-
range
a group of contiguous cells in a spreadsheet. Measure of dispersion equal to the highest score minus the lowest score, plus one
-
outliers
test takers who do much better or worse than most (more than three standard deviations above or below the mean.
-
Be able to name the measures of central tendency and measures of dispersion.
- Measures of Central Tendency: Mean, Median, and Mode.
- Measures of Dispersion: Standard Deviation, Variance, Semi-interquartile Range, and Range.
-
Know when we would expect to see a normal distribution (NRT, or certain CRTs), positive skewness, or negative skewness
- A Norm-Referenced Test (NRT) can be expected to have a relatively normal distribution with skewness and kurtosis relatively close to zero and the measures of central tendency close together.
- However, normal distribution is not expected in a Criterion-Referenced Test (CRT).
- A pre-test will have strongly positive skewness (with few high scores)
- while a post-test, given after students should have mastered the content, can be expected to have a negative skewness (with many high scores).
- The only times we should expect to see a normal distribution in a CRT are in cases where students at many different levels are taking the same test (e.g. in placement and proficiency testing). (Carr p. 231).
- POSITIVE = peak is on the left
- NEGATIVE = right peak.
-
correlation
the degree of relationship between two sets of numbers.
-
correlation coefficient
the mathematical estimate of the strength of relationship between two sets of numbers.
-
categorical/nominal variable
variables that can be grouped into categories, but the categories cannot be put in any sort of meaningful order(e.g. gender, native language, country of origin, and favorite color). All we can do is to count the number of people in each category.
-
ordinal variable
variables where different levels, ordered categories essentially, can be observed, and put into
-
interval variable
variables which are arranged in order, as with ordinal data, but are equal distances between each point on the scale.
-
continuous variable
a variable that can have any value within a particular range, with the range typically being fairly large.
-
ratio variable
similar to interval variables, and that there is an equal distance between each pair of points on a scale, the ratio variables also include a true zero point. Somewhat rare and language learning research and testing.
-
Pearson r
Pearson product moment correlation coefficient. Correlation coefficient used with two interval or ratio variables. Assumes variables are normally distributed
-
spearman R (rho)
correlation coefficient used in ordinal data, or for badly non-normal interval variables.
-
Know whether Pearson r or Spearman r would be more appropriate for correlating a particular set of variables (e.g., essay scores, scores on two long tests, test score rankings, scores on different item-based sections of a test, speaking test scores).
- earson R is used for correlations that are more linear and concrete; thus, it would be good for discrete-point, itemized tests.
- Since Spearman Rho is less linear focused (monotopic relationships), it would be best for subjective scores like essays and speaking test scores.
-
difference index
subtractive CRT discrimination index that requires giving the test twice, wants to Masters and wants to non Masters. Can be calculated as part of a differential group study or intervention study. As a rule of thumb, the D I should be. .40
-
item f
s a measure of association for two binary variables. Introduced by Pearson, this measure is similar to the Pearson correlation coefficient in its interpretation (Pearson R). In fact, a Pearson correlation coefficient estimated for two binary variables will return the phi coefficient.
-
differential groups study
approach to calculating the DI in which to existing groups are identified. One group is deemed to be at the mastery level in whatever ability is being assessed, and the other group is judged to still be non-masters
-
intervention study
essentially a pre-test/ post-test approach to calculating the DIi in which the test is administered to one group of students twice, before and after they have received instruction in whatever material is being assessed.
-
real-life domain
The actual, genuine TLU domain (e.g. a coffee shop, a doctor’s office, etc.)
-
language instructional domain
The classroom TLU were L2 learning is taking place, or is simulated to take place (e.g. a classroom in which students role-play a doctor’s visit, or practice reading a coffee-shop menu).
-
Commercial Test:
CELDT
- California English Language Development Test
- for K-12 students whose home language is not English
-
Commercial Test:
Oxford OPT
Oxford Online Placement Test
-
Commercial Test:
Cambridge English: First (FCE)
for proof of level. Recommended for level B2 on CEFR
-
Commercial Test:
Aptis
British-based English language proficiency test for businesses and organizations that assesses adults and teens A1 - C.
-
Commercial Test:
Cambridge English: Advanced (CAE)
proof of high-level achievement for level C1 on CEFR
-
Commercial Test:
Vantage Intellimetric/MyAccess!
- automated essay scoring used for academic, certification, or job screening purposes.
- MyAccess! Is an online tool from Vantage Intellimetric to help students with their writing and provides feedback
-
Commercial Test:
ACTFL OPI
- The American Council on the Teaching of Foreign Languages Oral Proficiency Interview
- assess how well a person speaks a language though a 20-30 minute one-on-one interview
|
|