with DIF.

Table A3.4. Average RMSE across all scoring models and design factors at N = 500. 6 Items Unconditional Score Impact-Only Score Impact-and-DIF Score 33% Small DIF 0.651 0.035 0.589 0.051 0.585 0.043 66% Small DIF 0.653 0.036 0.614 0.054 0.606 0.059 33% Large DIF 0.653 0.036 0.602 0.061 0.580 0.046 66% Large DIF 0.667 0.038 0.733 0.113 0.606 0.066 33% Small DIF 0.612 0.025 0.574 0.027 0.572 0.026 66% Small DIF 0.612 0.025 0.602 0.031 0.591 0.033 33% Large DIF 0.611 0.024 0.582 0.029 0.568 0.025 66% Large DIF 0.620 0.026 0.707 0.063 0.589 0.038 Large /Small Variance SD SD SD 33% Small DIF 0.601 0.020 0.547 0.022 0.545 0.022 66% Small DIF 0.598 0.022 0.575 0.028 0.557 0.026 33% Large DIF 0.600 0.021 0.561 0.026 0.546 0.022 66% Large DIF 0.604 0.021 0.694 0.062 0.562 0.028 12 Items Unconditional Score Impact-Only Score Impact-and-DIF Score 33% Small DIF 0.554 0.034 0.486 0.028 0.482 0.029 66% Small DIF 0.560 0.034 0.509 0.033 0.493 0.037 33% Large DIF 0.562 0.037 0.504 0.036 0.481 0.031 66% Large DIF 0.583 0.036 0.599 0.048 0.490 0.037 33% Small DIF 0.504 0.024 0.476 0.022 0.472 0.022 66% Small DIF 0.511 0.023 0.500 0.022 0.483 0.023 33% Large DIF 0.505 0.022 0.493 0.024 0.472 0.021 66% Large DIF 0.529 0.025 0.587 0.036 0.481 0.024 Large /Small Variance SD SD SD 33% Small DIF 0.489 0.019 0.461 0.019 0.456 0.018 66% Small DIF 0.492 0.018 0.484 0.022 0.462 0.019 33% Large DIF 0.488 0.018 0.484 0.023 0.456 0.018 66% Large DIF 0.506 0.019 0.579 0.037 0.465 0.019 24 Items Unconditional Score Impact-Only Score Impact-and-DIF Score 33% Small DIF 0.457 0.033 0.396 0.025 0.391 0.026 66% Small DIF 0.466 0.032 0.419 0.024 0.396 0.027 33% Large DIF 0.465 0.033 0.413 0.026 0.387 0.024 66% Large DIF 0.504 0.030 0.510 0.036 0.393 0.025 33% Small DIF 0.410 0.022 0.390 0.018 0.384 0.019 66% Small DIF 0.418 0.021 0.414 0.020 0.387 0.020 33% Large DIF 0.413 0.022 0.408 0.021 0.381 0.019 66% Large DIF 0.454 0.022 0.503 0.030 0.387 0.020 Large /Small Variance SD SD SD 33% Small DIF 0.391 0.017 0.377 0.017 0.369 0.016 66% Small DIF 0.401 0.017 0.405 0.020 0.377 0.017 33% Large DIF 0.390 0.017 0.401 0.021 0.368 0.016 66% Large DIF 0.429 0.016 0.504 0.029 0.377 0.017

Table A3.5. Average RMSE across all scoring models and design factors at N = 1000. 6 Items Unconditional Score Impact-Only Score Impact-and-DIF Score 33% Small DIF 0.649 0.0258 0.569 0.0233 0.565 0.0227 66% Small DIF 0.654 0.0255 0.596 0.0259 0.577 0.0279 33% Large DIF 0.651 0.0246 0.580 0.0269 0.563 0.0218 66% Large DIF 0.664 0.0259 0.696 0.0494 0.569 0.0263 33% Small DIF 0.610 0.0178 0.563 0.0172 0.560 0.0171 66% Small DIF 0.608 0.0175 0.588 0.0192 0.569 0.0181 33% Large DIF 0.608 0.0194 0.569 0.0193 0.557 0.0177 66% Large DIF 0.618 0.0182 0.685 0.0360 0.567 0.0182 Large /Small Variance SD SD SD 33% Small DIF 0.600 0.0147 0.537 0.0135 0.534 0.0136 66% Small DIF 0.596 0.0151 0.566 0.0183 0.544 0.0158 33% Large DIF 0.599 0.0151 0.550 0.0163 0.536 0.0144 66% Large DIF 0.602 0.0158 0.673 0.0386 0.546 0.0162 12 Items Unconditional Score Impact-Only Score Impact-and-DIF Score 33% Small DIF 0.552 0.0237 0.479 0.0186 0.473 0.0189 66% Small DIF 0.557 0.0241 0.498 0.0200 0.475 0.0200 33% Large DIF 0.558 0.0257 0.493 0.0214 0.469 0.0203 66% Large DIF 0.581 0.0240 0.581 0.0282 0.471 0.0196 33% Small DIF 0.504 0.0171 0.471 0.0153 0.466 0.0152 66% Small DIF 0.509 0.0164 0.494 0.0152 0.470 0.0152 33% Large DIF 0.505 0.0163 0.489 0.0156 0.465 0.0141 66% Large DIF 0.528 0.0173 0.576 0.0240 0.469 0.0152 Large /Small Variance SD SD SD 33% Small DIF 0.486 0.0129 0.453 0.0128 0.446 0.0120 66% Small DIF 0.491 0.0128 0.477 0.0135 0.453 0.0126 33% Large DIF 0.485 0.0130 0.476 0.0151 0.448 0.0116 66% Large DIF 0.505 0.0131 0.573 0.0247 0.456 0.0128 24 Items Unconditional Score Impact-Only Score Impact-and-DIF Score 33% Small DIF 0.454 0.0226 0.391 0.0169 0.384 0.0171 66% Small DIF 0.467 0.0235 0.415 0.0177 0.387 0.0186 33% Large DIF 0.465 0.0235 0.409 0.0173 0.381 0.0177 66% Large DIF 0.505 0.0234 0.501 0.0229 0.384 0.0185 33% Small DIF 0.406 0.0160 0.384 0.0133 0.376 0.0135 66% Small DIF 0.416 0.0144 0.408 0.0129 0.378 0.0127 33% Large DIF 0.412 0.0158 0.403 0.0145 0.374 0.0131 66% Large DIF 0.451 0.0157 0.494 0.0178 0.379 0.0135 Large /Small Variance SD SD SD 33% Small DIF 0.389 0.0113 0.372 0.0108 0.364 0.0107 66% Small DIF 0.398 0.0111 0.400 0.0120 0.368 0.0103 33% Large DIF 0.391 0.0109 0.396 0.0136 0.364 0.0107 66% Large DIF 0.427 0.0117 0.494 0.0205 0.369 0.0113

Table A3.6. Average RMSE across all scoring models and design factors at N = 2000. 6 Items Unconditional Score Impact-Only Score Impact-and-DIF Score 33% Small DIF 0.650 0.0180 0.564 0.0158 0.560 0.0150 66% Small DIF 0.654 0.0189 0.587 0.0172 0.564 0.0161 33% Large DIF 0.649 0.0186 0.571 0.0173 0.555 0.0153 66% Large DIF 0.666 0.0193 0.684 0.0294 0.559 0.0149 Medium /Medium Variance 33% Small DIF 0.609 0.0137 0.559 0.0123 0.556 0.0124 66% Small DIF 0.607 0.0123 0.582 0.0130 0.559 0.0113 33% Large DIF 0.608 0.0125 0.565 0.0125 0.553 0.0115 66% Large DIF 0.619 0.0135 0.675 0.0242 0.559 0.0119 Large /Small Variance 33% Small DIF 0.599 0.0101 0.534 0.0093 0.531 0.0091 66% Small DIF 0.594 0.0098 0.559 0.0118 0.535 0.0100 33% Large DIF 0.598 0.0110 0.544 0.0106 0.530 0.0098 66% Large DIF 0.602 0.0109 0.664 0.0247 0.538 0.0101 12 Items Unconditional Score Impact-Only Score Impact-and-DIF Score 33% Small DIF 0.552 0.0175 0.475 0.0137 0.468 0.0136 66% Small DIF 0.557 0.0173 0.494 0.0130 0.468 0.0129 33% Large DIF 0.555 0.0179 0.488 0.0143 0.463 0.0134 66% Large DIF 0.581 0.0183 0.576 0.0193 0.465 0.0137 Medium /Medium Variance 33% Small DIF 0.501 0.0116 0.468 0.0098 0.461 0.0101 66% Small DIF 0.508 0.0114 0.490 0.0107 0.465 0.0103 33% Large DIF 0.504 0.0114 0.485 0.0107 0.460 0.0102 66% Large DIF 0.526 0.0112 0.571 0.0165 0.463 0.0095 Large /Small Variance 33% Small DIF 0.486 0.0089 0.451 0.0087 0.444 0.0082 66% Small DIF 0.490 0.0094 0.475 0.0102 0.448 0.0087 33% Large DIF 0.484 0.0089 0.473 0.0100 0.444 0.0083 66% Large DIF 0.505 0.0102 0.566 0.0166 0.450 0.0091 24 Items Unconditional Score Impact-Only Score Impact-and-DIF Score 33% Small DIF 0.454 0.0157 0.389 0.0114 0.381 0.0117 66% Small DIF 0.467 0.0170 0.413 0.0126 0.383 0.0131 33% Large DIF 0.462 0.0170 0.406 0.0129 0.378 0.0125 66% Large DIF 0.503 0.0160 0.495 0.0148 0.379 0.0120 Medium /Medium Variance 33% Small DIF 0.406 0.0105 0.382 0.0091 0.374 0.0091 66% Small DIF 0.416 0.0108 0.405 0.0093 0.375 0.0095 33% Large DIF 0.411 0.0104 0.400 0.0092 0.372 0.0090 66% Large DIF 0.450 0.0104 0.491 0.0129 0.375 0.0089 Large /Small Variance 33% Small DIF 0.387 0.0080 0.370 0.0074 0.361 0.0073 66% Small DIF 0.397 0.0077 0.396 0.0084 0.365 0.0075 33% Large DIF 0.389 0.0080 0.392 0.0092 0.361 0.0075 66% Large DIF 0.426 0.0088 0.490 0.0145 0.366 0.0081