9
(b) The reader responds to each item by writing a response or selecting an option
from a predetermined set of options. Thus, the reader’s RCA is transduced to
the responses to the items.
(c) A rater judges the correctness of each item response. (This may be carried out
automatically, especially in the case of multiple-choice items.)
(d) An initial indication of a reader’s RCA is then given by the pattern of correct
and incorrect responses for the set of items, which might be summarized in
terms of the number (or percentage) of test items that the reader answered correctly, typically called the “sum score”. The sum score is then an indication at
the macro level of the reader’s comprehension of the reading passage, whereas
each individual item response is an indication at the microlevel of the reader’s
comprehension of the question asked in the item.
(e) This indication is, of course, limited in its interpretation to just the specific set
of items administered. A variety of methods are available to allow interpretation
beyond that specific test to generate a value on an instrument-independent RCA
scale (some of which are presented below).
In a typical educational context, once the measurement is completed, a teacher
would use interpretational curriculum materials keyed to the RCA scale value to
assign the reader to some specific reading instruction activities designed to be
appropriate for his or her level of RCA.
If the process just described has been successful, the basic input of the process is
the student’s RCA, and the basic output is the estimated value of the student’s reading comprehension ability. As in the case of temperature, other inputs are usually
present that could contribute to the output, such as distracting noises, mood of the
student, peculiarities of the text passages and/or the questions, and even specific
background characteristics of the reader.
A traditional method to generate RCA scale values is the norm-referenced
approach, which relates the RCA values to the sum score distribution for a chosen
reference population of readers. In this method, a representative sample of individual readers from a specified population (e.g., 5-year-olds in a given country) take a
test, and this generates a sample of results in the form of local values (“sum scores”)
on the test. Then some statistics are computed on the sum scores, such as the mean
and the standard deviation (or the median and the interquartile range, or the percentiles), and the public reference properties are taken to be the RCAs of readers at
those values. For example, if the mean and standard deviation were the chosen reference points, then a scale could be set by mapping the mean to, say, 500, and the
standard deviation to, say, 100: thus, following this scale formulation, a scale value
of 600 RCA units would be for a reader located at one standard deviation (100 RCA
units)
8
above the mean (500 RCA units). This is thus an ordinal scale, but it is often
treated as an interval-level scale in psychosocial measurement. An alternative
approach to the norm-referenced approach is where the RCA scale values are related
8 More typically, these would be called “points”: i.e., “100 points above the mean”.
1.2 Some familiar and not-so-familiar contexts for measurement
(b) The reader responds to each item by writing a response or selecting an option
from a predetermined set of options. Thus, the reader’s RCA is transduced to
the responses to the items.
(c) A rater judges the correctness of each item response. (This may be carried out
automatically, especially in the case of multiple-choice items.)
(d) An initial indication of a reader’s RCA is then given by the pattern of correct
and incorrect responses for the set of items, which might be summarized in
terms of the number (or percentage) of test items that the reader answered correctly, typically called the “sum score”. The sum score is then an indication at
the macro level of the reader’s comprehension of the reading passage, whereas
each individual item response is an indication at the microlevel of the reader’s
comprehension of the question asked in the item.
(e) This indication is, of course, limited in its interpretation to just the specific set
of items administered. A variety of methods are available to allow interpretation
beyond that specific test to generate a value on an instrument-independent RCA
scale (some of which are presented below).
In a typical educational context, once the measurement is completed, a teacher
would use interpretational curriculum materials keyed to the RCA scale value to
assign the reader to some specific reading instruction activities designed to be
appropriate for his or her level of RCA.
If the process just described has been successful, the basic input of the process is
the student’s RCA, and the basic output is the estimated value of the student’s reading comprehension ability. As in the case of temperature, other inputs are usually
present that could contribute to the output, such as distracting noises, mood of the
student, peculiarities of the text passages and/or the questions, and even specific
background characteristics of the reader.
A traditional method to generate RCA scale values is the norm-referenced
approach, which relates the RCA values to the sum score distribution for a chosen
reference population of readers. In this method, a representative sample of individual readers from a specified population (e.g., 5-year-olds in a given country) take a
test, and this generates a sample of results in the form of local values (“sum scores”)
on the test. Then some statistics are computed on the sum scores, such as the mean
and the standard deviation (or the median and the interquartile range, or the percentiles), and the public reference properties are taken to be the RCAs of readers at
those values. For example, if the mean and standard deviation were the chosen reference points, then a scale could be set by mapping the mean to, say, 500, and the
standard deviation to, say, 100: thus, following this scale formulation, a scale value
of 600 RCA units would be for a reader located at one standard deviation (100 RCA
units)
8
above the mean (500 RCA units). This is thus an ordinal scale, but it is often
treated as an interval-level scale in psychosocial measurement. An alternative
approach to the norm-referenced approach is where the RCA scale values are related
8 More typically, these would be called “points”: i.e., “100 points above the mean”.
1.2 Some familiar and not-so-familiar contexts for measurement