Research Methods Flashcards
What is nominal data?
Data that can be categorised
What is interval data?
Data that can be measure on equal interval scales e.g. height because it’s continuous
What is ordinal data?
Data that can be placed into rank order e.g. in a race
Steps of spearman’s ranks
- Data is categorised into a table
- Data is then ranked the lowest given a rank of 1 and so on (each column ranked individually). If there is a tied score the rank scores are added and divided (e.g. 10= 3+4 / 2 = 3.5)
- 4.+5. Complete the difference (d) between the ranks. Rank 1 - Rank 2 and then squared (d2) then all the values are added together to give the sum Ed2(squared)
- Rs = 1 -6Ed/n(n2-1)
- Level of significance - this requires looking at the critical values of the spearman’s rank test in the critical value table for this test
What is a one- tailed test, what is a two- tailed test?
One tailed is a directional hypothesis
Two tailed is a non- directional hypothesis
What is inferential testing?
Ways to verify your data
Can use chi square, sign test, Mann Whitney u test, the unrelated t test, the related t test, Wilcoxon matched pairs, spearman’s rank or Pearson’s rho
Steps to complete the sign test
- Data is categorised into a table of results
- Positive and negative signs need to be added. If conditions A is a yes and condition B is a no a plus is added and the opposite would be added.
- Count up each positive and negative sign assigned to each participants scores.
- The smallest of the total direction scores is the overall t test result
- Level of significance - this requires looking at the critical values of the sign test in the values tables.
Steps to complete the Chi square test
- Label 1,2,3+4 on the observed and expected values
- Calculate the degrees of freedom and expected values: degrees of freedom = (no. Of rows - 1) x (no. Of columns - 1)
Expected values (e) = row total x column total / N - Fill out the rest of the table ( e, o-e, (o-e)2, (o-e)2/e )
- Add the last column together to get E (the sum)
- Look at the table of values to determine whether your information is valid or not
How do you work out the expected values in the chi square test?
Row total x column total/ n
What is the difference between a natural and a quasi experiment?
Natural- naturally occurring independent variable and a natural setting.
Quasi- naturally occurring independent variable in a lab setting (non-natural)
What is a type I error?
Allow yourself to find a correlation where there isn’t one. Usually when the researcher uses 5% level of significance.
What is a type II error?
When you reject your hypothesis because you have been too harsh on yourself e.g. you used 1% significance instead of 5% where a relationship could be seen.
What is the difference between parametric and non parametric data?
Parametric data can fit in a graph (It’s interval or ratio) whereas non- parametric data can’t fit in a graph.
If you are looking at a difference between variables, have nominal data and are using an independent group design what inferential test should you use?
Chi square
If you are looking at a relationship between variables, are using a correlation design, and have ordinal data what inferential test should you use?
Spearman’s rank
If you are looking at an association between variables and have nominal data what inferential test should you use?
Chi square
If you are looking at a difference between variables, have nominal data and use a repeated measures design what test should you use?
Sign test
If you are looking at a difference between variables, have ordinal/interval data, use an independent group design, and is non parametric what inferential test should you use?
Mann Whitney U test
If you are looking at a difference between variables there is ordinal/ interval data, independent groups design and the data is non- parametric what inferential test should you use?
Unrelated t-test
If you are looking at a difference between variables, have ordinal/ interval data, use a repeated measures design and the data is parametric what inferential test should you use?
Related t test
If you are looking at a difference between variables have ordinal/ interval data, use repeated measures design and the data is non- parametric what inferential test should you use?
Wilcoxon matched pairs
If you are looking at a relationship between variables, use a correlational design, and have interval/ ratio data what inferential test should you use?
Pearson’s roe
What is alpha bias?
When the difference between the variables is exaggerated (e.g. Bowlbys nonproprietary theory)
What is beta bias?
Beta bias is when the difference between the variables is minimised (usually happens when studies generalise results from one gender to the whole population)
What is cultural relativism?
How relative somethings is to a certain culture
What is ethnocentrism?
Using one culture and generalising it over other cultures.
What is cultural bias?
Stereotyping a certain culture
What is emic research?
- culture specific
- potentially dangerous for exaggerating differences between cultural groups
- reductionist as research fails to account for cultural sub-groups (gender sexuality etc)
What is etic research?
- research from a specific culture is applied universally
- studies take samples from a single population and generalise the findings to all
- uses a between -group focus on research to study variations in behaviour
What did Bem say? What year?
Androgyny is desirable
1974
This research is ethnocentric
What did Brislin say? What year?
- intelligence testing in western cultures often involves completing tasks with a limited time
- ‘mental quickness’ is valuable in western culture
This research is etic
1976
What did Wober say?
People in Uganda categorise intelligence as a slow process of careful and deliberate thought
1974
This research is emic
What did Cochrane et al find?
That African-Caribbean population in the uk were 7 times more likely to be diagnosed with schizophrenia
What did Littlewood et al find out?
That the African-Caribbean patients were often prescribed stronger doses of medication than white patients whist exhibiting the same symptoms
What is free will?
The notion that humans can makes choices not based on biological factors
What is a content analysis?
A research tool used to indirectly observe the presence of certain words, images or concepts within the media (e.g. advertisements, books, films etc). For example content analysis could be used to study sex-role stereotyping usually carried out on secondary data.
How to conduct a content analysis
1- familiarise yourself with your data 2- generalise initial codes 3- searching for themes 4- reviewing themes 5- defining and naming themes 6- producing the report
What is a thematic analysis?
Brown and Clarke in 2006
- looks for the emergent themes in data, codes them and then interprets their meaning
How to check reliability
Inter- observer- reliability = observation
Inter- rater- reliability = questionnaire
Inter- interviewer- reliability = interviews
A null hypothesis
You use this when you want to prove that you should discontinue a research study (there is no relationship)
A directional hypothesis
You preempt the way your research will go in (the more/the less etc). What directional it will go in.
Non directional hypothesis
You say there is a relationship but you don’t say the direction
What is a paradigm shift?
A paradigm- a general law or principle that exists
Paradigm shift- throughout history there have been shifts in our general laws, brought into question by further research
What is the Harvard defence system?
Last name, First initial. (Year Published). Title. City: Publisher, Page (s)
What is the difference between correlations and experiments?
- correlations don’t manipulate the IV but look at the relationship between the two co-variables
- correlations don’t establish cause and effect. Look at the strength of the relationship known as the correlation coefficient
- an experimental hypothesis tests for a difference, whereas an alternative hypothesis tests for a relationship between two variables
In a correlational analysis can only be carried out on quantitative data. Essential to operationalise the co-variables, to enable they are clearly defined and measurable.
Steps to complete the Wilcoxon matched pairs signed ranks test
1- find the difference between each set of scores. Some of these differences may well be positive and negative differences
2- rank the differences ignoring the - or the + signs. Giving the highest difference rank 1
3- calculate the sum of - ranks and the sum of the + ranks. T = the smallest sum of the two ranks
4- calculate the n value (not the same as calculating the degrees of freedom ) n= number of difference.
5- level of significance - look at the critical value table to see whether the data is significant.
Steps to complete the Mann Whitney U test
1- data is ranked, the lowest given a rank of 1 and so on. If there is a tied score the rank scores are added and divided
2- total sum of rank 1
3- total sum of rank 2
4-
formula U1 = N1 N2 + (N1 (N1 +1)/2) - ER1
Formula U2 = N1 N2 + (N1 (N1 +1)/2) - ER2
U1 = 10x10 + 10(10+1)/2 - total sum of rank one
U2 = 10x10 + 10(10+1)/2 a total sum of rank two
5- Select whichever u value is the smallest as your calculated value
6- look at the critical table of values