Assessment Flashcards
Type II error
failure to reject the null hypothesis–you think that there is no relationship when there actually is
Type I error
rejected the null hypothesis when you shouldn’t have–you think there is a correlation when there isn’t
cluster sampling
unit is NOT the individual–but naturally occurring groups of individuals
-clusters are selected randomly for the experiment (classrooms selected for an experiment, naturally occurring, but randomly selected)
stratified sampling/proportional stratified sampling
randomly selecting the same proportion of individuals for the sample as they represent proportionally in the major subgroups in the population
quantitative non-experimental designs
- survey: response rate is low–cautious to generalize
- descriptive: describes an existing state of events
- comparative: investigates whether there are differences between two or more groups (no manipulation of conditions experienced by each group)
- correlational: uses the correlation coefficient to determine the degree of relationship between 2 or more variable
inductive research/deductive research
springs from theory—tries to determine what the relationships are between elements of the theory and may be experimental in nature
Sociometry
measures the structure and organization of social groups–requires revealing personal feelings about others
-sociogram: figure/map showing the interrelationships or structure of the group
Interests Tests
- Strong Interest Inventory
- Self-Directed Search
- Career Assessment Inventory
- Campbell Interests and Skills Survey
- O*Net Interest Profiler
Personality Projective Tests
- Rorschach
- Thematic Apperception Test (TAT)
- Rotter Incomplete Sentences Blank
- Draw a Person Test
Personality Inventory Tests
- Minnesota Multiphase Personality Inventory
- California Psychological Inventory (CPI)
- NEO Personality Inventory
- Beck Depression Inventory
- Myers-Briggs Type Indicator
Intelligence Tests
- Stanford-Benet Intelligence Scales
- Weschler Adult Intelligence Scales (WAIS-IV)
- Weschler Intelligence Scale for Children (WISC-IV)
- Cogitive Abilities Test
Ipsatively interpreted
comparing the results on the test within the individual
- comparing a first test to a second
- comparing scores within different parts of the tests
convergent validation/discriminant validation
occurs when there is no significant correlations between the construct under investigation and others
Standard Error of Measurement (SEM)
- confidence band/confidence limits
- helps determine the range within which an individuals test scores probably fall
- Score of 92, SEM=5.0
- 1 SD=67%=87-97
- 2 SD=95%=82-102
True and Error Variance
- want to measure true variance–the actual psychological trait or characteristic
- the amount of the variance measured in common is the correlation squared (.90^2=81%)(coefficient of determination)
-coefficient of non determination is the error variance (19%)
Kuder-Richardson formulas/Cronbach alpha coefficient
if it contains non dichotomous items (multiple choice)–cronbach alpha coefficient is applied
Internal Consistency
- split-half method–test divided into two halves–correlation between two halves are calculated
- may apply the Spearman-Brown formula to see how reliable it would have been if you didn’t split it in half
- measuring inter-item consistency–the more homogenous the items, the more reliable the test
Spearman-Brown formula
-rkk = k(r11) / [1 + (k – 1)* r11]
rkk = reliability of a test “k” times as long as the original test,
r11 = reliability of the original test(e.g. Cronbach’s Alpha),
k = factor by which the length of the test is changed.
To find k, divide the number of items on the original test by the number of items on the new test.
If you had 10 items on the original and 20 on the new,
k would be 20/10 = 2.
stability/equivalence
- test-retest reliability–using the same instrument on both occasions, looking at how the results correlate–two weeks is a good amount of time between testings
- alternate forms of the same test, are given and then scores are correlated
- intervening events and experiences might interfere with this
reliability
the extent to which a measure is free from error
- a correlation coefficient can be used
- if the reliability coefficient is high (.7), test scores have little error
correlation coefficient
- ranges from -1 to 1
- perfectly negative to perfectly positive
- a correlation between two variables is called bivariate–between two or more are called multivariate
- tells you NOTHING about cause and effect
t-scores
mean is 50
- SD is ten
- negative scores are eliminated
z-scores
mean is 0
SD is 1
range is -3 to 3(?)
-z-scores = zero is the mean
normal curve
scores are distributed into six equal parts (3 above the mean and 3 below the mean)
- 34% = 68% = 1 SD
- 13.5% = 95% = 2 SD
- 2% = 99% = 3 SD
Variance
the square of SD
standard deviation
the variability within a distribution of scores
-the mean of all the deviations from the mean
negative skew
- lump of scores are on the high end—outliers are on the low end of x axis
- (from left to right): mean median mode
Positive Skew
- lump of scores is on the low end–outliers are on the high end of the x-axis
- (from left to right): mode, median, mean
