Module 7: Test Development Flashcards
Test Development
an umbrella term for all that goes into the process of creating a test
Test Conceptualization
+ brainstorming of ideas about what kind of test a developer wants to publish
+ entail literature reviews and experimentation, creation, revision, and deletion of preliminary items
What kind of information is determined in test conceptualization?
- Construct
- Goal
- User
- Taker
- Administration
- Format
- Response
- Benefits
- Costs
- Interpretation
- Whether the test is norm-referenced or criterion-referenced
- How best to measure a targeted construct
Pilot work/study/research
preliminary reserch surrounding the creation of a prototype of the test
Pilot work/study/research
preliminary reserch surrounding the creation of a prototype of the test
Test Construction
stage in the process that entails writing test items, revisions, formatting, setting scoring rules
What kind of item should not be made?
It is not good to create an item that contains numerous ideas
Item Pool
reservoir or well from which the items will or will not be drawn for the final version of the test
Item Banks
relatively large and easily accessible collection of test questions
Computer Adaptive Testing
+ refers to an interactive, computer administered test-taking process wherein items presented to the testtaker are based in part on the testtaker’s performance on previous item
+ test administered may be different for each testtaker, depending on the test performance on the items presented
What does Computer Adaptive Testing reduce?
Reduces floor and ceiling effects
Floor Effects
occurs when there is some lower limit on a survey or questionnaire and a large percentage of respondents score near this lower limit (testtakers have low scores)
Ceiling Effects
occurs when there is some upper limit on a survey or questionnaire and a large percentage of respondents score near this upper limit (testtakers have high scores)
Item Branching
ability of the computer to tailor the
content and order of presentation of items on the basis of responses to previous items
Item Format
form, plan, structure, arrangement, and layout of individual test items
What are the different kinds of item formats?
- Dichotomous format
- Polychotomous format
- Category format
Dichotomous Format
offers two alternatives for each item
Polychotomous Format
each item has more than two alternatives
Category Format
a format where respondents are asked to rate a construct
Checklist
subject receives a longlist of adjectives and indicates whether each one if characteristic of himself or herself
Guttman Scale
items are arranged from weaker to
stronger expressions of attitude, belief, or feelings
Selected-Response Format
require testtakers to select
response from a set of alternative responses
What are the three elements of a multiple choice format?
- stem (question),
- a correct option, and
- several incorrect alternatives (distractors or foils)
Multiple Choice Format
Should’ve one correct answer, has grammatically parallel alternatives, similar length, alternatives that fit grammatically with the stem, avoid ridiculous distractors, not excessively long, “all of the above”, “none of the above” (25%)
What are the different kinds of distractors?
- Effective distractors
- Ineffective distractors
- Cute distractors
Effective Distractors
a distractor that was chosen equally by both high and low performing groups that enhances the consistency of test results
Ineffective Distractors
may hurt the reliability of the test because they are time consuming to read and can limit the no. of good items
Cute Distractors
less likely to be chosen, may affect the reliability of the test bec the testtakers may guess from the remaining options
Who are most likely to choose good distractors?
Good distractors has been chosen frequently by low scorers
Matching Item Format
Test taker is presented with two columns: Premises and Responses
Binary Choice Format
Usually takes the form of a sentence that requires the testtaker to indicate whether the statement is or is not a fact (50%)
Constructed-Response Format
requires testtakers to supply or to create the correct answer, not merely selecting it
Complete Item
requires the examinee to provide a word or phrase that completes a sentence
Short-Answer Format
Should be written clearly enough that the testtaker can respond succinctly, with short answer
Essay
allows creative integration and expression of the material
Scaling
process of setting rules for assigning numbers in measurement
Types of Selected-Response Format
- Multiple Choice
- Matching Items
- Binary Choice
Types of Selected-Response Format
- Multiple Choice
- Matching Items
- Binary Choice
Types of Constructed-Response Format
- Completion Item
- Short-Answer
- Essay
What are the primary scales of measurement?
- Nominal
- Ordinal
- Ratio
- Interval
Nominal
+ involve classification or categorization based on one or more distinguishing characteristics
+ label and categorize observations but do not make any quantitative distinctions between observations
Ordinal
+ rank ordering on some characteristics is also permissible
+ median
Ratio
+ contains equal intervals, has no absolute zero point (even negative values have interpretation to it)
+ Zero value does not mean it represents none
Interval
+ has true zero point (if the score is zero, it means none/null)
+ Easiest to manipulate
What are the comparative scales of measurement?
- Paired Comparison
- Rank Order
- Constant Sum
- Q-Sort Technique
What are the comparative scales of measurement?
- Paired Comparison
- Rank Order
- Constant Sum
- Q-Sort Technique
Paired Comparison
+ produces ordinal data by presenting with pairs of two stimuli which they are asked to compare
+ respondent is presented with two objects at a time and asked to select one object according to some criterion
Rank Order
respondents are presented with
several items simultaneously and asked to rank them in
order or priority
Constant Sum
respondents are asked to allocate a constant sum of units, such as points, among set of stimulus objects with respect to some criterion
Q-Sort Technique
sort object based on similarity with respect to some criterion
Continuous Rating
rate the objects by placing a mark at the appropriate position on a continuous line that runs from one extreme of the criterion variable to the other
e.g., Rating Guardians of the Galaxy as the best Marvel Movie of Phase 4
Itemized Rating
having numbers or brief descriptions associated with each category
e.g., 1 if your like the item the most, 2 if so-so, 3 if you hate it
Likert Scale
+ indicate their own attitudes by checking how strongly they agree or disagree with carefully worded statements that range from very positive to very negative towards attitudinal object
+ principle of measuring attitudes by asking people to respond to a series of statements about a topic, in terms of the extent to which they agree with them
Visual Analogue Scale
a 100-mm line that allows subjects to express the magnitude of an experience or belief
Semantic Differential Scale
derive respondent’s attitude towards the given object by asking him to select an appropriate position on a scale between two bipolar opposites
Staple Scale
developed to measure the direction and intensity of an attitude simultaneously
Summative Scale
final score is obtained by summing the ratings across all the items
Thurstone Scale
+ involves the collection of a variety of different statements about a phenomenon which are ranked by an expert panel in order to develop the questionnaire
+ allows multiple answers
Ipsative Scale
the respondent must choose between two or more equally socially acceptable options
Test Tryout
the test should be tried out on people who are similar in critical respects to the people for whom the test was designed
What is an informal rule of thumb for test tryouts?
An informal rule of thumb should be no fewer than 5 and preferably as many as 10 for each item (the more, the better)
What happens if there is a risk of using few subjects in a test?
Risk of using few subjects = phantom factors emerge
What kind of conditions should test tryouts be executed under?
Should be executed under conditions as identical as possible
What makes a good test item?
A good test item is one that answered correctly by high scorers as a whole
Empirical Criterion Keying
administering a large pool of test items to a sample of individuals who are known to differ on the construct being measure
Item Analysis
statistical procedure used to analyze items, evaluate test items
Discriminability Analysis
employed to examine correlation between each item and the total score of the test
Table of Specification
a blueprint of the test in terms of number of items per difficulty, topic importance, or taxonomy
Guidelines for Item Writing
Define clearly what to measure, generate item pool, avoid long items, keep the level of reading difficulty appropriate for those who will complete the test, avoid double-barreled items, consider making positive and negative worded items
Double-Barreled Items
items that convey more than one ideas at the same time
Item Difficulty
defined by the number of people who get a particular item correct
Item-Difficulty Index
calculating the proportion of the total number of testtakers who answered the item correctly; The larger, the easier the item
Item-Endorsement Index
for personality testing, percentage of individual who endorsed an item in a personality test
What is the optimal average item difficulty?
The optimal average item difficulty is approx. 50% with items on the testing ranging in difficulty from about 30% to 80%
What is the level of difficulty if the item difficulty range is 0.20-0.39?
Difficult
What is the level of difficulty if the item difficulty range is 0.0-0.19?
Very difficult
What is the level of difficulty if the item difficulty range is 0.40-0.60?
Average/moderately difficult
What is the level of difficulty if the item difficulty range is 0.80-1.0?
Very easy
What is the level of difficulty if the item difficulty range is 0.61-0.79?
Easy
Omnibus Spiral Format
items in an ability are arranged into increasing difficulty
Item-Reliability Index
provides an indication of the internal consistency of a test
What does it mean when the item-reliability index is high?
The higher Item-Reliability index, the greater the test’s internal consistency
Item-Validity Index
designed to provide an indication of the degree to which a test is measure what it purports to measure
What does it mean when the item-validity index is high?
The higher Item-Validity index, the greater the test’s criterion-related validity
Item-Discrimination Index
measure of item discrimination; measure of the difference between the proportion of high scorers answering an item correctly and the proportion of low scorers answering the item correctly
Extreme Group Method
compares people who have done well with those who have done poorly
Discrimination Index
difference between these proportion
Point-Biserial Method
correlation between a dichotomous variable and continuous variable
What does it mean when the correlation of an item is 0.40 and above?
Very good item
What does it mean when the correlation of an item is 0.30-0.39?
Good item
What does it mean when the correlation of an item is 0.20-0.29?
Fair item
What does it mean when the correlation of an item is 0.09-0.19?
Poor item
Item-Characteristic Curve
graphic representation of item difficulty and discrimination
Guessing
one that eluded any universally accepted solutions
What will happen if an item analysis is done for a speed test?
Item analyses taken under speed conditions yield misleading or uninterpretable results
How should item analysis be handled if it is for a speed test?
Restrict item analysis on a speed test only to the items completed by the testtaker
What should the test developer do when they are analyzing a speed test?
Test developer ideally should administer the test to be item-analyzed with generous time limits to complete the test
What are the types of scoring models?
- Cumulative Model
- Class Scoring/Category Scoring
- Ipsative Scoring
Cumulative Model
testtaker obtains a measure of the level of the trait; thus, high scorers may suggest high level in the trait being measured
Class Scoring/Category Scoring
testtaker response earn credit toward placement in a particular class or category with other testtaker whose pattern of responses is similar in some way
Ipsative Scoring
compares testtaker’s score on one scale within a test to another scale within that same test, two unrelated constructs
What is the process of test revision?
+ Characterize each item according to its strength and weaknesses
+ As revision proceeds, the advantage of writing a large item pool becomes more apparent because some items were removed and must be replaced by the items in the item pool
+ Administer the revised test under standardized conditions to a second appropriate sample of examinee
Cross-Validation
revalidation of a test on a sample of testtakers other than those on who test performance was originally found to be a valid predictor of some criterion; often results to validity shrinkage
Validity Shrinkage
decrease in item validities that inevitably occurs after cross-validation
Co-validation
conducted on two or more test using the same sample of testtakers
Co-norming
creation of norms or the revision of existing norms
Anchor Protocol
test protocol scored by highly authoritative scorer that is designed as a model for scoring and a mechanism for resolving scoring discrepancies
Scoring Drift
discrepancy between scoring in an anchor protocol and the scoring of another protocol
Differential Item Functioning
item functions differently in one group of testtakers known to have the same level of the underlying trait
DIF Analysis
test developers scrutinize group by group item response curves looking for DIF Items
DIF Items
items that respondents from different groups at the same level of underlying trait have different probabilities of endorsing a function of their group membership
Routing Test
subtest used to direct or route the
testtaker to a suitable level of items
Item-Mapping Method
setting cut scores that entails a histographic representation of items and expert judgments regarding item effectiveness
Basal Level
the level of which a the minimum criterion
number of correct responses is obtained
Computer Assisted Psychological Assessment
+ standardized test administration is assured for testtakers and variation is kept to a minimum
+ test content and length is tailored according to the taker’s ability
