Research Methods 2 Flashcards
1
Q
What is quantitative data?
A
Reports on data in numerical form quaNtitative
2
Q
Give some examples of quantitative data
A
- Percentages - Mean, mode, median - Range
3
Q
Give some examples of types of experiments which could produce quantitative data
A
- From a lab experiment (e.g. measuring testosterone) - Structured questionnaire/interview with closed questions (tally) - National statistics, e.g. data on crime rates
4
Q
What are the strengths of quantitative data?
A
- Large sample = generalisable - Higher chance of establishing cause and effect (objective) - Easy to analyse - Easier to make comparisons and see patterns and trends - Can repeat to test reliability
5
Q
What are the weaknesses of quantitative data?
A
- Statistics sometimes distort the truth and therefore it may lack the validity - Does not give context, i.e. conclusions may lack depth and detail
6
Q
State a question that would collect quantitative data
A
- Your patient has felt better since having physiotherapy following their stroke - strongly agree -> strongly disagree (take tally of answers) - On average, how many hours of physiotherapy does the patient receive per week
7
Q
What is qualitative data?
A
Reporting anything in word form with written language and context quaLitative
8
Q
What types of experiment would produce qualitative data?
A
- Transcripts of unstructured interviews with pps - Collation of answers from open ended questions in a questionnaire - Written report of free flowing observation
9
Q
What are the strengths of qualitative data?
A
- Depth gives detailed insight by allowing a range of responses/behaviours - increases validity - Taking the context into account makes results more valid - Chance for new or unexpected info with open ended questions and unstructured interviews
10
Q
What are the weaknesses of qualitative data?
A
- Harder to analyse/interpret - Behaviour in observations and interviews are open to interpretation (subjective) - Can be unreliable in terms of exactly how info is gathered - Unrepresentative if sample size is very small
11
Q
Give a question which would collect qualitative data
A
Since having their physiotherapy, what effects have you noticed in your patient?
12
Q
What is primary data?
A
Data collected first hand from the pps (original data) specifically towards a research aim which has not been published before
13
Q
What research methods would produce primary data?
A
- Interviews - Observations - Questionnaires - Physical testing - Diary extracts (from pps)
14
Q
What is a strength of primary data?
A
More reliable and valid than secondary data as it’s not been manipulated in any way
15
Q
What is secondary data?
A
Data originally collected towards another research aim which has been published previously
16
Q
What research methods would produce secondary data?
A
- Websites - Journal articles - Books - Government publications - Diary extracts (published)
17
Q
What is a strength of secondary data?
A
If drawn from several sources, can help to give a clearer insight into a research area
18
Q
What is a meta-analysis?
A
The collation and review of the results/findings of multiple research studies in the same area of study
19
Q
How do we conduct a meta-analysis?
A
We do not replicate the chosen studies, but trust in and review the findings
20
Q
What are the strengths of a meta-analysis?
A
- Allows for the identification of trends and patterns that would not be possible to see in smaller individual studies - Can improve reliability of findings because sample sizes can increase greatly. This in turn can increase validity
21
Q
What is content analysis?
A
A method that quantifies qualitative data, for example, creating a tally - translating words into numbers
22
Q
What types of qualitative data can content analysis quantify?
A
- Spoken interactions (conversations) - Written forms (texts/emails) - Media (books, magazines, TV)
23
Q
How is content analysis carried out?
A
- Identify key themes within the info and categorise data into meaningful units based on what you are trying to answer - sometimes known as “coding units”, e.g. references to positive behaviour in the info you are analysing - Re-read or listen to the info several times - Count the number of times a particular word or phrase is present, i.e. producing a tally, giving you quantitative data, e.g. if looking at stereotyping against the mentally ill, analysing the number of times terms like “crazy” or “mad” are used in the media
24
Q
What are coding units?
A
Required to categorise analysed material For example: - If the unit is word, a researcher might count the number of slang words used - If the unit is theme, a researcher might count the amount of violence on TV - If the unit is character, a researcher might count the number of female commentators in sports programmes on TV - If the unit is time and space, a researcher might count the amount of time (TV) and space (newspapers) dedicated to mental health
25
What are the strengths of content analysis?
- Easy to perform, inexpensive, non-invasive with no direct contact with pps - Complements other methods - verifies results from other methods, particularly being useful in longitudinal methods detecting changes over time - Reliable - very easy to replicate
26
What are the weaknesses of content analysis?
- Lacks descriptiveness - "what" but not "why" - Flawed results - if limited material is available then the analysis may not reflect reality, for example, negative events receive more coverage than positive ones in the media
27
What is thematic analysis?
Purely a qualitative analytical method identifying, analysing and reporting themes (patterns)
28
What does a thematic analysis do?
- Organises, describes and interprets data - A psychologist uses these themes to address the research or say something about an issue. This is much more than simply summarising the data; a good thematic analysis interprets and makes sense of it
29
What are the 6 stages of carrying out thematic analysis?
- Familiarisation with the data - Coding - Searching for themes - Reviewing themes - Defining and naming themes - Writing up
30
What is familiarisation with the data as a stage in thematic analysis?
Intense reading of data to become immersed in content
31
What is coding as a stage in thematic analysis?
Generating codes (labels) that identify features of the data important to answering the research question
32
What is searching for themes as a stage in thematic analysis?
Examining the codes and data to identify patterns of meaning (to find potential themes)
33
What is reviewing themes as a stage in thematic analysis?
Checking the potential themes against the data to see if they explain it and answer the research theme
34
What is defining and naming themes as a stage in thematic analysis?
Detailed analysis of each theme creating an informative name for each one
35
What is writing up as a stage in thematic analysis?
Combining together the information gained from the analysis
36
How do you calculate percentage decrease?
Calculate the difference between the starting point and how much it goes down by, then divide the difference by the original number and multiply by 100 to get a percentage E.g. percentage decrease in mean no of hours worked before and after Xmas - hours before = 30, hours after = 13 30-13 = 17 17/30 x 100 = 57%
37
How do you calculate percentage increase?
Calculate the difference between the end point and the start point, then divide by the original number (which is the lower number) and multiply by 100 to get your percentage For example: Percentage increase in hours worked at Xmas - hours before = 13, hours at Xmas = 30 30-13 = 17 17/13 x 100 = 131%
38
In central tendency, how do you calculate the mean?
It is the average score of the data - to calculate, add all of the scores together and divide by the total number of scores
39
In central tendency, how do you calculate the median?
It is the central/middle number in a list of rank ordered scores - odd number median = middle number, even number median = midpoint of the 2 middle scores (this may not be one of the original scores)
40
In central tendency, how do you calculate the mode?
It is the most common/frequent in a set of scores
41
What are the advantages of taking a mean?
- Uses all data in the calculation - Most accurate measure of central tendency - uses internal level of measurement when the units of measurement are all of equal size, for example, seconds in time
42
What are the disadvantages of taking a mean?
- Less useful if there is an extreme score which will then skew the data - The mean score could be one that isn't actually listed in your data set, which can create difficulties when applying to people, e.g. you can't have .1 of a person
43
What are the different types of statistical test?
- Sign test - Chi-squared - Spearman's Rho - Pearson's R - Wilcoxon - Related T-test - Mann-Whitney - Unrelated T-test
44
Which is the only statistical test you need to be able to calculate?
The Sign Test
45
Why do psychologists have to do statistical tests once they have collected their data?
- Research produces data which psychologists have to analyse to make sense of - You can purely describe what the data shows in several ways (e.g. mean, median, mode, graphs, tables, etc) - A more sophisticated form of analysis is to carry out inferential statistical tests
46
What is meant by the term "inferential (statistical) testing"?
A psychologist can make informed decisions about whether differences or relationships in data are significant ones, i.e. beyond the boundaries of chance, and can then be applied to whole target populations that the sample represents
47
What questions should you ask to determine which test you should choose?
1. Is the research looking for a difference or a relationship (correlation)? HINT: IF IT SAYS IT'S LOOKING FOR A RELATONSHIP, IT'S A CORRELATION - What experimental design has been used? (repeated measures, matched pairs, independent groups) 3. What type of data did the psychologist collect? (nominal/ordinal/interval)
48
"A psychologist wanted to test whether listening to music improves running performance" Is the research looking for a difference or a relationship (correlation)?
A difference (between running performance with or without music)
49
"All participants were asked to run 400 metres as fast as they could on a treadmill in the psychology department. All participants were given standardised instructions. All participants wore headphones in both conditions. The psychologist recorded their running times in seconds. The participants returned to the psychology department the following week and repeated the test in the other condition." What experimental design has been used? (repeated measures, matched pairs, independent groups)
Repeated measures (participants ran both with and without music)
50
How can "types of data" also be phrased?
"Levels of measurement"
51
What are the three types of data?
- Nominal - Ordinal - Interval
52
What is nominal data?
Each item can only appear in one category (discontinuous) It is the lowest level of measurement/the simplest data collection and is the most uninformative type of data
53
Give an example of nominal data
Blood groups
54
What data is produced by nominal data?
Qualitative data - it cannot be measured numerically
55
How is nominal data represented?
In a bar chart
56
What is ordinal data?
You can automatically rank this data into place order A problem with the data is that the gap between each data point may be different
57
Give an example of ordinal data
Finishing places in athletics (tells us which athletes are better than others but doesn't say how much better 1st place is to 2nd) Use of the Likert scale (strongly agree - strongly disagree)
58
What data is produced by ordinal data?
Quantitative and qualitative data can be used
59
How can ordinal data be represented?
By ranking/scales
60
What is interval data?
Standardised measurement UNITS like weight, temperature, time and distance - units like g, degrees C, s, m "Highest level of measurement" Most informative as the distance between each data point (or the interval) is exactly the same - most informative when drawing conclusions
61
Give an example of interval data
How long, heart rate, etc
62
What type of data is produced by interval data?
Quantitative only
63
How can interval data be represented?
On a number line
64
What type of data is time taken to sort cards into categories?
Interval data as time has a unit
65
What type of data is people's choice of the Sun, The Times or the Guardian?
Nominal data as people can only go in one category, the Sun, the Times or the Guardian
66
What type of data is participants' sense of self-worth, estimated on a scale of 1-10?
Ordinal data as pps' estimations of self worth can be ranked
67
What type of data is judges in a dancing competition giving marks for style and presentation?
Ordinal data as the marks given by the judges can be ranked
68
What type of data is participants' reaction to aversive stimuli measured using a heart rate monitor?
Interval data as heart rate has a standardised unit
69
What type of data is a set of medical records classifying patients as either chronic, acute or "not yet classified"?
Nominal data as patients can only go in one category - chronic, acute, or "not yet classified"
70
Give a mnemonic for remembering the statistical test table
Cats Share Carrots Meanwhile Wobbly Shrek Unleashed Rabbid Parrots
71
A psychologist is testing the effect the number of hours of sleep has on cognitive function. He tests two groups of pps in a simple word recall task. Pps in group 1 sleep 6 hours per night, whereas pps in group 2 sleep 10 hours. State the correct statistical test to analyse the data collected and justify your choice (4 marks)
The study is testing a difference between effect on cognitive function after having 6 hours or 10 hours of sleep The study is using an independent groups design as participants only experience one of the conditions - they either sleep for 6 hours or 10 hours The study is collecting ordinal data as the word recall scores produced by the participants can be ranked from best to worst Therefore the statistical test needed to analyse the data collected is the Mann-Whitney test
72
During an investigation on schizophrenia a psychologist aims to measure the relationship between pps' numbers of delusions per months and the time of onset for their symptoms. He uses a content analysis method to gain info from local hospitals State the correct statistical test to analyse the data collected and justify your choice (3 marks)
The study is testing a correlation between participants' numbers of delusions per month and the time of onset for their symptoms The study is collecting ordinal data as the numbers of delusions can be ranked from most to least Therefore the statistical test needed to analyse the data collected is Spearman's Rho
73
What do you do if there are two types of data collected in a study?
Always go to the lowest data level of the two, e.g. if the two were interval and ordinal, you would go with ordinal as it is the lowest level
74
A researcher is conducting a study to find out if a positive view towards Donald Trump correlated with a negative view towards Mexicans in society. He uses interview methods to compare pps' views in each case State the correct statistical test to analyse the data collected and justify your choice (3 marks)
The study is testing a correlation between positive views on Donald Trump and negative views on Mexicans in society The study is collecting nominal data as each participant can only fit into one category - they either have a positive or a negative view towards Donald Trump and a positive or negative view towards Mexicans in society and there are no cases of pps being in between categories Therefore the statistical test needed to analyse the data collected is the Chi-Squared test
75
Can researchers ever be 100% certain they have found statistical significance?
NO - it is usually up to 5% possible that the wrong hypothesis may be accepted
76
What is a Type I error?
- When the null hypothesis is rejected and the experimental hypothesis is accepted when it should have been the other way around as the null hypothesis is true - Referred to as an optimistic error or false positive as the researcher claims to have found a significant difference or correlation when one does not exist
77
What is a Type II error?
- When the null hypothesis is accepted but it should have been the experimental hypothesis as the experimental hypothesis is true - This is a pessimistic error or false negative
78
When are we likely to make a Type I error?
If the significance level is too lenient, e.g. 0.1 or 10%
79
When are we likely to make a Type II error?
If the significance level is too stringent, e.g. 0.01 or 1% as potentially significant values may be missed
80
Which significance level do psychologists favour and why?
The 5% level - it best balances the risk of making a type I or II error
81
What is probability?
A measurement of the likelihood that an event will occur, i.e. chance - this can be represented numerically
82
What does it mean if the probability is 0?
There is no chance it will happen, i.e. statistically impossible (a person getting to the moon in a second) - also represented as 0%
83
What does it mean if the probability is 1?
Statistically certain (Christmas will happen on 25th December) - also represented as 100%
84
What is meant by the term "significance level"?
The point at which a null hypothesis can be accepted or rejected - psychologists ideally want to be able to reject the null hypothesis and accept their experimental one
85
How is the accepted significance level (0.05) represented and what does this translate to?
p < 0.05 The probability that the results of the experiment are due to chance is less than 5% You are 95% certain that the results show a significant difference/relationship
86
How is the 0.01 significance level represented and what does this translate to?
p < 0.01 The probability that the results were due to chance is less than 1% You can be 99% certain there was a significant difference - this is VERY high
87
How is the 0.1 significance level represented and what does this translate to?
p < 0.10 The probability that the results were due to chance is less than 10% You can only be 90% certain there was a significant difference - this is too uncertain or lenient
88
Why might a researcher in a clinical trial of a drug (which can only be done once due to side effects) use the 1% level of significance?
They can be 99% certain that they have found a significant difference between symptoms after using the drug vs not using the drug (or a placebo), and as it can only be done once, there is not room for uncertainty. The findings can have far reaching consequences for a person's health and so being 95% certain is just too uncertain here
89
Which two values do you need to be able to work out in the sign test?
S (the calculated value) N (number of pps showing a significant difference)
90
What is the first step in using the sign test?
Control result - experimental result Record the sign produced - if the value is positive, write a +, if it is negative, write a -, and if it is the same result, write a 0
91
How do you calculate the S value in the sign test?
Count up the number of times the + and - occurs - the least frequent sign is the S value (calculated value)
92
How do you calculate the N value in the sign test?
Count the total number of + and - signs (don't count the 0s) - the result is the N value
93
What is the fourth step of the sign test?
Determine if the hypothesis is one or two tailed
94
How do you get your critical value in the sign test?
Look at the level of significance column (usually 0.05 unless otherwise stated) under the heading of the hypothesis used in the experiment (one or two tailed) and follow it down to meet your value of N. The value here is your critical value
95
How do you determine if a result is significant using the sign test?
Compare your S value (calculated value) to the critical value obtained from the table - the S value must be equal to or less than the critical value to be significant
96
What is the final step of calculating the sign test?
State the conclusion - state the calculated value of S, if it is less, more or equal to the critical value, state the critical value, determine if the finding is significant or not, state which hypothesis you accept and which you reject, write conclusion of whether the researcher's hypothesis was correct or not (e.g. if using a proposed worksheet will improve exam scores or not)
97
What 3 factors do you need to determine which critical value is the correct one to use from a critical values table (not sign test)?
- What kind of hypothesis was used (one tailed or two tailed/ directional or non-directional) - Number of participants (the N value) or degrees of freedom (the df value) - Level of significance (usually 0.05 unless otherwise stated)
98
How do you determine your critical value from a table other than the sign test one?
Look at the column of the significance level under the heading of the type of hypothesis used and match this to the number of participants (N) or the df - the value crossing both is the critical value - them compare your calculated (S) value to the critical value
99
What is the formula to calculate df?
(rows - 1) x (columns - 1) - substitute in the number of rows and columns in the results table where raw data has been entered and then use the formula
100
What is the rule of R?
Statistical tests with a letter "R" in their name are those in which the calculated value must be more than or equal to the critical value
101
In which statistical tests must the calculated value be equal to or more than the critical value for there to be statistical significance?
- Chi-squared - Unrelated T-test - Spearman's Rho - Related T-test - Pearson's R
102
In which statistical test must the calculated value be equal to or less than the critical value for there to be statistical significance?
- Sign Test - Wilcoxon - Mann-Whitney
103
What is a correlation?
The relationship between two co-variables
104
How is a correlation represented graphically?
The relationship is plotted on a scatter graph
105
What is a correlation co-efficient?
- Represents the strength of the correlation by placing a numerical value between +1 and -1 - +1 represents a perfect positive correlation and -1 represents a perfect negative correlation - The closer to 0 the weaker the correlation
106
How could a weak correlation be statistically significant?
Due to a large data set
107
How are correlation coefficients calculated?
Using a Spearman's Rho or Pearson's R statistical test
108
What is the number above or equal to which a correlation is seen as strong?
+ or - 0.8 "It ain't great if it ain't 0.8"
109
What are the strengths of correlations?
- Can study relationship between variables that occur naturally (we can measure things that cannot be manipulated experimentally) - Can suggest trends that can lead to experiments
110
What are the weaknesses of correlations?
- It is not possible to say that one thing causes another - There may be another variable connecting the two variables
111
What do descriptive statistics do?
Give us a way to summarise and analyse numerical data in order to draw meaningful conclusions
112
What are the 3 measures of central tendency?
- Mean - Median - Mode
113
What are the strengths of using the median?
- Not affected by extreme scores - Easy to calculate - Can be used with ordinal (ranked) data whereas the mean can't
114
What are the weaknesses of using the median?
- It is less sensitive than the mean as not all scores are included in the final calculation - Can be un-representative of small sets of data in particular
115
What are the strengths of using the mode?
- Unaffected by extreme values - Only method that can be used when data are in categories
116
What are the weaknesses of using the mode?
- Loses meaning if there is more than one mode - It does not use all scores in its description
117
What is a measure of dispersion?
Summarises and describes data by showing us the spread of a set of data - how far the scores vary from one another
118
What are the two measures of dispersion?
- Range - Standard deviation
119
How do you calculate the range?
Largest value - smallest value
120
What is a strength of using the range?
Easy to calculate
121
What are the weaknesses of using the range?
- Only takes into account the two most extreme values, which may be unrepresentative of the data set as a whole - Doesn't indicate whether most numbers are closely grouped around the mean or spread out evenly
122
What does it mean if you have a low range value?
Data is consistent
123
What does it mean if you have a high range value?
Data is inconsistent
124
What is standard deviation?
A more sophisticated measure of dispersion - a single value that tells you how far scores deviate (move away from) the mean
125
What does a low standard deviation value mean?
Data is consistent (clustered around the mean) A smaller dispersion of data - all pps responded in a similar way to the IV
126
What does a high standard deviation value mean?
Data is inconsistent (spread from the mean) A greater dispersion of data - pps not affected by IV in the same way
127
What does a standard deviation of zero mean?
All of the data was the same
128
What are the strengths of using standard deviation?
- More sensitive than the range as includes all values in final calculation - Less likely to be distorted by a single extreme value
129
What are the weaknesses of using standard deviation?
- Harder to interpret if the data is not in a normal distribution curve - Hard to calculate
130
What do you need to put in a summary table if you are asked to draw one?
You do not need to show the raw data, but you do need to calculate the mean, median/mode, range and show the totals for each condition
131
What two things do you always need on a graph?
- X and Y axis labelled and operationalised - A title operationalised
132
How should you use a bar graph?
- Used to compare categories - Uses discontinuous/discrete data - bars must be separate - X axis = IV (what you change) - Y axis = DV (what you measure) For example, data for different blood groups would be represented on a bar graph
133
How should you use a histogram?
- Bars touch each other - data is continuous rather than discrete - X axis = equal sized intervals of a single category (e.g. score) - Y axis = frequency within each interval - If there was zero frequency for one of the intervals, it remains without a bar
134
How should you use a line graph?
- Continuous data - Dot to mark middle top of where each bar would be and each dot is connected by a line - X axis = IV (what you change) - Y axis = DV (what you measure)
135
How should you use a scattergram?
- Used when doing correlational analysis - do not depict differences, but associations between co-variables - It does not matter which variables go on the X and Y axis - Each cross on the graph corresponds to the X and Y positions of the co-variables
136
What are distributions?
A way of visually showing where the mean, median and mode lie
137
What is a normal distribution?
- Classic bell-shaped curve - The predicted distribution when considering an equally likely set of results
138
How can we identify a normal distribution?
- Mean, median and mode are all in exact midpoint - Distribution is symmetrical around this midpoint - Dispersion of scores/measurements either side of the midpoint = consistent and can be expressed in standard deviations
139
What always goes on the Y axis of a distribution graph?
Frequency
140
How are the measures of central tendency plotted on a skewed distribution graph?
Mode = at the top middle (peak) Median = in the middle of the mean and the mode (either to the left or right) Mean = at the bottom (either to the left or right)
141
When does a skewed distribution occur?
When scores are not distributed equally around the mean - the distribution appears to lean to one side or the other
142
What is a positive skew?
Where most of the distribution is concentrated towards the left of the graph (tail lies to the right) - occurs, for example, when students take a very difficult test and most people got low marks with only a handful getting to the higher end
143
How can we identify a positive skew?
- The tail lies at the positive (right) end of the number line - Mode has highest frequency (peak) - Median is next mark - Mean is to the right as it has been affected by the few high scores and pulled upwards
144
What is a negative skew?
Where most of the distribution is concentrated towards the right of the graph (tail lies to the left) - would occur, for example, if students took a very easy test and most people got high marks with only a handful getting low marks
145
How can we identify a negative skew?
- The tail lies at the negative (left) end of the number line - Mode is the highest frequency (peak) - Median is the next mark - Mean is to the left as it has been affected by the few low scores and pulled downwards
146
How do you calculate decimal places?
- Count no of decimal places you need to go after the decimal point - this can include 0s - Look at number after the decimal place you are asked to go to - if it's 5 or above, round up
147
How do you calculate significant figures?
- 1 sig fig - 1st no you count must not be zero, e.g. 0.145, the 1 is your first sig fig - Use same rounding rule - if the fig after is more than 5, round up - 2 sig fig - count along from first figure which is not a zero
