1
Q
What are the four sample types for surveying?
A
- Convenience sample (haphazard sample)
- Purposive sample
- Snowball sample
- Deviant case sample
2
Q
What is a convenience sample?
A
- Choose individuals who are readily available.
- e.g., surveying in a shopping mall
- Problems - often unrepresentative
- Biased: favors certain outcomes.
3
Q
What is a purposive sample?
A
- Chose based on their display of characteristic that we want in our study
- e.g., surveying families about work-life balance and purposively oversampling women
- often unrepresentative; biased: favors certain outcomes.
4
Q
What is a snowball sample?
A
- Ask initial participants to point out or help recruit other participants
- E.g., interviewing persons experiencing homelessness in SJ
- can often result in a very narrow and interconnected sample
- best when you are looking at a phenomenon in a tight knit group who has knowledge of one another
5
Q
What is a deviant case sample?
A
- sampling specially for extreme or counter cases for examination
- e.g., surveying people who have been charged with violating cannabis laws post-legalization.
- only useful in cases where we hope to understand an extreme.
6
Q
What are the four forms of probability sampling?
A
- Simple random sample (SRS)
- Systematic sample
- Stratified random sample
- Cluster sampling
7
Q
What is the simple random sample? (SRS)
A
- Sample chosen by chance
- Every member of the population has a chance to be chosen.
- uses a random procedure for choosing the sample
- there are apps and tables that can help with this too.
- e.g., understanding voting behaviors of UNB students: registrars’ list
- Need an accurate/complete list or risk under coverage
- Our sampling frame must not be biased (e.g., choosing a magazine subscription list that only older people buy)
8
Q
What is systematic sampling?
A
- The same as SRS but uses a random entry point with intervals
- interval = desired
- N = 5000
- Population size N = 50,000 = 50,000 every 10 people.
- understanding voting behaviors of UNB students: registrars’ list
- Potential that every case selected could be an odd case
- e.g., every 10th student could be different, but generally it is a good technique.
9
Q
What is a stratified random sample?
A
- Classifies individuals into groups of similar individuals (aka strata) and run a separate SRS for each strata, then combine results for your sample list.
- e.g., rural vs. urbanities voting preferences
- same as with SRS
10
Q
What is cluster sampling?
A
- sample in stages
- used when the sample frame
- e.g., all high school students in Canada are too large, yet somewhat specific
- Persons discharged from hospitals last month in Canada (firstly randomly sample hospitals, then randomly sample their lists)
- study becomes geographically concentrated or limited.
11
Q
Define Internal validity
A
- the confidence you can have in make cause-and-effect conclusions from the results of your study.
- 3 factors to check: IV & DV change together
- IV precedes DV change
- No other plausible explanations
- e.g., studying smoking and lung cancer –> internal validity depends on extent to which study can control for other factors that cause lung cancer.
12
Q
Define external validity
A
- The confidence you can have in generalizing the findings of your study to people, settings, and times not included in your study.
- consider study’s sample
- consider experimental setting
- e.g., studying new medication –> want to apply results beyond sample to people in population who need this medication.
13
Q
What are the factors that improve internal validity?
A
- random selection/sampling
- random assignment
- blinding
- having a strict protocol
14
Q
What are the factors that improve external validity? (RSR)
A
- Random selection/sampling
- Similarity between experimental & real-world
- Replication, replication, replication.
15
Q
What are the threats to internal validity? (AHME)
A
- Attrition
- Historical events
- Maturation
- Experimenter bias
16
Q
What are the threats to external validity?
A
- Sampling bias
- History
- Situational factors
17
Q
What are frequency counts?
A
They are used to graph and run frequencies counts in order to display the distribution of a variable.
18
Q
When do we use frequencies counts?
A
we use them for:
- numerical distributions –> what values the variable takes and how often it takes it.
- Categorical distributions –> provides the categories of the variable and the count and/or percent (aka frequency) of individuals that fall into each category.
19
Q
Name the four types of questions that are used to design surveys.
A
- direct questions
- indirect questions
- structuring questions
- Follow-up questions
20
Q
What is a direct question?
A
- ‘Do you find it easy to make an appointment with your doctor?
- often need to probe to elicit meaning ‘why’?
21
Q
What is an indirect question?
A
- ‘What do most people in your community think about access to health care?’
- it is best to follow these with a question about their views ‘is this what you think too?’
22
Q
What is a structuring question?
A
- Move on to a different topic.
- ‘I am now going to switch gears to ask you a bit about your experience with emergency department use…’
23
Q
What are follow-up questions?
A
- getting the interviewee to elaborate his/her answer.
- ‘could you tell me a bit more about that?’; ‘What do you mean by long wait time?’
24
Q
What is a 5 number summary?
A
- shows us a reasonably complete description of spread and center
- not the most common way to measure spread
- minimum Q1, Q3
- M, maximum
- We graph the five number summary in a boxplot
- used to compare distributions of different variables.
25
What are the quartiles?
- range: the smallest and the largest observation
- doesn't account for outliers
- quartiles: the three data points that divide your data into 3 equal groups
- First (Q1): Median of the observations to the left of the median (25%) - Lower
- Second (Q2): the median point in the data (50%) - median
- Third (Q3): median of the observations to the right of the median (75%) - upper
26
When do we use the mean, median, mode, and standard deviation?
- the mean and median indicate the 'center' of the data points.
- the mode is the value or values that occur most frequently.
- standard deviation measures how far the data 'deviates' from the center, on average.
27
How do we calculate the mean?
To find the mean for a particular variable, add up the observation values and divide by the # of observations.
- example: I have 10 students in my class.
- we collected how many minutes each student spent studying the night before. they responded: 10 + 0 + 40 + 20 + 10 + 20 + 0 + 20 + 40 + 0 = 160
- 160/10 = 16
28
How do we find the median?
M is the median symbol
M = (n+1)/2
Arrange all observations in order from the smallest to largest.
Find the locations by counting observations up from the bottom of the list.
0, 0, 0, 10, 10, 20, 20, 20, 40, 40.
if there are two medians you add them together and divide by 2
median = 15
29
How do we find the mode?
the mode is the most frequently occurring value.
30
How do you calculate the standard deviation?
- You have to find the variance
- the SD is the variance square rooted
31
Define the measure of central tendency.
- the mean
32
What are the descriptive statistics you can produce by level of measurements?
- Nominal --> Mode
- Ordinal --> Mode, median; Range, Quartile/five number summaries.
- Interval --> mode, median mean; Range, variance, standard deviation.
- Ratio - mode, median, mean; Range, variance, standard deviation.
33
What is the 68-95-98 rule?
- it is the rule for a normal distribution.
- 68% of our observations should be within 1 SD of the mean
- 95% should be within 2 SD
- 99% should be within 3 SD
34
What larger concept do the variance and standard deviation measure?
- how far away our observations are from the mean.
35
how to calculate the variance and standard deviation?
variance = calculate the mean
- subtract the mean and square the result for each data point to get differences/deviations (sum of deviations should = 0)
- calculate the average of those differences
- take the variance and square root it., that is the SD
36
What are bimodal and unimodal distributions?
- a camel and a normal bell curve
- bimodal = two bell curves
- unimodal = a normal bell curve
37
What is a normal distribution?
- each standard deviation will hold a certain proportion of cases (this applies for all normal curves)
38
What is a left skew?
left skew is when the left side extends out further than the right.
deviations to the pattern.
39
What is a right skew?
right skew is when the right side extends out further than the left.
40
How are they related to the mean and median?
the median is greater than the mean to be right skewed.
the median is less than the mean to be left skewed.
41
What are outliers?
an individual value that falls outside the overall pattern.
- was there an error in data entry
- can something explain this
example: life expectancy after diagnosis of dementia - average 8 yrs.
42
How do we represent missing data in SPSS?
- captured using 99 (missing)
- 98, 97, 96 (usually used for no answer, no opinion, refusal, etc.)
- reasons for missing data: missing at random (MAR): there is no statistical explanation for why the data may be missing.
missing not at random (MNAR): missing data is associated with other variables in your study (e.g., wealthy people may not want to report their income, women may not want to report their weight.)
Quant. Methods
flashcards
Decks in class (3)
# Cards
