Question

 

Use below terms to fill in blanks in the sentences that follow. Each term is used only once.

contingency table

control variable

correlated

cross-products

dependent variable

direction of association

expected values

gamma correlation coefficient

higher order gammas

independent variable


negative

Nominal


ordinal


positive


spurious association

statistical elaboration

statistically significant

strength of association

1. When conducting a two-way chi-square test to decide whether one variable probably influences a second variable, we compare the actual (observed) frequencies to the (1) ________________________, which are based on the pattern of responses for the overall group. When the actual (observed) values differ enough from these values, the difference between the two are determined to be (2) ______________________________. 

2. When we observe data for two variables arrayed in a 2 by 2 (3)_________________________________, we may suspect that the variables are (4) _________________________________ or related to each other if two cells on one diagonal contain a disproportionately large frequency of cases. Multiplying A*D and B*C produces two (5)________________________________, the larger of which reveals which diagonal is dominant. Multiplying A*D and B*C is the first part of the formula for computing the (6)_______________________________. 

3. When the frequencies in the A cell multiplied by the frequencies in the D cell are greater than the frequencies in B times the frequencies in C, the A-D diagonal is dominant and the direction of association is (7)_______________________. When the frequencies in the B cell multiplied by the frequencies in the C cell are greater than the frequencies in A times the frequencies in D, the B-C diagonal is dominant and the direction of association is (8)_______________________. 

4. The sign for a zero order gamma reveals the (9)_________________________________ and the absolute value tells us the (10)___________________________________. 

5. When we are studying hypothesized association between educational level and income level and we assume educational level (no college, graduated college) influences income level (low income, high income) educational level is the (11)__________________________ and income level is the 
(12)___________________________. Income level is an example of a 
variable measured at the (13)_________________ level of data. If we introduce the variable gender as a control, gender is measured at the (14)____________________ level of data.

6. The process of using a (15)___________________________ to test an apparent relationship between two variables measured at the nominal/ordinal level is referred to as (16)__________________________. 

7. When a control variable is introduced, two (17)_____________________________ are computed, one for each of the two sub-sets. For an apparent association to be reclassified as (18)______________________________, measured association in each sub-set must drop below an absolute value of 0.2. 


Answer 1