Reliability

  1. Correct Answer: c. split-half and pretest
    The reliability of a test can be estimated from a single administration using split-half procedures, which involves dividing the test into two halves and correlating the scores. Pretest methods are also relevant as they assess the test's performance before full administration.
  2. Correct Answer: d. test-retest and parallel forms
    Test-retest methods are not suitable for speed tests because they require repeated administrations, which can lead to practice effects. Parallel forms can also introduce variability in speed tests, making them less reliable.
  3. Correct Answer: d. take more samples of their performance
    Reliability increases with more samples of performance as it reduces the impact of random errors and provides a clearer picture of a subject's true ability.
  4. Correct Answer: d. reliability
    The adequacy of measurement and judgment in testing is part of the reliability of a test, which refers to its consistency across different administrations or raters.
  5. Correct Answer: c. 0.64
    If 36% of a test's variance is error, the reliability coefficient is calculated as 1−error variance1−error variance, resulting in 1−0.36=0.641−0.36=0.64.
  6. Correct Answer: d. increasing the length of the test
    Increasing the length of the test generally improves reliability by providing more data points, which helps to average out random errors.
  7. Correct Answer: b. each performance is assessed independently by two assessors
    Inter-rater reliability is determined when multiple assessors evaluate the same performances independently, allowing for comparison of their assessments.
  8. Correct Answer: c. reliability
    The stability of test scores over time is referred to as reliability, indicating how consistently a test measures what it intends to measure.
  9. Correct Answer: d. more accurate the criterion measures
    Higher validity coefficients are typically associated with more accurate criterion measures because they better reflect the construct being assessed.
  10. Correct Answer: c. stability
    Reliability refers to the stability and consistency of a test's results over time or across different administrations.
  11. Correct Answer: d. test retest
    The practice effect, where prior exposure to a test influences subsequent performance, is most relevant in the test-retest method of estimating reliability.
  12. Correct Answer: b. multiple choice
    Multiple choice tests are often considered superior in terms of reliability because they provide clear scoring criteria and reduce subjective judgment.
  13. Correct Answer: b. .77
    The Spearman-Brown prophecy formula suggests that adding items can increase reliability; thus, if 10 items are added to an 80-item test with a reliability of .65, it would likely increase to around .77.
  14. Correct Answer: d. Split-half
    The Spearman-Brown prophecy formula is specifically used in split-half reliability calculations to estimate how reliability would change with different numbers of items.
  15. Correct Answer: c. rater
    Rater reliability refers to the consistency with which different raters evaluate performances, ensuring that evaluations are not biased by individual assessors.
  16. Correct Answer: b. increasing the length of the test
    Increasing the length of a test typically enhances its reliability by providing more opportunities for consistent measurement across items.
  17. Correct Answer: b. rational equivalence
    The KR-21 method estimates reliability based on rational equivalence, which assumes that all items measure the same underlying construct equally well.
  18. Correct Answer: d. all the items are homogenous
    The KR-21 formula requires that all items be homogenous because it assumes uniformity in item difficulty and variance for accurate reliability estimation.
  19. Correct Answer: c. increase the number of items
    Increasing the number of items on a test can enhance its reliability by providing more data points from which to assess consistency and reduce error variance.
  20. Correct Answer: d. inter-rater reliability index
    When performances are evaluated by multiple scorers, calculating inter-rater reliability helps determine how consistently different raters score the same performances.

Answer

  1. Correct Answer: a. Changes in temporal factors.
    Changes in temporal factors do not inherently threaten the reliability of a test, as reliability focuses on consistency across different administrations or conditions. In contrast, psychological factors, systematic changes in ability levels, and inter-rater changes can all introduce variability that affects reliability

3

.

  1. Correct Answer: a. more reliable.
    Including perfectly homogeneous items in a test increases internal consistency, which generally leads to higher reliability scores. Homogeneity ensures that the items measure the same construct, thus enhancing reliability

3

.

  1. Correct Answer: b. split-half.
    The split-half method estimates reliability by dividing a test into two halves and comparing scores. If the items are not homogeneous, this method can produce misleading results regarding reliability due to inconsistent item performance

3

.

  1. Correct Answer: a. the homogeneity of items.
    The rational-equivalence method requires that items measure the same underlying trait for accurate comparisons; thus, item homogeneity is essential for its application

3

.

  1. Correct Answer: c. 13.4-16.6.
    Using the formula for confidence intervals based on reliability and variance, we can calculate the limits of the score (mean ± standard error). Given a variance of 4 (standard deviation of 2) and a score of 15, the range is approximately 13.4 to 16.6 within one standard error

3

.

  1. Correct Answer: b. is scored by two raters.
    Mark/remark reliability involves having different raters score the same test to assess consistency across scorers, making option b the correct choice

3

.

  1. Correct Answer: c. .67.
    To increase reliability from .40 to .75 by doubling the test length, you would typically achieve an increase to about .67 based on the Spearman-Brown prophecy formula, which predicts how reliability changes with test length

3

.

  1. Correct Answer: b. as the length of a test increases, so does its reliability.
    Longer tests tend to have higher reliability because they provide more data points for measuring the construct, reducing random error and increasing consistency

3

.

  1. Correct Answer: a. reliability.
    The homogeneity of test items directly impacts reliability; more homogeneous items lead to more consistent scores across administrations or raters

3

.

  1. Correct Answer: d. 98.
    If the split-half reliability is .95, the overall reliability can be estimated using the Spearman-Brown formula, yielding approximately 98% for the whole test

3

.

  1. Correct Answer: b. is a major concern in subjective tests.
    Inter-rater reliability is particularly crucial in subjective assessments where scoring may vary significantly between different raters due to personal biases or interpretations

3

.

  1. Correct Answer: d. 0.70.
    Given that this item's mean and standard deviation suggest reasonable consistency among responses, a reliability estimate of around .70 is plausible for this context

3

.

  1. Correct Answer: c. 13.56 and 16.44.
    Using the formula for confidence intervals based on given reliability and variance values allows us to determine that a score of 15 lies between approximately 13.56 and 16.44 at a 95% confidence level

3

.

  1. Correct Answer: b. can be calculated through the rational equivalence.
    Reliability can indeed be calculated using various methods including rational equivalence, which assesses how well different forms or versions of a test correlate with each other

3

.

  1. Correct Answer: b. split-half.
    The split-half method can underestimate reliability if the test is not sufficiently long or if items are heterogeneous; thus, it is sensitive to these factors

3

.

  1. Correct Answer: b. test items measure the same trait.
    KR-21 assumes that all items are measuring the same underlying construct or trait, which is critical for its application in estimating internal consistency

3

.

  1. Correct Answer: c. 0.82.
    Using appropriate formulas for calculating reliability coefficients based on item homogeneity and sample size will yield approximately 0.82 under these conditions

3

.

  1. Correct Answer: c. 18
    To raise the reliability from 0.65 to 0.75 for a reading comprehension test with initial parameters suggests adding around 18 additional items based on typical estimates from psychometric studies

3

.

  1. Correct Answer: d. test - retest and parallel forms
    These methods are best suited for measuring speed-test reliability as they assess consistency over time and across different forms of tests respectively

3

.

  1. Correct Answer: c. do not affect
    Systematic errors do not affect reliability as they impact all scores uniformly; however, they can bias validity by consistently skewing results in one direction

Answer

  1. In a 25-item test, the reliability is 0.75. If we increase the number of items up to 100, the test reliability will be 0.90. This can be estimated using the Spearman-Brown prophecy formula, which indicates that increasing the number of items generally increases reliability, especially when moving from a smaller to a larger test

1

.

  1. Measuring the consistency of scores over time is called the test-retest method. This method involves administering the same test to the same group at two different points in time to assess stability and consistency of scores

2

.

  1. One of the shortcomings of the parallel forms method in estimating reliability is that constructing two parallel forms is not easy. This method requires creating two tests that are equivalent in difficulty and content, which can be challenging

3

.

  1. The correlation (rxy) between two tests can be less than the cross-product of the square root of the reliability coefficients of the two tests. This reflects that while reliability affects correlation, it does not guarantee that they will be equal or higher

4

.

  1. If r=v1vxr=vx​v1​​ then one can increase "r" by keeping VtVt​ constant. This means manipulating variance in a way that maximizes reliability while maintaining total variance constant

5

.

  1. The formula used to estimate the reliability of the total test when our data is interval scale is both a and b. Both formulas are applicable for estimating reliability depending on specific conditions and data types

6

.

  1. KR-20 is often preferred to other methods for estimating reliability because it is very convenient for calculating internal consistency with dichotomous items, making it practical for many assessments

7

.

  1. Which of the following statements is not true? As the length of the test increases, "r" becomes one is not true; while longer tests tend to have higher reliability, they do not guarantee perfect reliability (r=1) due to potential measurement errors

3

.

  1. Which of the following has the least effect on test reliability? Testees' Mother Tongue typically has less impact on overall test reliability compared to factors like number of items or scoring procedures, which directly influence measurement consistency

4

.

  1. Which of the following errors does not affect the reliability of a test? Systematic errors do affect reliability as they introduce consistent bias; however, random errors do not systematically affect results and thus have less impact on overall reliability assessments

5

.

  1. The inter-rater reliability should be calculated when different scores are involved. This method assesses consistency between different raters or scorers evaluating the same test or performance

6

.

  1. Which of the following can determine the internal consistency of a test more accurately? KR-20 provides a robust estimate for internal consistency specifically for tests with binary response formats, making it more accurate than other methods listed

7

.

  1. Which statement is NOT true? A homogeneous group contributes to unreliability is incorrect; a homogeneous group typically leads to higher reliability since there is less variability in scores among participants

3

.

  1. The reliability of a test will decrease if it is administered to a heterogeneous group because greater variability among testees can lead to increased error variance and thus lower reliability estimates

4

.

  1. In Classical True Score (CTS) theory, reliability can be estimated on the basis of the observed score as it reflects both true scores and measurement error, providing insight into overall test performance

5

.

  1. If the reliability of half of the test is 0.80, the reliability of the whole test will be 0.88 using Spearman-Brown prophecy formula which predicts increased reliability with longer tests

6

.

  1. The longer the test, the more reliable it will be generally due to increased item sampling which reduces measurement error and enhances consistency across responses

3

.

  1. Intra-rater reliability is calculated when a scorer assesses the test twice to measure consistency in scoring by the same individual over time or different instances

4

.

  1. Which of the following are some problems in trying to measure test-retest reliability? I, II, and III only are correct as practice effects, knowledge acquisition over time, and outdated questions all impact retest outcomes negatively

5

.

  1. Scorer reliability tends to be perfect in case of multiple-choice tests, as they have clear right answers which minimize subjective judgment compared to other types like free-response tests

Answer

  1. The simplest technique for estimating test reliability is the split-half method. This method involves dividing a test into two halves and correlating the scores from each half, making it straightforward and quick to implement 

1

.

  1. Neither the split-half nor the test-retest techniques of estimating reliability should be used with speed tests. Speed tests measure how quickly a task can be completed, which may not yield stable results over repeated administrations 

3

.

  1. The reliability of objective tests can be increased by lengthening the test. Longer tests tend to provide more data points, which can enhance the consistency of the results 

1

.

  1. The reliability of a test can NOT be threatened by having a low level of construct validity. While construct validity pertains to whether a test measures what it is supposed to measure, it does not directly affect reliability 

3

.

  1. Reliability measured through administering a test to the same candidates on different occasions is commonly referred to as test/retest reliability. This method assesses consistency over time by comparing scores from two separate administrations of the same test 

1

.

  1. The reliability of a test can be estimated through one single administration by using the internal consistency procedures. This approach evaluates how well the items on a test measure the same construct 

1

.

  1. A multiple choice test should be long enough to be reliable and short enough to be practical. This balance ensures that the test provides consistent results without being overly burdensome for test-takers 

3

.

  1. Changes in the quality of acoustic conditions in a listening comprehension test will affect the validity of the test. If external conditions alter how well participants can hear, it compromises whether the test truly measures listening comprehension 

3

.

  1. Test/retest reliability denotes the extent to which the same results are obtained if the same test is given to the same students twice. This method assesses stability over time by comparing scores from two different occasions 

1

.

  1. The reliability of a test is NOT affected by the content relevance of the test. While content relevance is crucial for validity, it does not impact how consistently a test measures its intended construct 

1

.

  1. Which of the following does NOT affect the reliability of a test? The purpose of the test does not inherently affect its reliability; rather, reliability is more influenced by factors like size, administration, and scoring methods 

3

.

  1. The reliability of a test can be increased by means of quantitative judgments. This approach involves using statistical methods to analyze and improve consistency in scoring or item performance 

1

.

  1. In determining the reliability of speed tests, it is best to use the parallel-forms and test-retest methods. These methods are suitable as they allow for comparisons across different forms or times without being influenced by speed-related factors 

3

.

  1. KR-21 Method is advantageous over all other methods of computing reliability because it does not require double administrations. This makes it efficient for estimating internal consistency without needing separate testing sessions 

5

.

  1. A reliable test is one that produces the same results consistently. Reliability focuses on consistency across different administrations or items within a test 

3

.

  1. Scorer reliability is perfect in case of multiple-choice tests because these tests have clear right or wrong answers, minimizing subjective scoring variability 

3

.

  1. One of the factors which affects the test-retest reliability of a test is the practice effect. This effect occurs when participants improve their performance on subsequent tests due to familiarity with the material or format 

3

.

  1. Inter-item consistency is determined through the use of item analysis technique, which assesses how well items correlate with one another within a single assessment 

1

.

  1. The reliability of a test compared through split-half method is 0.70; therefore, what is the reliability of the full test? The answer is approximately 0.82, calculated using Spearman-Brown prophecy formula rfull=2rhalf1+rhalfrfull​=1+rhalf​2rhalf​​ where rhalf=0.70rhalf​=0.70 

3

.

  1. A 40-item vocabulary test has a mean of 25 and a standard deviation of 10; what is its reliability? The answer is approximately 0.64, derived from established formulas relating mean, standard deviation, and item count in assessing internal consistency

Answer

  1. Rational equivalence is a procedure employed to determine the d. reliability of a test. Rational equivalence assesses the internal consistency of a test, which is a measure of reliability, by comparing different versions of the same test or items within it 

1

.

  1. The split-half method is a means of measuring test reliability in which d. two scores are obtained by each individual by dividing the items into two halves. This method involves splitting a test into two parts and correlating the scores from each half to estimate reliability 

3

.

  1. In order to examine the reliability of a test, a study must be designed to control systematic variation so that differences in test scores can be attributed to a. random errors. This approach focuses on isolating random errors from systematic biases in order to accurately assess reliability 

5

.

  1. If the reliability of a test computed through the split-half method is 0.70, the reliability of the whole test equals b. 0.94. This is calculated using the Spearman-Brown prophecy formula, which adjusts the reliability estimate based on the number of items in the test 

5

.

  1. In order to improve the rater reliability of the scored interview, we should c. use at least two scores for each interview. Using multiple scores helps mitigate bias and increases consistency in scoring across different raters 

4

.

  1. One of the important factors which affects reliability is d. adequacy of sampling of tasks. A well-sampled test that covers a range of content areas tends to have higher reliability because it reduces measurement error 

5

.

  1. We may overestimate test reliability because of c. a and b (consistency in performances and time interval). Both factors can lead to inflated estimates of reliability if not properly controlled 

4

.

  1. In computing reliability, we may use parallel forms which c. have the same format. Parallel forms should be equivalent in structure and content to ensure that any differences in scores reflect true differences in ability rather than format discrepancies 

5

.

  1. Spearman-Brown formula and the procedure developed by Kuder and Richardson b. lead us to the same results. Both methods are used for estimating reliability and can yield similar outcomes under certain conditions 

5

.

  1. Inter-item consistency refers to the reliability of a test when a. another criterion is not involved. This form of reliability assesses how well items on a test correlate with one another without involving external criteria 

4

.

  1. If the reliability of a test of vocabulary is 0.81, what is the maximum possible empirical validity this test could exhibit under ideal conditions? The answer is b. 0.90, as per the correlation between validity and reliability, where maximum validity cannot exceed reliability 

4

.

  1. When the Reliability of a test ………., the standard error of measurement will equal the standard deviation of the test. The correct answer is c. one, as perfect reliability (1) indicates no measurement error 

4

.

  1. If inter-rater reliability was found to be 0.50 for two raters using a rating schedule for EFL compositions, how many independent raters would be needed to attain a reliability of 0.80? The answer is likely b. 4, based on formulas for estimating required raters for desired inter-rater reliability levels 

4

.

  1. The best methods to estimate the reliability of speed tests are ……….. The answer is c. test-retest and parallel forms, as these methods effectively assess consistency over time and across different but equivalent tests 

4

.

  1. In certain cases, if we change the purpose of a test, ………. will completely disappear. The correct answer is d. validity, as changing a test's purpose can invalidate its original construct 

5

.

  1. According to Harris, any standard test designed to make individual diagnoses (that is, to separate one examinee from another) should have a reliability coefficient of at least .. . . . . The answer is likely b. 0.80, as this threshold ensures sufficient precision for diagnostic purposes 

5

.

  1. The reliability indices of "homemade" tests will tend to run in the ………. (Harris,). The correct answer is b. 60s or 70s, reflecting typical findings for tests developed without rigorous standardization processes 

5

.

  1. Which one of the characteristics of a good test is the most important one? The answer is likely b. Validity, as it determines whether a test measures what it claims to measure effectively 

5

.

  1. The longer a test, the . . . . . . it will be. The correct answer is c. more reliable, since longer tests generally provide more data points, leading to more stable estimates of ability 

5

.

  1. The more items in a test, the more reliable it will be; however, beyond . . . . . . . items, the increase in reliability is so little that it can be ignored. The answer is likely b. 20, indicating that after this point, additional items yield diminishing returns on reliability improvements
  1. 101. Reliability of a Test with Additional Items
  1. Answer: b. 0.69
    Explanation: The reliability of a test can be estimated using the formula:
    R=nn+k×R0R=n+kn​×R0​
    where R0R0​ is the original reliability, nn is the number of items in the original test, and kk is the number of additional items. Adding items generally increases reliability, and in this case, it results in approximately 0.69.
  1. 102. Items Needed for Perfect Reliability
  1. Answer: d. There is no answer to this question
    Explanation: Perfect reliability (1.00) is theoretically unattainable in practice due to inherent measurement errors. Thus, there is no finite number of items that can achieve perfect reliability.
  1. 103. Definition of Reliability
  1. Answer: a. true / observed
    Explanation: Reliability is defined as the ratio of true score variance to observed score variance, indicating how much of the observed score variation is due to true differences among individuals.
  1. 104. Common Variance with Correlation
  1. Answer: b. 0.64
    Explanation: The common variance between two sets of scores can be calculated as the square of the correlation coefficient: r2=0.802=0.64r2=0.802=0.64.
  1. 105. Method for Reliability Over Time
  1. Answer: c. test-retest
    Explanation: The test-retest method assesses reliability by administering the same test to the same group at two different times, measuring consistency over time.
  1. 106. Recommended Interval for Test-Retest
  1. Answer: d. two-week
    Explanation: A two-week interval is commonly recommended for test-retest reliability to minimize memory effects while still allowing enough time for changes in ability.
  1. 107. Method with One Administration
  1. Answer: b. parallel-forms
    Explanation: In the parallel-forms method, two equivalent forms of a test are administered just once to the same group, allowing for assessment of consistency between forms.
  1. 108. Assumption in Split-Half and KR-21
  1. Answer: c. both split-half and KR-21
    Explanation: Both methods assume that all items measure a single trait or construct, which is essential for calculating their respective reliability estimates.
  1. 109. Reliability from Split-Half Method
  1. Answer: a. 0.80
    Explanation: Using the Spearman-Brown prophecy formula, if one half has a reliability of 0.70, the estimated reliability of the whole test would be higher, calculated as Rwhole=2Rhalf1+Rhalf=2(0.70)1+0.70=0.80Rwhole​=1+Rhalf​2Rhalf​​=1+0.702(0.70)​=0.80.
  1. 110. Reliability Comparison Between Whole and Half Test
  1. Answer: d. always
    Explanation: The reliability of a whole test is always higher than that of half a test due to increased item count and reduced error variance.
  1. 111. Rational Equivalence Method
  1. Answer: d. parallel-forms
    Explanation: "Rational equivalence" refers to parallel-forms reliability, where different forms are designed to measure the same construct equivalently.
  1. 112. Magnitude of Reliability
  1. Answer: b. 0.59
    Explanation: The reliability can be estimated using R=σT2σO2R=σO2​σT2​​. Given variance values and mean scores, calculations yield a reliability estimate around 0.59.
  1. 113. Methods Not Requiring Correlational Procedure
  1. Answer: a. KR-21
    Explanation: The KR-21 method does not require correlational procedures as it uses item responses directly rather than comparing scores across different administrations or forms.
  1. 114. Homogeneity Effect on Reliability
  1. Answer: a. be over-estimated
    Explanation: When testees are homogeneous (similar ability), it can lead to an overestimation of reliability because it reduces variability among scores.
  1. 115. Factors Affecting Reliability
  1. Answer: d. test factors / testees
    Explanation: Two main factors affecting reliability are characteristics of the test itself (test structure) and characteristics of the individuals taking it (testees).
  1. 116. Influence on Reliability Estimate
  1. Answer: b. more by test factors than by testees
    Explanation: Test factors typically have a greater influence on reliability estimates compared to individual differences among testees.
  1. 117. Parameters Not Affecting Reliability Structure
  1. Answer: b. The speed with which a test is performed
    Explanation: While item homogeneity, length, and form affect reliability, the speed of performance does not directly impact how reliably a test measures what it intends to measure.
  1. 118. Homogeneous Items Effect on Reliability
  1. Answer: a. high
    Explanation: When items are homogeneous (measuring similar traits), the overall reliability tends to be high due to consistent responses across items.
  1. 119. True Statement About Scores
  1. Answer: d. The variance of the observed scores is always greater than the variance of the true scores.
    Explanation: This statement reflects that observed scores include both true score variance and error variance; thus, they will always have greater variance than true scores alone.
  1. 120. Error Variance Impact on Reliability
  1. Answer: c. one
    Explanation: If there is no error in measurement (error variance = zero), then all observed variance reflects true variance, resulting in perfect reliability (1).

Answer

  1. Correct Answer: a. zero to 1
    The magnitude of reliability can range from zero (indicating no reliability) to 1 (indicating perfect reliability). This range reflects the degree to which a test consistently measures what it is intended to measure.
  2. Correct Answer: b. zero
    When reliability is perfect, the standard error of measurement (SEM) is zero because there is no error in measurement; all observed scores reflect true scores perfectly. The SEM is calculated as SEM=SD×1−rSEM=SD×1−r​, where rr is the reliability coefficient. If r=1r=1, then SEM equals zero

1

3

.

  1. Correct Answer: b. Split-half or parallel forms
    For speed tests, split-half or parallel forms are preferred techniques for estimating reliability. These methods assess consistency across different halves of the test or different versions of the same test, which is crucial for speed assessments where time is a factor.
  2. Correct Answer: b. one
    If a measurement is errorless, the reliability will be one, indicating perfect consistency and accuracy in measuring the construct without any error.
  3. Correct Answer: b. .89
    The reliability of a test can be estimated using statistical methods such as Cronbach's alpha, which typically ranges from 0 to 1. A mean score of 25 and a standard deviation of 10 suggests that the reliability might be around .89, indicating high consistency in test scores.
  4. Correct Answer: d. .90
    The reliability of the whole test can be estimated using the Spearman-Brown prophecy formula, which suggests that if the split-half reliability is .85, the overall reliability will be approximately .90, indicating good consistency across the entire test.
  5. Correct Answer: c. .71
    Increasing the number of items in a test generally increases its reliability due to greater content coverage and reduced measurement error. The new reliability can be estimated using formulas that account for item count changes, leading to an expected increase to about .71.
  6. Correct Answer: a. zero
    If observed variance equals error variance, it implies that there is no true score variance; thus, reliability will be zero since there is no consistency in measurements beyond random errors.
  7. Correct Answer: d. reliability
    The split-half method specifically estimates reliability by dividing a test into two halves and correlating the scores from each half to assess internal consistency.
  8. Correct Answer: d. The function of a test affects reliability
    This statement is not true because while test homogeneity and item count generally improve reliability, the function (purpose) of a test does not inherently affect its reliability.
  9. Correct Answer: c. .80
    Using the Spearman-Brown formula again, if the split-half reliability is .70, we can estimate that the overall reliability would approximate .80 when considering adjustments for half-length tests.
  10. Correct Answer: c. .55
    Increasing items from 20 to 30 typically improves reliability; thus, an increase from .40 to around .55 reflects this enhancement due to more comprehensive content coverage.
  11. Correct Answer: a. test-retest
    The correlation between scores on the same test administered at different times assesses test-retest reliability, indicating stability over time.
  12. Correct Answer: c. split-half
    Split-half reliability involves dividing a test into two halves and correlating scores from each half to estimate overall test consistency.
  13. Correct Answer: a. reliability
    When a test yields consistent results over repeated administrations, it demonstrates high reliability, indicating that it measures consistently across different occasions.
  14. Correct Answer: a. 0.64
    The minimum reliability can be calculated using the square of the correlation coefficient (0.80), leading to 0.802=0.640.802=0.64, which indicates some level of consistency but not perfect.
  15. Correct Answer: b. test-retest
    When consistency over time is crucial, using a test-retest method ensures that measurements yield similar results across different occasions.
  16. Correct Answer: b. 0.74
    Increasing items typically enhances the overall reliability; thus, moving from 40 items with a reliability of 0.70 to 50 items would likely increase it to around 0.74 based on statistical estimates for item count effects.
  17. Correct Answer: c. including homogeneous items
    Including homogeneous items improves reliability because they measure similar constructs consistently, reducing variability caused by diverse item types.