Please enable JavaScript.
Coggle requires JavaScript to display documents.
quant W5-7 research skills pt1 - Coggle Diagram
quant W5-7 research skills pt1
levels of measurement
we distinquish between quant variables by their levels of measurement
nominal data
refer to categorical data
gender, ethniticity. job type
numbers are given to distinquish between categories. no order to rank importance
interval data
puts scores in order, however the differences between the numbers are equal
temperature- centigrade or fahrenheit= difference between 0-10 is the same as 10-20
no absolute 0 where variable being measured doesnt exist
ratio data
the same features as interval datsa
difference between numbers are equal but there is an absolute zero
height- cant have height less than 0
descriptive statistics
used to describe your sample
summaries data using number of graphs, a value that represents a whole distribution
measure of central tendency (averages)
mean
sum of all scores divided by the number of scores in the sample, most commonly reported
pros- ease of calculation, a good estimate of population mean, ideal basis for inferential statistics
cons- sensitive to extreme scores, can tbe used for nominal data
gives an idea of the typical or middle values in a set of data (most common form of descriptive statistics)
median
middle score, less commonly reported than the mean
pro- not sensitive to extreme scores, only requires ordinal level data
cons- cant be used for nominal data, not ideal as a basis for inferential stats
mode
used for categorical variables, most frequently occurring score, bimodel= 2 modes, multimodal= several modes
pros- cant be used with any type of data
cons- may not represent central tendency at all (if distribution is skewed), not ideal as a basis for inferential stats
population mean and sampling error
population mean and sampling error
a typical score in a population
sampling error- the difference between the sample statistic and the population stats
graphical description of data
a way of summarising your data visually
bar chart
used to summarise a categorical variable
x axis- categorical variable. y-axis frequency average
histograms
type of bar chart for continuous variables- bars not separated and equal, all values representated even if empty
used to graphically illistrate your whole dataset to a reader
x axis: details of score on our variable, Y axis: f=frequency of occurence of scores
large sets of data from a distribution
normal distributions of histograms
key features: peak in the middle, tails of symmetrically , bell shaped curve
measures of variability
how to spread out your scores are- how much variation is there?
the range
the distance between the lowest and highest score in a sample
subtract bottom value from top value
problem- sensitive to outliers
interquartile range
distance between the upper and lower quartile in a set of data
appropriate for ordinal level data and non normal data
less affected by outliers/ extreme scores
standard deviation
an estimate of the average deviation of the scores from the mean
appropriate for interval and ratio level data
2 ways to calculate:
corrected used to estimate population SD
SPSS uses this formula
the mean deviation
the mean of all absolute deviation values in a data set
what does SD tell us- an indicator of how closely scores are clustered around the mean
advantages- used in population parameter estimates, take exact account of all values in the data set and are the most sensitive of the dispersion measures
disadvantages- they are sensitive to extreme values
hypothesis
alternate hypothesis
the population means from the 2 groups/ conditons are not equal
there is a difference/effect/relationship between the variables we are studying
hypothesis phrasing ex- there will be a difference between emotion recognition than people without a diagnosis of depression
one tailed hypothesis
specificed the direction of relationship/difference between conditions, directional hypothesis
directional hypothesis
suggests the direction of the effect (tails)
non-directional hypothesis
doesnt specify the direction of the difference/effect
aim is to confirm/disprove the null hypothesis rather than finding evidence that support the alternative hypothesis
null hypothesis- state there is no difference/effect/relationship between the variables we are investigating
empirically testable proposition about facts, behaviour, relationship, usually based on theory, states an expected outcome resulting from specific conditions or assumptions
scientific hypothesis- falsifiable, testable, precise
two-tailed hypothesis
you have predicted that there will be a relationship/difference between variables but you have not predicted the direction of that relationship/difference
non directional or bi-directional hypothesis
issues standard error
small samples have larder confidence intervals
larger samples have narrower confidence intervals
larger the sample size the better the estimate of the population
population mean- mean in the population
