PSYC/SOC - Research Methods Flashcards
Dependent variables depend on the _____ variable
independent
Operationalization
connecting your hypothesis with the operations you are going to do
Factor, explanatory variable, and predictor are all words for
independent variable
the explanatory variable is also known as the _____
independent variable
Categorical variables are categories. What two variables fit into this section?
Nominal and ordinal
Nominal
categorical variable that is a name
Ordinal variable
classes data in ordered categories but does not convey the degree or magnitude of difference between the categories.
what type of variable isa rate your pain on a scale of 1-10
ordinal
Interval variables
quantitative
Ratio variable
what we think of as a normal number
what type of variable is it when there is a fundamental zero?
Ratio variable
What type of a variable is it if there is an arbitrary zero?
interva
Why recruit and sort randomly?
To make it likely that differences will be spread out between the experimental group(s) and control group(s). People are weird so spread them out so that the weirdness spreads out
When a conclusion is based on a sample, it is called a _______. If the whole population were to be measured, that’s a parameter. Statistics are estimates of parameters
statistic
Parameter is a number describing a whole _____
population
a statistic is is a number describing a sample and is an estimate of the ____
parameter
cluster random sampling is usually used for ____
convenience
Stratified sampling sorts a population that behave ____
differently
Cluster random sampling sorts a population that behaves ___
similarily
what is non-probability sampling?
snowball
Descriptive statistics describes ____. It does not seek relationships within it.
data
What are potential problems with matched pair design or block design?
We may nit know that age and sex are important variables. Maybe we can’t pair exactly
A block design is when we ___ first, then randomize.
group. I.e male and female and then randomize
Unblinded is when ____ knows which participants are in which group
everyone
Blinding refers to who knows whether a participant is in the experimental or ____ group
control
Single blind is when the participant doesn’t know which group they’re in, but the ____ knows
researcher
Double blind is when neither the _____ or the ______ knows which group the participant is in. Only a third party (research coordinator) knows.
participant, researcher
What is the limitation of a matched pair design? It requires a knowledge of important ______ before the study begins
variables
Observational studies are when scientists observe variables without ____ them
manipulating
Observational studies are usually ______ studies, meaning they seek to correlate two or more variables.
correlation (move in the same direction, not necessarily cause and effect)
Cross-sectional studies study a sample at ____ point in time. Example: 5000 people’s LDL cholesterol and average daily step counts are testes. IS there a correlation between step count and LDL cholesterol
one. These are pretty easy to do. Your daily step count and LDL are linked but this is not establishing a cause and effect. Only correlated
Longitudinal studies study one sample over _____. Example: 500 people’s step counts and LDL concentrations are tested in the present time. Two years later, these are tested again. Are changes in step count and LDL cholesterol correlated?
time
Case-control studies study two groups: one that has a certain outcome (skin cancer) and one that does not. Then the values of another variable (sun exposure) are compared between groups. This is used for _____ outcomes. These are very ____ to do.
rare, difficult
Quasi-experimental studies or natural experiments are when an _______ is applied, but not applies randomly. Ex/ one city’s infrastructure changed to encourage walking while another stays constant. Does this change correlate with LDL cholesterol concentration of the residents.
interventions
Case studies dig deeply in a ____ cases.
few
A mixed methods approach combines multiple methods in a way that pains a more ____ picture.
complete
all normal distributions have the same properties. 1) they have a ____ shape and are ______. 2) the mean u(mu) is in the center of the distribution. 3) the area under the curve is __.
bell, symmetrical
1
the 50th percentile means that you beat 50% of people and 50% people beat you. Splitting into percentiles means to split into ___ pieces
100
quartiles allows us to split the data into four portions. the second quartile is the same as the ____
median
The interquartile range is Q3-___
Q3-Q1
For 5, 10, 11, 12, 15, 18, 20 the quartiles are. knowing Q2 is 12
Q1 is 10
Q2 is 12
Q3 is 18
The measures of centreal tendency are mean, median, and ___
mode
An ______ can dramatically skew the mean
outlier
when the data is left skewed the mean is furthest left then the median and then the mode. When the data is right skewed the mean is the furthest right and then the median and the mode is the least ____
impacted
The ___ gets pulled the hardest by the skew.
mean
Measures of dispersion are ___, interquartile range, variance/standard deviation
range
the ____ is calculated by Biggest-smallest
range
Interquartile range is calcualted by
Q3-Q1
variance is the standard deviation ____
squared
the boxplot provides a graphical display of the 5 number summary
min num, Q1 Q2 Q3 and max
the box does not show the mean it shows the ____ at Q2
median
High Accuracy is hitting the ____ of the board
middle
Precision is being ______.
consistent, so off target but at the same place
is accuracy or precision related to measure of central tendency?
accuracy
precision or accuracy concerned with tendency of dispersion
precision
dispersion is related to ___
precision
A systematic error is also known as a __
bias
if we dont know the Actual number we cant determine
____
accuracy
if we have a 95% CI and we want a 99% CI we need to _____ the range
expand
standard error is about the ______ of the distribution
spread
A low p value is unlikely to have occurred due to ___
chance
the hypothesis that it is just due to chance is the ____ _____
null hypothesis
the ______ hypothesis tells us something is going on
alternative
The null hypothesis is like saying nothing is going on. Change is due to __ alone
chance
a p-value of 0.03 means (for example) that there is a 3% probability that these observations (or more extreme ones) occurring due to ____
chance
If P is ___ than 0.05 then the results are significant
less than
if p-value is less than the alpha, the we reject
H0
Type 1 errors is a _____ positive
false
Type 1 errors (false positive) is rejected when it is actually ____
true
type 2 error is a false ____
negative
Type 2 error (false negative) if the null hypothesis is not rejected when it ___
should be
an alpha of 0.01 vs 0.05 is better because it has a ____ type 1 error
lower
when alpha 0.05 is used the probability of a type 1 error is a maximum of ___%
5%
how to remember type I and type II error?
type I you can make a P (FP)
type II you can make an N (FN)
We will only use a correlation coefficeint to compare two numbers
simple linear regression describes how one variable is associated with another and is an extension of ____
correlation
a lower Rsquared value means there is more spread
Residual is what is left over
observed-predicted
for r (correlation coefficient) we need to compare # and # and needs to be between -1 and 1
A chi squared test is ____
categorical
A test test is used to compare a number between two categories
i.e life expectancy different b/w Canadian and Americans
ANOVA is like a t-test more more than two groups
a number between two categories
Internal validity is for ____ and effect
cause and effect
external validity is the ability to do it in other situations
ethnographic studies are likely to have strong ___ validity
external
which study tends to have the highest internal validity?
experimental studies
standard deviation are show as error bars on a scatter plot
n
