Week 1 - Lecture 1 Flashcards
what are the 2 data types?
Quantitative (measurement)
- scores are meaningful
- e.g. scores 1-5 of anxiety
Qualitative
- can’t use a meaningful number system
- e.g. vote liberal, labour or green
what is a variable vs a construct?
- variables represent a construct
- e.g. the amount of words able to be unscrambled in 1 min could represent verbal ability or visual acuity
- e.g. response time (variable) could represent efficiency/inattention (construct)
what is a raw score and how do we know whether this score is good or bad?
- raw score is the value you scored
- we need to know the average & standard deviation to compare the raw score to and give it meaning
what is a sample vs a population?
- a **sample **can help *make inferences *about a **population **
- population:* all members of the group that you are interested in
- can only get population parameters when we have data for the entire population of interest –> cannot always get these parameters
Therefore, take a sample instead of testing whole population - subset of members from that population
- use that sample data to form inferences about* population parameters* we are interested in
what symbols represent population parameters vs sample data?
- greek symbols “mu” or “sigma” = population parameters
- “x-bar” or “s” = sample data
Why is the sample mean (X-bar) and population mean (mew) slightly different?
Estimates from sample are not perfectly accurate
**sampling error: **
- by chance, there is random variation between any sample and the population
sampling bias:
- due to the specific methods in a study, samples may tend to differ from the population in consistent ways
- avoidable through random sampling
- not always a problem, depending on the research question and the nature of the bias
- eg. some groups may be over or under represented
define distribution vs normal distribution
**distribution **
- graphical representation that associates a frequency/probability with each value of a variable
**normal distribution **
- bell shaped
- unimodal & symmetrical
- tails extend indefinitely
- area under the curve = 100% (everybody is included in the data
what type of distribution is required for parametic testing?
- parametric testing is assumed to have a normal distribution (dependent variable often assumed to be normally distributed)
- important because the probability of falling at different places along the distribution is known, which allows us to make inferences from our sample about population parameters
what is a standard score (z-score)?
- when you measure something, the scale is often arbitrary (e.g. weight)
- ## a z-transformation transforms a normal distribution into a standard normal distribution (puts everything on the same scale)
what is a z-transformation?
- if you convert your raw scores into z-scores and plot the z-scores on a frequency distribution, it’s called a standard normal distribution
- mean = 0
- SD = 1
- a z-transformation alters the mean and SD of a variable, but not the relative location of scores to each other –> visually the plotted scores look the same
what is the purpose of the standard normal distribution?
- allows comparison of performance across different tests/ different distributions
- tells you how many people score above or below you on a certain measure
- allows you to make inferences concerning the probability that different scores will be obtained
how to find areas under the curve
- “mew” and ‘sigma” can be used to calculate the probability that values will lie within a specified interval
- use the table of standard normal distributions (z-scores) to find these probabilities
what % of people fall within the IQ range: 85-115 (-1 to +1 SD)?
68.26%
Marshmallow experiment example question:
- let’s say you waited 125 seconds
- mew = 115 seconds
- sigma = 15
- you want to find out how long you waited to eat the marshmallow compared to other children your age
- convert 125 to z-score
- use tables to find area below your score
- look up the larger portion
purpose of comparing a sample with a population
- use sample data to make inferences about population parameters
- if nothing else is known, the statistics of a sample (e.g. the mean) are the best estimates of the population parameters (e.g. height of UQ students based on this class)
- but samples may fail to provide good estimates of the population due to sampling error and sampling bias
- compare a mean (X-bar) for the sample with the population mean (mew)
practise question: What percentage of people are within the IQ range of 100-130 (0 to +2 SD)?
A: 34%
B: 13.5%
C: 47%
D: 68.26%
- the total percentage of data within +-2 SD is 95.44%
- the percentage of data within -2 SD to ) is half of that: 95.44% / 2 = 47.72%
- since we only want the range from 0 to +2 SD, we take half of the data from 0 to +2 SD: 47.72%
Thus, the percentage of people within an IQ between 100 and 130 is 47.72%
(answer = c)
