Midterm 2 Review Flashcards

Vocabulary flashcards summarizing key concepts from the lecture notes.

SEM (Standard Error of Measurement)

Tells us how confident we are in our estimate of a student’s ability; the uncertainty in our inference of their true ability based on their responses.

Computer Adaptive Testing (CAT)

A method where a computer adjusts the test based on a student's performance, updating the ability estimate using MLE or Bayesian methods.

Classical Test Theory (CTT)

A theory that posits every tester has a True score (T) and random error score (E) that altogether form observed score (X): X = T+E.

Item Response Theory (IRT)

A theory that estimates a student's latent ability (θ) on a continuous scale and the difficulty, discrimination, and guessing behavior of each item.

Discrimination (IRT)

How well an item distinguishes ability.

Difficulty (IRT)

The point on θ where the item is 50% likely to be answered correctly.

Guessing (IRT)

The chance a low-ability student gets an item right.

Rasch Model

A part of IRT that assumes EQUAL discrimination and only measures difference in difficulty

Logit Scale

Takes log-odds of raw scores → makes scale linear; linear with respect to the difference between person ability and item difficulty. RASCH MODEL

Item Difficulty (Item Analysis)

Proportion of people who got an item right.

Item Discrimination (Item Analysis)

Whether an item differentiates high vs. low scorers.

Angoff Method

A method for determining cut scores where subject matter experts examine each test item and estimate the probability that a minimally competent candidate would answer it correctly.

Adequate Yearly Progress (AYP)

A component of the No Child Left Behind Act where states had to show progress, and schools that failed faced penalties.

Value Added Modeling (VAM)

A regression estimating a teacher’s/school’s “added value” to student test scores, controlling for prior performance/demographics.

NAEP (National Assessment of Educational Progress)

"Nation’s report card" conducted by US DOE, standardized across all states, for comparison.

Selection Bias

A problem in comparing student performance where students aren't randomly assigned, leading to skewed results.

problems with

External factors (e.g., SES, parental education) that differ between groups being compared, affecting the validity of comparisons.

Coleman Report (1966)

A public v private performance analysis using NAEP weighting scores based on socioeconomic status (SES).

No Excuses Schools

Charter schools aimed at serving low-income students of color with high expectations, strict discipline, and data-driven instruction.

Formative Assessment

Assessment used for feedback and improvement during the learning process.

Summative Assessment

Assessment used to judge performance at the end of a learning period.

Minimum Competency Exams (MCE)

Exit exams designed to ensure students have a minimal level of functional literacy/numeracy.

Standards-Based Exit Exams (SBEs)

Exit exams based on state academic standards.

Remediation

Targeted instruction given to students who failed a test, intended to help them meet the required standard on a retest.

Construct Validity

The extent to which a test actually measures the theoretical trait (construct) it's intended to measure.

Sensitivity (in Testing)

Ability to correctly identify true positives.
tpr = tp / (tp + Fn)

Specificity (in Testing)

Ability to correctly identify true negatives.

CAT (Computerized Adaptive Test)

All questions come from a calibrated item bank using IRT. Even though test forms differ, all students are evaluated against the same metric (θ), with equal opportunity to demonstrate mastery.

pushback against brown v board

Stell v. Savannah-Chatham County Board of Education (1963) -GEORGIA

Nomothetic Approach

Seeks to discover general laws that apply to many people when studying personality.

Idiographic Approach

Seeks to understand the uniqueness of a single person when studying personality.

Face Validity

Items are written to reflect the content of a concept based on (i.e., looks like it measures what it should).

Criterion-Based Selection

Items are selected based on their ability to differentiate between known groups (e.g., depressed vs. non-depressed).

Construct-Based Selection

Items are selected and validated through theoretical models and factor analysis, aimed at measuring latent psychological constructs.

Trait

DIMENSION that exists on a scale (high, mediu, low ) - ex) big 5

Type

category (ex. introvert vs extro

Objective Tests

Objectively scored. Results don’t hinge on the examiner’s interpretation —only on the keyed answer sheet

Projective Tests

Uses images or storytelling to see how someone “projects” inner motives and conflicts into the response, which the clinician interprets.

Content Based

constructed by defining a theory first, and then writing questions that directly represent that theory. Relies on face validity: the questions look like they measure what they say.

IQ testing racism cases

diana vs state board of education; LARRY P vs Riles 1979; hobson v hansen 1967

debra p vs turlington, florida

Does requiring students to pass a graduation test—when they were never previously taught the material—violate their constitutional rights to due process and equal protection?

CONtrast group

2 groups w proficiency level relative to the standard is already known through some external, reliable means (e.g., teacher ratings, supervisor evaluations, performance on a different benchmark).

bookmarking

Items are then arranged in an "Ordered Item Booklet" (OIB) from easiest to hardest based on these IRT difficulty values. The booklet shows each item along with its statistical informatio

American schools are failing; we’re falling behind global competitors—especially Japan and the USSR.

1983: nation at risk

provided federal grants to states and local districts—if they developed "standards-based reform" plans. This means creating clear academic standards and aligning curricula and tests to them.

1994: Goals 2000

• States set their own standards but had to show Adequate Yearly Progress (AYP)

• Schools that failed to meet AYP faced penalties, restructuring, or closure\

• Teaching to the test

2001 : No child left behind (NCLB) george bush

Make standards consistent across states; Detailed grade-by-grade standards

2010: common core standards (CCSB) / obama

3 ways to evaluate school success

1. student achievement scores

2. VAM

3. operational (graduation rates teacher turnover)

types of exit exams

MCE, Standards-Based Exit Exams (SBEs), end of course exams (EOC

con. of exit exams

remediation

teaching TO the test vs teaching THE test

Narrow focus on item formats, drill, tricks vs Teaching the core skills the test claims to measure

what happens when u raise cut score for exit exam

less people fail → less likely to pass underqualifed ppl → increases specificity ( catch more underprepared students ) → more ‘ready’ ppl who can’t graduate

SEM remains high when

poorly matched/low discrimination items

low sensitivity

= high false neg = low true positive

dubers religious coping (defining latent traits)

construct based