Term
| 5 Steps to Data Processing |
|
Definition
1) Editing
2) Coding
3) Data Entry
4) Cleaning
5) Tabulation and Statistical Analysis |
|
|
Term
|
Definition
1) field edit
2) central office edit |
|
|
Term
|
Definition
check for interviewer/respondent mistakes
(Inspection/Correction) |
|
|
Term
| In coding, numbers are... |
|
Definition
1) associated to closed-ended questions
2) better kept to raw numbers than collapse into categories
3) mutually exclusive unless specified otherwise |
|
|
Term
| When coding you should define... |
|
Definition
- variable names
- value labels
- missing value codes
all of which are for entry into a computer |
|
|
Term
|
Definition
1) ID each respondent
2) use only numerical values
3) standardize codes for no answer (example: 8 = don't know; 9 = no answer) |
|
|
Term
| When data is put into software it is called a |
|
Definition
|
|
Term
|
Definition
|
|
Term
|
Definition
| Computer Assisted Telephone Interviewing |
|
|
Term
|
Definition
| Computer Assisted Personal Interview |
|
|
Term
| Other types of Automatic Data Entry |
|
Definition
- internet
- optical scanning |
|
|
Term
|
Definition
- generate frequencies for categories
- check to see if codes are within specified range (check for outliers) |
|
|
Term
| What is "simple tabulation?" |
|
Definition
| counting a single variable |
|
|
Term
|
Definition
1) Locate mistakes, outliers, and unusual observations
2) Determine degree of item non-response
3) Determine empirical distribution
4) Calculate summary stats |
|
|
Term
|
Definition
substitute missing values by:
regression
or
mean |
|
|
Term
|
Definition
| used for ordinal variables |
|
|
Term
|
Definition
| Study relationships between variables |
|
|
Term
|
Definition
| are either naturally nominal or ordinal, or they represent categorizations of continuous (ratio scale) measures |
|
|
Term
| lack of initial (2-way) cross tab relationship... |
|
Definition
| doesn't mean there isn't a 3-way relationship |
|
|
Term
| How to best present tabular data |
|
Definition
1) condense into one table
2) use no more than one decimal in percentages |
|
|
Term
| the greater the difference between expected and observed values... |
|
Definition
| the more we suspect the variables are related; the closer the difference to zero, the more we suspect the variables are independent (not related) |
|
|
Term
|
Definition
|
|
Term
|
Definition
|
|
Term
|
Definition
|
|
Term
| If you reject the Null, how do you calculate the strength of the effect? |
|
Definition
1) Phi Coefficient
2) Contingency Coefficient |
|
|
Term
|
Definition
- measures strength in 2x2 table
- it takes the value of zero when there is no association, which would also mean a chi-square value of zero
- when values are perfectly associated, phi equals 1 (observations fall on the main or major diagonal) |
|
|
Term
|
Definition
- table of any size
- varies between 0 and 1
- should only be used to compare tables of the same size |
|
|
Term
| Cross Tabs are good for... |
|
Definition
|
|
Term
|
Definition
(Also called categorical)
- 5 different genotypes, can be reordered and not effected
- mean/average is meaningless |
|
|
Term
|
Definition
- The values simply express an order
- The spacing between variables may not be the same (example 3 to 5 =/= 5 to 7) |
|
|
Term
|
Definition
| The distance between 2 values is meaningful (temperature) |
|
|
Term
|
Definition
| Has all the characteristics of interval, but it also has a clear zero. (height/weight) |
|
|
Term
|
Definition
|
|
Term
|
Definition
|
|
Term
|
Definition
- variable is normally distributed
- mean is known
- population variance is est from the sample |
|
|
Term
|
Definition
|
|
Term
|
Definition
|
|
