Transcript Appraisal in Counseling

Appraisal in Counseling
Session 2
Schedule




Finish History
Statistical Concepts
Scales of measurement
Norm-referenced versus Criterionreferenced instruments




Measures of Central Tendency
The Normal Curve
Measures of Variability
Types of Scores
Statistics in Appraisal
1.
2.
3.
4.
5.
6.
7.
A language, communication, from one
professional to another.
How much confidence can I put in this
instrument’s scores?
Will I consistently get the same scores
every time I administer the instrument?
Do the scores really measure the concept
or construct I want to measure?
What can this instrument tell us? What
can it not tell us?
What is the meaning of the score(s)?
Compared to…?
How can scores be explained?
Assessment



Standards for Educational and
Psychological Testing – any method used to
measure characteristics of people,
programs or objects.
Anastasi & Urbina– objective, standardized
measure of a sample of behavior
Cronbach – systematic procedure for
observing behavior and describing it with
the aid of numerical scales or fixed
categories.
Samples of Behavior



Usually measuring constructs such as
intelligence, interest, etc. which cannot be
measured directly.
Since only samples are gathered must
determine if sample is how person usually
acts and are inferences we are making
correct.
Use objective, standardized and systematic
measures to gather samples.
Assessment Integral to
Counseling




Assessing client problems.
Conceptualizing and defining client
problems
Selecting and implementing effective
treatment
Evaluating counseling.
Benefits of Using Testing




Aids focusing on developmental issues
Aids problem solving
Helps decision making
Psychoeducational tools

Data indicate about 73% of counselors
responded tests very important or
important; only 9% said never used.
Minimum Competency for
Proper Test Use
Proper testing environment
 Examinee follows
directions so test scores
are accurate
 Avoid coaching or training
Scoring errors
 Avoid scoring & recording
errors
 Keep scoring keys & test
materials secure
 Know your instrument’s
psychometric properties
Counseling and
interpretation
 Establish rapport
 Know the instrument and
the construct completely
 Willing to interpret &
offer guidance
 Avoid labeling people on
basis of test scores
 Careful about assuming
norms for one group
apply to other groups
 Do your homework
Statistics

Scales of Measurement




Nominal
Ordinal
Interval
Ration
Nominal
categories, not ranked
 manipulating the numbers
arithmetically makes no
sense
 Example: Continent of Birth
1 Asia
2 Africa
3 Europe
4 Latin America
5 North America …
What is the arithmetic mean?

Continent of Birth
Continent
Frequency
Asia
Africa
Europe
Middle East
Latin America
North America
US America
Caribbean
New Zealand
Total
%
Cumulative %
151
22
31
8
7
1
7
5
1
64.8
9.4
13.3
3.4
3.0
.4
3.0
2.1
.4
64.8
74.2
87.6
91.0
94.0
94.4
97.4
99.6
100.0
233
100.0
100.0
Frequency Distribution
160
140
120
100
80
60
40
20
0
Asia
Africa
Europe
Middle East
Latin America
North America
US America
Caribbean
New Zealand
Ordinal




categories in rank order
manipulating the numbers arithmetically may not make
sense.
Example Opinions…Likert-like scale
Median, Mode, Range, Percentile. Not the arithmetic mean
Instructions: Please mark the number from the scale that best corresponds to your
answer.
1------------------------------2---------------------------3---------------------------4
Strongly
Disagree
Agree
Strongly
Disagree
Somewhat
Somewhat
Agree
I think of myself as being U.S. American.
1------------------------------2---------------------------3---------------------------4
Ordinal Data
US Ident 1 N=233
Code
1.00
2.00
2.50
3.00
4.00
Total
Missing
Total
Freq
130
56
1
33
12
232
1
233
Median 1.0000
Mode
1.00
Minimum 1.00
Maximum 4.00
Range
4.00 (Max-Min)+1
Percent
55.8
24.0
.4
14.2
5.2
Strongly
Strongly
Strongly
Strongly
Cumulative Percent
56.0
80.2
80.6
94.8
100.0
99.6
.4
100.0
Disagree
Disagree
Disagree
Agree
Interval/Ratio Data

Interval and ratio are similar




Interval




Ratio


continuous variable
can be represented numerically
Mean, median, mode, frequency, range, percentage,
percentile
may not make sense when we talk about ratio
for instance…IQ…..is someone who has an IQ of 50 half
as intelligent as someone with an IQ of 100? Or does a
zero IQ mean an absence of intelligence?
Examples: IQ or scores on an achievement test or the
GRE, MAT, SAT
continuous data with true zeros.
Height, weight, age
Norm referenced vs. Criterion
referenced
Measures of Central Tendency



Mode
Median
Mean
Measures of Central Tendency

Mode
 is the most frequent score
 can describe multimodal curves
 least stable of the measures because
it is not readily repeatable
 it is affected by the number of
people taking the test
 not affected by extreme scores
 NOIR
Measures of Central Tendency

Median
 middle score achieved
 50% of the scores are higher than this score
and 50% are lower than this score.
 If there is no middle, we create it.
 The median can tell us more about data which
is skewed versus data which is “normal”
 It is not effected by extreme scores
 No indication of multimodalness.
 IR
Measures of Central Tendency

Mean
 algebraic arithmetic average
 sum of the scores/number of observations
 it is affected by extreme scores
 most frequently used Measure because it
lends itself readily to more complex
manipulations.
 Used most extensively to describe normal
distributions.
 NOIR
Variability
Measures of Dispersion

Measure of the spread of scores

Range


Variance


The range is the highest score minus the lowest score
plus 1.
How widely spread are the scores from the mean
Standard Deviation

Takes the variance and standardizes it so that the
standard deviation is described in the same units as the
original scores.