Tutorial 2 extra practice material based on Lec.2&3 Flashcards
(17 cards)
What is a nominal scale? And why do we use it?
Simplest level of measurement in statistics. It is used for categorizing data without any order or ranking.
Examples:
- Gender: male, female, non-binary
- Hair color: black, brown, blonde, red
- Marital status: single, married, divorced
- Types of fruit: apple, banana, orange
We use it to
- name or group data
- calculate the mode (most frequent category)
What is an ordinal scale? And why do we use it?
An ordinal scale is a level of measurement used to categorize and rank data, where the order matters, but the differences between values are not precisely measurable or equal.
Examples:
- Survey responses:
-Very dissatisfied, dissatisfied, neutral, satisfied, very satisfied
- Education level:
-High school, bachelor’s, master’s, PhD
- Pain level:
-None, mild, moderate, severe
- Class rankings:
-1st place, 2nd place, 3rd place
We use it to:
- Determine which value is higher or lower
- to calculate the median or mode (not the mean)
What is interval scale?
An interval scale is a level of measurement where:
- Values are ordered,
- The differences between values are meaningful and equal,
- But there is no true zero point, zero does not mean the absence of the quantity
Examples:
- Temperature in Celsius or Fahrenheit: The difference between 20°C and 30°C is the same as between 30°C and 40°C, but 0°C doesn’t mean “no temperature.”
- IQ scores: A person with an IQ of 120 is higher than someone with 100, but not necessarily “twice as smart.”
We use it to:
-Perform many statistical operations: mean, standard deviation, correlation etc.
What is a ratio scale?
A ratio scale is the highest and most informative level of measurement. It has all the characteristics of an interval scale, plus a true, absolute zero point — which means you can perform all arithmetic operations (add, subtract, multiply, divide).
Examples:
- Height (e.g., 180 cm is twice as tall as 90 cm)
- Weight (e.g., 0 kg means no weight at all)
- Age (e.g., someone who is 40 is twice as old as someone who is 20)
- Income (e.g., €0 means no income, and €60k is three times more than €20k)
- Distance, time, speed, volume, etc.
We can use it for:
- All statistical operations: mean, median, mode, standard deviation, variance, ratios and percentages
What are different names for 1-0 coded variables?
Dummy variables or binary variables
What is the interviewer error and what are possible solutions? (1)
Interviewers may not be giving the respondent the correct instructions.
- Provide very specific constructions on how the interview should be conducted
What is the omissions error and what are possible solutions? (3)
Respondents often fail to answer a single question or a section of the questionnaire, either deliberately or
inadvertently.
- Data imputation
- Remove observations:
1) You need to keep track of
data changes
2) you need to check whether
there is a “systematic” reason
why a single question is not
answered.
3) you need to justify why an
observation is removed - Contact the respondent again
What is the ambiguity error and what are possible solutions? (2)
A response might not be legible or it might be unclear (which of two boxes is checked in a multiple-response
system).
- Copy-edit and proof read surveys before they are sent out
- Conduct pre-tests
What is the Inconsistencies error and what are possible solutions? (3)
Sometimes two responses can be logically inconsistent. For example, a respondent who is a lawyer may have checked a box indicating that he or she
did not complete high school.
- Copy-edit and proof read surveys before they are sent out
- Conduct pre-tests
- Correct and edit the data (but, keep track of data changes and provide strong
justification)
What is the lack of cooperation error and what are possible solutions? (3)
In a long questionnaire with hundreds of attitude or image questions, a respondent might rebel and check the same response (in a agree-disagree
scale, for example) in a long list of questions.
- Make sure that surveys are not too long
- Have incentives for respondents to take
every single question seriously - Conduct pre-tests
What is the Ineligible respondent error and what are possible solutions? (2)
An inappropriate respondent may be included in the sample. For example, if a sample is supposed to include only women over 18, others should be excluded.
- Observations may be removed (keep track of changes and provide reasoning)
- Provide estimation results for the entire data set and the restricted data set (as a robustness check)
What are the most
common procedures for statistically adjusting the data in order to enhance its quality for data analysis?
- weighting
- variable specification
- scale transformation
What is weighting and when do we use it?
Weighting is a procedure by which each response in the database is assigned a number according to some prespecified rule. Most often, weighting is done to make the sample data more representative of a target population on specific characteristics. Categories
underrepresented in the sample are given higher weights, while overrepresented categories
are given lower weights. Weighting should be applied with caution, and the weighting procedure should be documented and made a part of the project report.
What is variable specification and when do we use it?
Variable specification: is a procedure in which the existing data are modified to create new variables, or in which a large number of variables are collapsed into fewer variables. The purpose of this procedure is to create variables that are consistent with the study’s objectives.
What is scale transformation and when do we use it?
Scale transformation involves the manipulation of scale values to ensure comparability with other scales. In the same study, different scales may be employed for measuring different variables. Therefore, it would not be meaningful to make comparisons across the
measurement scales for any respondent. Even if the same scale is employed for all the
variables, different respondents may use the scale differently. Some respondents may consistently use the lower end of a rating scale, whereas others may consistently use the upper end. These differences can be corrected by appropriately transforming the data.
What are the advantages (4) and disadvantages (5) of personal interviews?
Advantages:
* High flexibility
* Reactions to questions by respondents possible
* Questioning can be enhanced by visual aids
* Reduction of nonresponse rate
Disadvantages:
* Time consuming
* Administratively difficult
* Costly
* Need for qualified interviewers
* No anonymity which has an effect on socially desired answers and more personal
answers
Name some typical error sources of personal interviews (5)
- Over reporting: respondents state a more positive attitude than they actually have
- Interviewer bias: a partiality towards a preconceived response based on the structure, phrasing, or tenor of questions asked in the interviewing process.
- Bias because of question order
- Halo-effect: when the answer of one question influences the answers of other questions
- Socially desired answers because of non-anonymity
