07.

17 PA Refresher Course Revised

Maximal Ability Test/Ability Test – overlap

Testing vs Assessment sometimes
 Psychological Testing – the process of  Intelligence Test
measuring psychology-related variables by o General potentiality (potential to
means of devices/procedures designed to solve problems, adapt to changing
obtain a sample of behavior circumstances, and profit from
 Psychological Assessment – the gathering experience)
and integration of psychology-related data o Concerned with past and future
for the purpose of making a psychological o Binet-Simon Test: only administered
evaluation that is accomplished using tools
in one-on-one basis; school children
such as tests, interviews, case studies,
 Aptitude test
behavioral observation, and specially
o Focuses on informal learning/life
designed apparatuses and measurement
experiences
procedures
o Measures potential to learn
o Evaluation (final decision); the one
o Concerned with the future
reflected in psychological report
 Achievement test
o Integration: present ONLY in
o Designed to measure
assessment
accomplishment
 Psychological Test – a device; to quantify
o given after
psychological variables; measures a construct
o Construct – indirectly observable, we o focuses on formal learning
believe that they exist and can be o only concerned with the past
measured o measures what has learned

Testing Assessment Interest test – dislikes/likes; does not tell whether

To obtain a score To answer a referral you will excel in a certain area
question
Diagnostic Test – used in screening, for diagnosis
Individual or by group Highly individualized
(clinical); for education, to identify
- Clinical setting
strengths/weaknesses
observing the individual
closely
Types of Psychological Tests
Highly objective (as is) Has some degree of
 Personality Test (or Typical-performance
subjectivity (or
tests)
maturity/clinical
o measure typical behavior (traits,
judgment)
temperaments, and dispositions)
Tester is NOT key to the Assessor is key to the
o related to the overt and covert
process process of selecting
tests and/or other tools dispositions of the individual
of evaluation and in  Structured (objective)
drawing conclusions o provides self-report statement
o choose between two or more
alternative responses
  Ex. NEO-Pi-r
 Projective – ambiguous test stimulus; Level C: require substantial understanding of testing
response requirements are unclear and supporting psychological fields together with
supervised experience in the use of these devices
Typical Ability Test – no right or wrong answer  RPsy
 Attitude  Unstructured personality test
 Interest  Individually administered intelligence tests
 Personality Flynn Effect
 raw scores mean on intelligence tests has
Army Alpha – dependent on the ability to read and been increasing over the years
write  Not observable in less advanced countries
Army Beta – illiterate adults  Present in highly developed countries
 Contained tasks such as mazes and picture Sampling Techniques
completion  Simple Random - Equal chance of being
selected
The war created a demand for large-scale group  Systematic – begins by listing all people in
testing the population, then randomly picking a
 Robert Yerkes – army requested for his starting point on the list. Sample is then
assistance obtained by moving down the list, selecting
every nth name.
Informed Consent  Stratified Random – identify the specific
Necessary for: subgroups (or strata) to be included in the
 Risks and benefits sample. Select equal-sized random samples
 How the obtained data would be stored from each of the pre-identified subgroups,
 Limits of confidentiality using the same steps as in simple random
sampling
For persons who are legally incapable of giving o merong different groups, and make
Informed Consent, psychologists nevertheless sure that each group is equally
 Provide an appropriate explanation represented
 Seek the individual’s assent o recognizes that there are subgroups
 Obtain appropriate permission from a legally o approach is randomized recruitment
authorized person  Cluster – random selection of groups instead
of individuals from a population; selecting
Consent is obtained except when groups NOT individuals
 Testing is mandated by law or governmental  Convenience – use participants who are easy
regulations to get
 Testing is conducted as a routine educational, o Weakest way of recruiting
institutional, or organizational activity respondents
 One purpose of the testing is to evaluate o Doesn’t recognize subgroups
decisional capacity o Quota
Levels of Psychological Tests  type of convenience
Level A: can adequately be administered, scores, and sampling; identify specific
interpreted with the aid of the manual and a general subgroups to be included in
orientation the sample and then
 Lowest level establishing quotas for
 No license required individuals to be sampled
 Ex. Teacher-made tests from each group
Level B: require some technical knowledge of the test  similar with stratified, but
construction and use and of supporting psychological manner of recruitment is
and educational fields convenient
 RPm Nuggets: Consider the profile of the target
 Structured personality and intelligence tests population; sample should represent the interest of
the population  Zero doesn’t mean absence
 Ex. Fahrenheit and Celsius
Assumptions of Psychological Testing 4. Ratio
1. psychological traits and states exist  Interval scale with the additional
2. psychological traits and states can be feature of an absolute zero point
quantified and measured  Ratios of numbers do reflect ratios of
3. test-related behavior predicts non-test- magnitude
related behavior  Ex. Number of students, pets,
4. tests and other measurement techniques amount of water in a container,
have strengths and weaknesses length, weight.
5. various sources of error are part of the  Zero means wala talaga
assessment process
6. testing and assessment can be conducted in a
fair and unbiased manner
7. testing and assessment benefit society
Results should not be interpreted in isolation
Isolation = not considering other factors
Levels of Measurement
1. Nominal Mean – can be distorted by extreme values (outliers)
 consists of a set of categories that Mode – most recurrent number in the distribution
have different names Median – not sensitive to outliers that’s why this is
 measurements label and categorize recommended to use as an alternative to mean
observations, but do not make any Relative Size Measures of Variability
quantitative distinctions between  Variance = SD2
observations  Range = highest score minus lowest score
 ex. Favorite color, brand of shoes,
gender, cellphone number, jersey Understanding the Standard Deviation
no., plate no. IQ Average = 100, SD = 15
 categories; do not have proper order,  100 + 15 = 115 (1 SD above)
pantay-pantay lang; may be  100 – 15 = 85 (1 SD below)
numerical
2. Ordinal Therefore, the average range for the IQ is 85 to 115.
 Organized in an ordered sequence
 Measurements rank observations in
terms of size or magnitude
 There’s a proper way of arranging the
observations
 Likert scale (per item): ordinal ; Likert
scale (total scores/overall): interval
 Hindi pantay-pantay yung gap
 Ex. Sizes of t-shirt, ranking in an
exam, top 3 singers, severity of
disease
3. Interval
 Intervals are of the same size
 Equal differences between numbers
on scale reflect equal differences in
magnitude
 Zero point on an interval scale is
arbitrary and does not indicate a zero
amount of the variable being
measured
Characteristics of the normal curve 1
 Mean, median, mode is in the center of the  Positive z-score is obtained when the raw
bell curve score is greater than the mean.
 Symmetrical. Left and right portion contains  Negative z-score is obtained when the raw
50% each. The entire curve covers 100% of score is smaller than the mean.
the distribution.  We cannot compare test takers using raw
 Most of the test takers’ score are close to scores; it doesn’t tell the whole story
average  Converted raw score to locate the score in a
 Low scores = close to the left tail distribution
 High scores = right tail  Location in a distribution
 It is not skewed.  For comparing average
The so-called normal range within which  Z-score = 0 = average
approximately two-thirds of all scores fall is located
within one standard deviation above and below the Z-Score Formula
mean.

Nuggets:

Nuggets:
 Ayaw sa z-score kasi may negative score.
Hence, misinterpretation.
Percentile Ranks – answers the question, “What
percent of the scores fall below a particular score?”
 99th percentile = person outscored 99% of
the test takers
- Ilang percent ng testtakers, ilan sa kanila ung  1st percentile = outscored 1% of the test
makaka-average? 68 out of 100 test takers. takers
- Rule:  50th percentile = outscored 50% of the test
o 1 SD = 68% takers
o 2SD = 95%
Percentile – ilang percent of test takers ung natala
Standard deviation – how the scores, in general, mo
scatter around the mean
 Higher SD = mas kalat yung score

Z-scores
 Direction and degree that any given raw
score deviates from the mean of a
distribution on a scale of standard deviation
units.
o Indicates how far an individual raw
score falls from the mean of a
distribution
 The z distribution has a mean of 0 and a SD of 0 z-score = 50th percentile
Z = 1.00 = 84th percentile international personal data sheet
Note: always compute the left area of the curve The Woodworth Psychoneurotic Inventory: for
personal psychological adjustment
Educational Assessment
 Formative Assessment: an assessment aimed
at facilitating learning and evaluating it.
 Summative Assessment: an assessment that
has a purely evaluative function
 Classical Test Theory: observed score = true
score plus error
 Error: nagpapalayo ng observed score from
the true score
 Higher reliability = lower error = mas malapit
ang observed score sa true score
Normal VS Skewed Distribution  Parallel forms: similarity in the results
obtained using two different forms of the test
o Concern is equivalence
o Computation of Pearson r for the
results obtained from both forms of
the same test
 Inter-rater reliability: consistency in the
ratings given by the two or more evaluators
o Dapat same ung diagnosis among all
raters
 Test-retest reliability: hindi gumagamit ng
Positively Skewed Distribution magkaibang form ng test
 Mean > Median > Mode o Practice effect – kaya tumataas yung
 the mean is greater than the median. Median score sa second take
is greater than the mode o Time sampling error
o Trait being measured does not
Negatively Skewed Distribution change over time (used in stable
 Mode > Median > Mean observations)
 The mode is greater than the median. o Traits, time, concern is stability
Median is greater than the mean o Establish stability over time kaya
Direction of the Relationship ginagamit lang sa mga traits that do
 Positive relationship = direct relationship not change over time
 Negative relationship = indirect relationship o Two administrations
Pearson-r vs Spearman’s Rho  Alternate test: reliability reduces practice
 Pearson-r – used in correlating variables in effect
interval and ratio scale o Content sampling error – reliability
 Spearman’s Rho – used when one or both of the test can be compromised or
variables are of ordinal scaling undermined if the items in the
 Both for correlation second form of the test do not match
Dependent t-test: comparing 2 observations from the items in the original form of the
the same group of respondents test; may discrepancy sa items in 2
Independent t-test: comparing data from two forms of test
different groups (for separate groups) KR 20
One-way ANOVA: deals with different groups  When computing internal reliability
ANOVA Repeated Measures: dealing with the same  used for dichotomous items
group  has right or wrong answers
ANOVA: deals with 3 or more groups/observations Coefficient alpha
Wechsler-Bellevue Intelligence test: adult  used for tests composed of items with no
 right or wrong answers.  Convergent – hahanap ng theoretically
Internal Consistency related with the one you’re measuring
 1 administration, 1 set of tests  Divergent – discriminant; 2 constructs na
 High if all items in a test measure the same magkaiba/walang kinalaman with one
construct another
 Cronbach’s alpha o Relationship should be weak
 MacDonald’s ω – good alternative compared ---End---
to Cronbach alpha; measure of internal
consistency
 Split-half reliability
Very high reliability = problematic = redundant
Domain sampling - sampling items from the
population of possible items that could be used in a
test

The greater the number of items = higher reliability,

kasi mas malawak yung coverage = tends to yiled
more reliable results

Face validity
 the assessment of the validity from the
perspective of the test taker
 simplest and least scientific
 demonstrated when the superficial
appearance of a measurement measures
what it is supposed to measure.

Content Validity
 items are examined if they are accurately
measure what supposed to measure
 ensuring inspection of test items
 educ setting; coverage of the test
 When test developers ask experts to rate the
items they included in a newly constructed
test as useful, useful but not essential or not
useful

Criterion Validity
 correlating with other measure; you stick
with the same variable
 Correlating test scores with the
scores/observations recorded using an
external measure.
3 types:
1. Concurrent - other measure to correlate with
is available in the present
2. Predictive
3. Known groups

Construct Validity: Correlating a newly created test

and another measure that quantifies a similar
construct