Chapter 1: Stats and Research Methods Review

Question 1

Variable

Answer 1

anything that can take on different values

Question 2

types of variables (levels of measurement)

Answer 2

categorical (binary, nominal, ordinal)
continuous (interval, ratio)

interval: equal intervals, no true 0 (F)
ratio: equal intervals, true 0

Question 3

measurement error

Answer 3

difference between observed and true value

Question 4

reliability & validity

Answer 4

Reliability refers to the consistency of a measure (whether the results can be reproduced under the same conditions). Validity refers to the accuracy of a measure (whether the results really do represent what they are supposed to measure)

Question 5

experimental vs correlational design

Answer 5

experimental: manipulate the IV to see affect on DV, can infer causation
- true experiment: random assignment
- quasi-experiment: no randomization, used in applied settings

correlational: no manipulation, used in applied settings. measures variable of interest to see if pattern of covariation is consisten with hypothesis/explanation. cannot infer causation

Question 6

methods of data collection

Answer 6

between-subjects: between groups, independent design, none of the same people in each group

within-subjects: within group, repeated measures design, same people in each group measured at different times, they are their own control

Question 7

types of variation

Answer 7

systematic variation: variation that can be explained by the model

unsystematic variation: variation not explained by the model

Question 8

randomization

Answer 8

random assignment: makes groups equal on extraneous variables, greatly improves causal inference

counterbalancing: eliminate order effects (practice, boredom), improves causal inference, determines which goes to 1st and 2nd condition first

Question 9

frequency distributions

Answer 9

normal, bimodal, rectangular, positively skewed (tail at high end), negatively skewed (tail at low end)

Question 10

4 qualities of a distribution

Answer 10

1. central tendency (representative values): mean, median, mode
2. dispersion (spread): range, variance, SD
3. skewness: positive, negative
4. kurtosis: more or fewer extreme scores relative to a distribution
   - platykurtic: low degree of peakedness (kurtosis<0)
   - mesokurtic: normal distribution (kurtosis = 0)
   - leptokurtic: high degree of peakedness (kurtosis >0)

Question 11

measures of central tendency

Answer 11

describe typical, average, representative scores
convey info about the center of a distribution
mean, median, mode

Question 12

Mode

Answer 12

score that occurs most frequently
useful descriptive statistic w/ nominal (categorical) variables

major/minor mode sometimes used

Question 13

median

Answer 13

score at the 50th percentile of a distribution
- for odd sets, its the middle number
- for even sets, its the average of the two middle numbers

Question 14

mean

Answer 14

average

Question 15

measures of dispersion

Answer 15

describes the spread of scores
are scores clumped together around the mean or are they dispersed?
range, variance, SD

Question 16

range

Answer 16

higest-lowest score
simple, sometimes useful for communicating spread
little value as an inferential statistic

Question 17

variance

s^2

Answer 17

quantifies the amount of dispersion of scores
useful as an inferential statistic
not easy to interpret by our minds
deviance, sum of squares (SS)

interpretation: On average, the number of FB friends people had different from the mean (95) by about 56.5 friends

Question 18

standard deviation

s

Answer 18

quantifies the amount of dispersion of scores
useful for our minds
it is the amount the scores vary from the mean on average

Question 19

z scores

Answer 19

qualtify the location of a value in a distribution in SD units (usually -3 to 3)
indexes an individual score’s place in a distribution in SD units as a ruler
z of 1 indicates the raw score is 1SD above the mean

z scores can be converted to percentiles if you believe the variable is normally distributed

z = X -Xbar / SD (one score - mean / sd)

Question 20

reporting data

Answer 20

• journals, conferences, books
• apa style
• choose mode of presentation that optimizes understanding of the data:
• 3 or less numbers: use a sentence. 4-20 numbers: use a table. 20+ numbers: use a graph