Lecture 2 Flashcards
4 important things in research question?
- question mark at the end
- indicate all the constructs being investigated
- contain the population
- verb that indicate the type of relationship proposed (associate, predict, difference)
3 verbal forms of relationship?
- associate
- predict
- difference
in RQ:
“is there an association between studying research methods and happiness in psychology students?”
identify if this is a good RQ (use the 4 checklist for RQ requirement)
- question mark
- all constructs are listed explicitly: studying research methods and happiness
- population: psychology students
- verb: association
population vs sample
population: complete set of all individuals relevant to RQ. size: large. often unknown
sample: subset of individuals selected from population. size: known.
assumption in RQ regarding population & sample?
only one possible population and many possible samples drawn from that population
what is a construct?
a construct is an unobservable attribute of people that we use in both theories and research to explain human behaviour, cognition, affect
aka: attribute / theory
eg: need for cognition
what is a measure?
method for measuring people on a construct to obtain a score
aka: method
eg: need for cognition scale
what is a score?
numerical value on a construct measure assigned to an individual
aka: numerical value
eg: summed responses to items in need for cognition scale
what is raw observed scores?
values obtained directly from construct measure
often indicated by capital letter X, Y, etc
what is a deviation scores?
transformed value from raw observed scores
value obtained by subtracting mean score form each raw observed individual score
what does it mean when deviation score = 0?
means that raw score obtained is equal to mean score.
because deviation score is raw score subtracted by mean score
benefit of deviation score?
helps us to identify people below and above the mean.
characteristics of deviation score?
- mean of deviation score = 0
- normally lower case letter x or y
standard deviation = standard deviation of raw scores
what is z score?
standardised score whereby the deviation score is divided by the standard deviation
characteristics of deviation score?
- mean of z score = 0
- standard deviation = 1
- scaling is in standard deviation units
what is a standardised score?
more general version of z scores
value obtained by transforming raw score to have a PREDEFINED mean AND scaling for each unit of standard deviation.
mean and sd for standardised score are unknown. if mean = 0 and sd = 1, then it’s called z score.
hence, z score is a particular type of standardised score
eg: IQ scores (mean = 100, sd = 15)
what is continuous random variable
any numerical value within defined interval (eg: 0-100)
any possible real number within the range
what is discrete random variable
finite number of distinct values (eg: 1, 2, 3, 4, 5, 6)
typically indicated in integers. termed a ‘factor’ in R
also known as categorical variable (if it’s character by nature)
what kind of variable is measured in this RQ:
“is there an association between subjective well being and optimism in students enrolled in RMHI and ARMP in 2019?”
swb and optimism are CONTINUOUS VARIABLE
what kind of variable is measured in this RQ:
“is there an association between subject enrolment and type of student among enrolled students in RMHI and ARMP in 2019?”
subject enrolment and type of student are CATEGORICAL VARIABLE
what differentiates the analysis involving continuous and categorical variables?
(hint: correlation / contingency table)
continuous = correlation to answer RQ categorical = contingency table to answer RQ
summary characteristic
- our analysis to answer RQ produce summary characteristic
- a kind of aggregation on individual scores in one or more variables to produce a single quantity (eg average)
- eg: mean, sd
2 types of summary characteristics
population parameter and sample statistics
how to distinguish???
look at whom they are referring to (population or sample)
what is population parameter
aggregated summary characteristics involving ALL individual from a population
characteristic of population parameter
there is only ONE value of a population parameter
what is sample statistic
aggregated summary characteristics from single sample drawn from a population
characteristic of sample statistics
can have many sample statistics (one for each sample)
sample statistics vary from one to the next hence sample statistics are themselves random variables.
value of sample statistic is an estimate of the unknown population parameter.
2 kinds of distributions of scores
population distribution and sample distribution
characteristic of population distribution
there is only ONE
typically exist theoretically
characteristic of sample distribution
many, one for each sample
what is univariate distribution
distribution of one variable
what is bivariate distribution
distribution of 2 variables
correlation
measure of LINEAR SYMMETRIC association between two CONTINUOUS variables
most commonly used: pearson
will it make a different which variable is assigned to x axis or y axis in correlation?
no.
cor(X, Y) = cor(Y,X)
hence, measure of association would be the same regardless of the variables assigned to which axis
what does it mean to have positive association
high scores on variable 1, high scores on variable 2
what does it mean to have negative association
high scores on variable 1, low scores on variable 2
what does it mean to have no association
variables are uncorrelated
what are the characteristics of correlation
- absolute value of correlation indicates strength of association (ignore negative or positive)
- range is 0 - 1
- weak correlation: nearer to 0
- stronger correlation: nearer to 1
what is a correlation matrix
a table/matrix that indicates correlation of 2 or more variables. the rows indicate the set of variables column indicate the set of variables in the other variable.
what is correlation plot
table/matrix in which each individual cell refers to the correlation between 1 variable and another.
colour intensity signifies the strength (darker means stronger)
colour (red or blue) signifies direction (red is positive)
helpful to see pattern (instead of numbers)
population vs sample correlation coefficient
p (rho) for population correlation
r for sample correlation (vary from one sample to the other). used as estimate for population.
what is the effect of sample size to sampling distribution?
shape and spread of distribution (aka graph) depends on the sample size.
the bigger the sample size, the smaller the variability, less spread.
what is standard error
measure of variability of a sample statistic.
