Term
| What is the primary goal of factor analysis? |
|
Definition
| To identify a measurement model for a latent variable (a variable that can't be directly measured) like: being burned-out or intelligence |
|
|
Term
Describe how researchers use factor analysis for evaluating research measures
(3 things) |
|
Definition
1) to understand the structure of a set of variables
2) to construct a questionnaire to measure a latent variable
3) to reduce a data set to a more manageable size while keeping as much of the original info as possible |
|
|
Term
| Explain what is meant by “factor extraction” |
|
Definition
| Factor extraction is the # of factors that you are going to keep in your research. You keep factors based on the scree plot and also with Eigen values greater than 1 |
|
|
Term
What role does a correlation matrix play in extraction phase?
(3 things) |
|
Definition
1) The martrix represents the relationship between variables
2) The linear parts (factors) of the martrix are calculated by figuring out the eigenvalues
3) The eigenvalues then determine which variables load onto which factors |
|
|
Term
| Explain the meaning of an “eigenvalue” and its relation to % of Variance. |
|
Definition
It is an index that represents the extent to which a factor explains variance across the variables
Meaningful factors will have eigenvalues greater than 1 and will explain most but not all of the variance in the data |
|
|
Term
| Explain the logic behind the “eigenvalues greater than 1” rule |
|
Definition
"Kaiser" came up w/ this method.
The "greater than 1" logic stemmed from the idea that eigenvalues represent the amount of variation explained by a factor and "1" represents a large amount of variation in data |
|
|
Term
| How do you interpret a scree plot to help judge the number of factors to keep? |
|
Definition
Cattell came up w/ this method
y-axis is the eigenvalue and x-axis is the factor
The shape of the line will be curved w/ a long tail
Typically, you will see a few high eigenvalues at the top (significant factors that you should keep) and then a trailing-off of the other factors
[image] |
|
|
Term
| What is the goal/purpose of “factor rotation?” |
|
Definition
| Goal: to clarify the interpretability of the the factors (once you rotate factors, you will usually be able to discriminate between factors) -You look at the clusters to see which variables are being loaded onto which factors |
|
|
Term
| How to interpret 2-dimensional factor plots? |
|
Definition
-You will look at the clusters that have the most variables loaded onto the factors
-Then you will label the factors based on the pattern of common qualities (the researcher will have to make a judgement on the pattern) |
|
|
Term
| How do you interpret the value of a “factor loading?” |
|
Definition
Use the critical value table to see if the value is significant (a loading aka cluster of an absolute value greater than 0.3 is significant)
Stevens (2002) created this critical value table
You can also use R2 |
|
|
Term
| How do you use the table of factor loadings to help define the meaning of the factors? |
|
Definition
-You look at sample size and loading data based on an alpha of .01 and 2-tailed
1) a sample size of 50 with a loading of .722 is significant
2) a sample size of 100 with a loading of .512 or greater is significant
ETC. |
|
|
Term
| How does an “oblique” rotation differ from a more standard “orthogonal” (varimax) solution? |
|
Definition
1) Oblique rotation: the factors are allowed to correlate and it should only be used if there are good reasons that the factors could be related
2) Orthogonal (Varimax) rotation: factors are rotated while keeping them independent from one another (unrelated), this method makes sure the factors are uncorrelated |
|
|
Term
| Explain how consideration of “theoretical meaningfulness” aids in factor interpretation, and how it can influence the decision on the number of factors that underlie the variables |
|
Definition
The researcher must choose which rotation method he/she will used based on their own judgement (whether or not you think that the underlying factors should be related)
-If you expect the factors to be independent, you would use varimax
-If you expect the factors will correlate, you would use oblique aka oblimin |
|
|
Term
| Explain why precision in measurement is important to psychological research |
|
Definition
precision is
usually estimated using a standard error, that is, the amount of chance fluctuation (or lack of
precision) we can expect in sample estimates. We can use the standard error as an estimate of the precision of a statistic in two ways: descriptively or inferentially |
|
|
Term
| Distinguish between a psychological construct and an operational definition of the construct |
|
Definition
Construct refers to any complex psychological concept. Examples would be a person's motivation, anger, personality, intelligence, love, attachment, or fear. A Construct's height, weight or depth cannot be measured because constructs are not concrete materials in the visible world.
An operational definition is a clear, concise definition of a construct so that it can be measured more easily
|
|
|
Term
| What does it mean to say a measure is valid? |
|
Definition
Validity: established by collecting studies and evaluating them to see how the scores relate to your topic (are the variables correctly defined in relation to your DV?)
-A valid measure means you are measuring the RIGHT thing |
|
|
Term
| What do deficiency and irrelevance mean? |
|
Definition
-Deficiency: stuff your leaving out of your study that should be there
-Irrelevance: (1) Contamination, aka measuring the wrong things, outliers, observer bias, and (2) Measurement error, randomness |
|
|
Term
|
Definition
X = T + e
X = observed "raw" score on the measure
T = the person's true score on the measure
e = measurement error
Note: if X = T + e, then e = X - T |
|
|
Term
Common Sources of Error
(4 things) |
|
Definition
1) Test-Taker (fatigue, attention, etc.)
2) Examiner/Researcher (clarity of instructions, etc.)
3) Situation/Test environment (noise, lighting, distractions, etc.)
4) Item idiosyncracies (language, clarity, etc.) |
|
|
Term
Types of error
(4 things) |
|
Definition
Test-retest rxx:
e = instability over time
Alternate forms rxx:
e = inconsistency across different item sets
Internal consistency rxx:
e = inconsistency across different individual items
Inter-rater rxx:
e = inconsistency across different raters/observers
|
|
|
Term
|
Definition
Most common measure of reliability, a measure based on the correlations between different items on the same test (or the same subscale on a larger test). It measures whether several items that propose to measure the same general construct produce similar scores
(Scores from 0 to 1, closer to 1 means higher reliability) |
|
|
Term
| Pearson r between items for consistency |
|
Definition
One beneficial feature of the Pearson correlation coefficient is that the scores on the rating scale can be continuous in nature (e.g., they can take on partial values such as 1.5) can only be computed for one item at a time
|
|
|
Term
| Cronbach's Alpha (aka coefficient alpha) |
|
Definition
rxx= K(rij)
-----------------------
1+(K-1)rij
rij= Average r of all the items in the test
K= # of items in the test
|
|
|
Term
| standardized coefficient alpha |
|
Definition
k rbar
---------------------
1+(k-1) rbar |
|
|
Term
|
Definition
-Compares 2 or more groups
-is a test of the hypothesis that the variation in an experiment is no greater than that due to normal variation of individuals' characteristics and error in their measurement |
|
|
Term
| Between Groups and Within Groups Variance |
|
Definition
Between group variability and within group variability are both components of the total
variability in the combined distributions. What we are doing when we compute between
and within variability is to partition the total variability into the between and within
components. So: Between variability + within variability = total variability |
|
|
Term
| 3 Assumptions of Oneway ANOVA |
|
Definition
-Independence
-Normality
-Homogeneity |
|
|
Term
|
Definition
| is used to test if k samples have equal variances. Equal variances across samples is called homogeneity of variance |
|
|
Term
| Levene's Test Null Hypothesis |
|
Definition
| H0: σ12 = σ22 = ... = σk2 |
|
|
Term
| Fmax aka ratio of group variances |
|
Definition
-Assumes normally distributed
-Fmax is used in the analysis of variance to verify that different groups have a similar variance, an assumption needed for other statistical tests.
-largest variance/smallest variance; worry if it’s a 4; you have to worry if it’s a 10 |
|
|
Term
|
Definition
-Helps to evaluate the normality assumption
-Normal distribution will have 0 skewness and 0 kurtosis
-Negative kurtosis= platykertic
-Positive kurtosis= leptokurtic |
|
|
Term
|
Definition
- Eta: Eta indicates the ability of a predictor, using the given categories, to explain variation in the dependent variable.
-Eta square: Eta2 is the correlation ratio and indicates the proportion of the total sum of squares, explained by the predictor |
|
|
Term
|
Definition
| It tests 'overall' significance of the model. It tells you that among your means, some comparison/condition differs significantly from each other. However, it doesn't tell you which exactly. This is why we conduct Post Hoc tests after we obtain a significant omnibus, to determine where those meaningful differences are exactly. |
|
|
Term
|
Definition
-A common solution for the alpha inflation problem
-If omnibus test is significant, you are safe
-Multiple t-tests can be done |
|
|
Term
|
Definition
-uses the studentized range distribution
|
|
|
Term
|
Definition
| a special case of regression where the IV's are a mix of continuous and categorical variables |
|
|
Term
|
Definition
-Reduction in the error term (within group variance) for the ANOVA (partialling out factors that increase error)
-Elimination of confounds, especially when random assignment can't be used
-Test for hypothesized mediating factors |
|
|
