Term
| field research includes 3 things |
|
Definition
| naturalistic observation, case studies, archival research |
|
|
Term
| field research has little or less |
|
Definition
| manipulation, control, random assignment |
|
|
Term
|
Definition
| higher external validity, improve generalizability, determine effects of events happening in real world |
|
|
Term
| quasi experimental field research |
|
Definition
| causal hypothesis, at least 2 levels of an IV, usually pre existing, specific procedures for testing, some control for threats to validity ex:double blind |
|
|
Term
| non-equivalent control group design in quasi experimental field research |
|
Definition
| pre existing control/experimental groups, groups can be made similar on important variables, pre tests to determine if groups are different |
|
|
Term
| interrupted time series design for quasi-experimental design field research |
|
Definition
| one group tested repeatedly, within subject design, helps see pattern of variability |
|
|
Term
| randomized control group design |
|
Definition
| uses randomization to select control group. good for protecting against confound |
|
|
Term
| pretest-posttest design program eval |
|
Definition
| no control goup, commonly used, not desirable, could be used for political motivation |
|
|
Term
|
Definition
status- describes characteristic of populatin
survey-gathers status + other info to make relationships between variables |
|
|
Term
| types of survey questions |
|
Definition
| demographics=factual, attitudes=content, abilities=intelligence |
|
|
Term
| format of survey questions |
|
Definition
| open, multiple choice, likert, dichotomous, physical analogue |
|
|
Term
|
Definition
| make non-normal distributions normal, for conceptual reasons, fits our model better |
|
|
Term
if piled up on the left then it's_____skewed?
piled up on right then it's_____skewed |
|
Definition
|
|
Term
| if right skewed which transformations can be used? |
|
Definition
| sq root of X, LogX, or the inverse 1/X |
|
|
Term
| if left skewed which transformations can be used? |
|
Definition
| square X^2, cubic X^3, Quartic X^4 |
|
|
Term
|
Definition
| data without order, Male or female |
|
|
Term
|
Definition
order is important but not evenly placed or equal interval
ex: not at all, disagree, agree, totally agree. another example of decent to completely awesome is more than one step |
|
|
Term
|
Definition
| distribution-free tests where the null hypothesis is about the location of the median. no assumptions made about the distributional shape. |
|
|
Term
|
Definition
equal interval but no absolute 0
ex: temperature |
|
|
Term
|
Definition
order, equal interval and absolute 0.
time or weight |
|
|
Term
|
Definition
| scales persons responding to question and those items on same scale, by definition equal interval |
|
|
Term
| b in item response theory stands for... |
|
Definition
|
|
Term
|
Definition
|
|
Term
|
Definition
| what define the scale by which b and theta are measured on |
|
|
Term
item response theory xs are____
number sideways lined up above Xs are____ |
|
Definition
|
|
Term
| assumptions of chi square |
|
Definition
| not normal dist, not homogeneous varaiances, independent cases, 1 response per case/person |
|
|
Term
|
Definition
for one way responses. one cell for each response
Hypothesis says there is a pattern |
|
|
Term
| independence Two-way chi sq |
|
Definition
| hypothesis says there is an association between.... |
|
|
Term
| how to get expected for one cell |
|
Definition
E= (R/N)(C)
R=cells row count
C= cells column count
N=total # of persons |
|
|
Term
|
Definition
|
|
Term
| how to calculate who contributed most |
|
Definition
standardized residuals
calculate R, treat like z score, compare to z table. |
|
|
Term
|
Definition
|
|
Term
|
Definition
| a measure of how influential a data part is. not R^2 |
|
|
Term
|
Definition
|
|
Term
| multiple regression used to... |
|
Definition
| used to see which of the IVs is a good predictor of the Dv |
|
|
Term
| Hierarchical method of multiple regression |
|
Definition
| planned, incremental. done in steps and planned by YOU |
|
|
Term
| stepwise method of multiple regression |
|
Definition
unplanned, incremental, computer selects best correlated IV with DV succesively
examine R^3 and its change |
|
|
Term
| 2 types of factor analysis |
|
Definition
| exploratory and confirmatory |
|
|
Term
|
Definition
| goal is to extract common variance of the variables with their factors |
|
|
Term
|
Definition
we know how many factors we want to extract, and we know what relationship the factors should have with the variables.
This type of analysis is also known as structural equation modelling, or latent variables modelling |
|
|
Term
|
Definition
| consistency in measurement |
|
|
Term
| Reliability measure: test-retest |
|
Definition
| measure same people twice on same scale, compute a correlation coefficient between the 2 |
|
|
Term
| reliability measure: split-half |
|
Definition
| split test in half and correlate the two. can get reliability estimate with 1 test administration but measuring the same trait through scale |
|
|
Term
| reliability measure: cronbach's alpha |
|
Definition
| avg of all possible ways of splitting a test. meaning the more items there are the more reliable a test. so if high on a study with few items it is meaningful |
|
|
Term
| reliability measure: item-total correlation |
|
Definition
| take item 1's score and correlate with sum of all others. if score high on this item and high on total test then the relationship should be high |
|
|
Term
|
Definition
|
|
Term
|
Definition
|
|
Term
| need to transform data when... |
|
Definition
skewness/stand error of skewness is not between -2 and 2
or kurtosis/SE of kurtosis isnt btwn -2 and 2 |
|
|
