Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
48 Cards in this Set
- Front
- Back
|
Try and name 6 characteristics of quantitative data |
1. Measurement 2. Control 3. Causation and prediction 4. Objectivity 5. Reductionist 6. Replication of effects
|
|
|
Also...can you name 6 qualities of qualitative data? |
1. More about meaning 2. open-ended and exploratory 3. Anti-reductionist 4. in-depth interpretation 5. Subjective 6. Each study has its own value |
|
|
4 examples of quantitative approaches are...? |
Experiments, questionnaires, quantifiable observation, statistical analysis |
|
|
4 examples of qualitative approaches are...? |
Interviews, focus groups, textual analysis, descriptive observation |
|
|
What is a theory? |
A logically organised set of statements that explains observed phenomena and the relationship between them (describe relationship between phenomena, explain why they are related and the meaning, and predict research outcomes- that is, form a hypothesis) |
|
|
What are some opposite characteristics of scientific vs. non-scientific thinking, e.g. empirical vs. intuitive? Write down scientific first, and aim for 5 more pairs. |
skeptical vs. accepting systematic vs. casual observation objective/unbiased vs. subjective/biased accurate vs. ambiguous testable vs. untestable |
|
|
The characteristics of a good theory are: __________ (provides a clear account); _____________ (evidence can be found to falsify the theory; and _____________ (offers the simplest possible explanation for the phenomena). |
precision; testability; parsimony |
|
|
A hypothesis is a ______________ explanation of the relationships between variables, which usually translates into a ____________ as to the outcome of research. |
tentative; prediction |
|
|
Hypotheses will be untestable if they are imprecise. Each statement is an example of a different type of untestable hypothesis- can you name them all?
1) "People will be cheerful in warm weather" 2) "People with better memories will remember more words" 3) "People who are stroppy are under the spell of Satan" |
1) Inadequate definition of concepts i.e. what is 'cheerful'? What is 'warm'? 2) Circularity, i.e. better memory leads to remembering more words, because...better memory?? 3) Appeal to unscientific notions, i.e. religious ideas, which aren't empirically testable |
|
|
Causal or correlational?
There are 4 ways that A and B need to interact to say that A causes B. What are they? |
1) A needs to happen before B 2) Every time A happens, so does B 3) If A doesn't happen, neither does B 4 ) The theory logically supports that A would cause B
So, A is 'hearing Let It Go'. B is 'my daughter telling me to stop what I'm doing and watch her do her full Elsa routine'. She never does it until she notices the song is on, she does it EVERY DAMN TIME the song is on, she never does it without the song on, and the theory supports a causal relationship as she loves Frozen, and she's a showoff. Causal relationship! |
|
|
What is an operational definition? |
Some concepts are not directly observable, so an operational definition defines a concept in terms of how it can be measured, e.g. you can't measure 'aggression' generally, but you might choose body language indicators, or certain verbal cues, as operational definitions of 'aggression' |
|
|
What type of data represents involves naming or categorizing, i.e. each observation fits into a category and there is no way of distinguishing between 2 members of the same category? |
Nominal |
|
|
What is ordinal data? And what do you need to remember about analyzing it? |
Ordinal data provides info about the ranked position (that is, order) of observations, e.g. 1st, 2nd, 3rd in a marathon. It doesn't give information about the relationship between rankings, so in this example, we wouldn't know how far 2nd was behind 3rd. Note: it may not be useful to work out the mean of the raw data |
|
|
Interval and __________ data are very similar. __________ data has equally spaced data points, i.e. we know the difference between 10 and 20 is the same as from 20- 30. ___________ data is the same but has a 'true zero', e.g. mass or length |
ratio; interval; ratio
So remember a ratio data scale has a true zero! |
|
|
How can you remember the order of different data types, from the type that gives the least info to the type that gives the most? |
Nominal; Ordinal; Interval; Ratio- NOIR |
|
|
What is sampling bias? Give an example. |
This occurs when participants have been selected in a way that has a systematic effect on the outcome of the study, e.g. measuring attitudes to the 50% tax rate for high earners by speaking to people who pay that rate- likely views are negative?? |
|
|
There are 3 types of sampling that are all variations on 'random'. What are they and what do they mean? |
Simple- every member of a population has an equal chance of being selected Systematic- Selecting, for example, every 5th person on a list, or knocking on every 3rd door. The starting point must be randomised Stratified- selecting randomly but from within subgroups, e.g. if a workforce is 50% full-time and male, 25% full-time and female, and 25% part-time and female, then the sample should reflect this
|
|
|
What is cluster sampling? |
Selecting groups of participants to represent sub-groups in the population, e.g. if you were comparing history to psychology students, you might simply use a entire class of each |
|
|
Quota sampling? |
Often used is market research- similar to stratified sampling but the selection methods are not random |
|
|
Self-selecting? |
As seen in naturalistic observation studies, the participants fits criteria to be observed e.g. stopping at a traffic light. The researcher has no control over who becomes a participant |
|
|
Opportunity/convenience? |
Just whoever you can conveniently get e.g. psychology students. Very common sampling method!
|
|
|
Haphazard? |
Inviting participation from whomever is at a given location, e.g. a library or shopping centre. Also very common! |
|
|
Snowball? |
Participants recruiting others by word-of-mouth. Very useful with sensitive subjects or vulnerable groups. |
|
|
What can we measure though observation? 4 basic things...
Number of times something happens: ____________ How long something happens for:_____________ Time between events: _______________ How good/bad something is:_________________ |
frequency; duration; latency; ratings |
|
|
What is inter-observer reliability and why does it matter? |
It's the consistency between observations recorded by 2 or more observers- it needs to be checked to ensure operational definitions are being understood and applied correctly to observations |
|
|
What is observer bias and how does observer 'blindness' help? |
Observer bias is the name for systematic errors in observation when expectations cause observers to focus on particular behaviors. If observers are 'blind' to the hypothesis of the study this effect can be reduced or eliminated |
|
|
What do inferential statistics actually calculate?? |
If the null hypothesis is true, what is the probability of getting my results purely by chance? e.g. I say hamsters run faster than cheetahs. If actually cheetahs run faster than hamsters, what's the chance that my results are due to dumb luck? |
|
|
If the probability that our results are due to chance is less than ____ in ____, we conclude our hypothesis is __________. |
1; 20; supported
1 in 20 = 5%, which is .05
p < .05 means it's sufficiently statistically unlikely result is a fluke, so we consider our hypothesis supported. But this isn't proof and FFS don't write 'proof' on the exam anywhere, there's no such thing! |
|
|
Describe a Type I and Type II error. |
A Type I error is a false positive (stat test detecting an effect that isn't really there).
A Type II error is a false negative (stat test failing to detect an effect that actually does exist). |
|
|
Chi-square is a type of inferential stats analysis run on categorical data. There are 5 assumptions about the data that need to be met for chi-square to be useful; what are they? |
1) One or more categories 2) Independent observations 3) Sample size of at least 10 4) Random sampling 5) Expected frequency of at least 5 in each cell |
|
|
How should you report a p value reported on SPSS as .000? |
p cannot be zero, so should be reported as
p < .001 |
|
|
There are 7 sections t a quantitative research report- what are they (remember, you wrote one!)? |
1) Abstract 2) Introduction 3) Hypothesis 4) Method 5) Results 6) Discussion 7) References |
|
|
What is a rationale? |
It's an explanation/justification for the direction your research is taking, referencing and building on previous research |
|
|
Write out the chi-square standard reporting equation in words, so anyone could get the info from an SPSS table and write the equation correctly |
The equation is:
chi-square (df value, N = no. of valid cases) = value shown as Pearson's chi-square, rounded up, p = sig. value, rounded down
e.g. X2 (1, N = 188) = 6.27, p = .001 |
|
|
The mean and the median are both measures of central tendency. How do you calculate them? |
The mean: add up scores and divide by number of scores
The median: arrange scores in order of size and take the value in the middle |
|
|
If SPSS shows '95% confidence interval for mean, with an upper boundary of 20 and a lower boundary of 18...what does it mean? |
It means, for this sample, we are 95% sure that the true population mean lies between 18 and 20. Confidence intervals narrows as sample size increases
|
|
|
What does a normal distribution curve look like? |
You know the one, that bell-shaped thingummy |
|
|
Variance and standard deviation (SD) are measures of ______________. |
Dispersion, i.e. how spread out the scores are around the mean. Don't get worried about this- SD is usually reported, and it's the square root of the variance |
|
|
How would you work out the interquartile range of a data set? |
Put the scores in order, and divide in half. Take the mean of the bottom half and the mean of the top- the difference between them is the interquartile range.
Basically, you chop of the bottom and the top 25%. |
|
|
Skewness: how _______________ is the data set? Kurtosis: how ________ _____ is the data set? |
symmetrical; squashed up
I'd look at some examples if I were you. |
|
|
Types of kurtosis:
Platykurtic: ________ and ___________ Leptokurtic: ___________ and ____________ |
flat; rounded, tall; pointy |
|
|
If you have a box plot and one bit of data has gone loco and is miles away from everything, it may be considered to be ..... |
An outlier. But will you take it out? More to come... |
|
|
What's the case for removing outliers? |
Leaving them in can lead to error and impacts the mean; also inaccurate representation of the task |
|
|
What's the case for leaving outliers in? |
Removing it might reveal others (err nerrr!), effect in large sample is minimal, shouldn't discard data |
|
|
How do you decide about removing an outlier or not? |
Consider the nature of the task, and the possibility of Type I and II errors.
Also, does the outlier help you to reject the null hypothesis? If it does, taking it out makes it harder for you to support you own hypothesis- which is what you want, because empirical research = academic masochism (don't write that on the exam) |
|
|
Testing for normality: how likely is it that the data came from a normally distributed population?
1) look at the __________ 2) check _______ and _________ stats 3) Tests of __________ |
histogram; skewness and kurtosis; normality |
|
|
Tests of normality: if N < 50, you use ______ - ______. If N > 50, you use _________ - ________. |
Shapiro-Wilk; Kolmogorov-Smirnoff
If p < .05, the data are different to what you'd expect in a normally distributed population
These can be inaccurate, esp. with large samples, so common sense and judgement are needed (in which case I'm screwed) |
|
|
Z-scores are ___________and can be expressed as the number of standard __________ they are aware from the mean |
standardized; deviations
Z - scores larger than 2 (+ or -) are considered extreme- used in measures of depression, anxiety etc.) |
