2 Measurement Errors Correct Answers - Random/chance error
- Systematic/constant error
2 Sampling Strategies Correct Answers 1. Nonprobability
sampling
2. Probability sampling
2 Types of Criterion-Related Validity Correct Answers -
Predictive Validity
- Concurrent Validity
3 Aspects of Reliability Correct Answers 1. Test-Retest
Reliability
2. Internal Consistency
3. Interrater/Interobserver Reliability
4 Kinds of Validity Correct Answers 1. Content Validity
-- Face Validity
2. Criterion-Related Validity
3. Construct Validity
4 Levels of Measurement Correct Answers - Nominal
- Ordinal
- Interval
- Ratio
Accessible Population Correct Answers One that meets study
criteria & is available
,Alpha Correct Answers - Researcher sets significance level, b/f
conducting tests
- Commonly set at 0.05 (5%) or 0.01 (1%)
- Used to determine whether to reject of accept
- Reflects probability of making mistake by rejecting null when
its true (Type 1 error)
- Significance level of 0.05 indicates risk or probability is less
than 5% that finding due to chance alone
Bar Chart Correct Answers - Nominal or ordinal level data
- Distinct or separate
Bivariate Statistics Correct Answers - Describe relationship
b/w variables
- rxy= correlation b/w x & y
Central Tendency Correct Answers - Mean is very sensitive to
extreme scores; can skew or distort findings
- Mean is middle in set of values; median is middle point in set
of cases
- Extreme scores don't affect median
- Mode is least precise
Central Tendency - When to Use Correct Answers - Mode:
nominal
- Median: ordinal, interval, or ratio (when have extreme scores)
- Mean: interval (don't incl. outliers)
Characteristics of Normal Distribution Correct Answers -
Mean, median, & mode approx. equal
- Symmetrical
, - Tails asymptotic
-- Scores that fall in middle are most likely to occur than those
in tails
- Fixed % of cases fall within certain distance from mean (1 or -
1; 2 or -2,...)
Cluster Sampling Correct Answers - Successive random
sampling of clusters that meet criteria
- Proceeds from large to small
Components of Power Analysis Correct Answers - Alpha level
- Effect size
- Power (0.8 is conventional standard; 20% risk of making type
2 error)
- Sample size
Concurrent Validity Correct Answers - Degree of correlation of
2 measures of same construct admin. at same time
- High correlation indicates high agreement
Confidence Interval Correct Answers - Estimated range of
values that provide measure of certainty about sample findings
- Most commonly reported in research is 95% or 99% degree of
certainty (95% CI: 95% of time, findings will fall within range
of values given as CI)
- Along w/ p-value, CIs used & reported to further interpret
findings
Construct Validity Correct Answers - Extent to which test
measures a theoretical construct or trait; complex approach