Research methods Flashcards
Explain paradigms as a feature of science
its a shared set of assumptions about the subject matter and the methods appropriate to study it. idea developed by Kuhn. 3 stages in the development of science
pre science- no paradigms
normal science - paradigm established
revolution - when a paradigm is challenged and disproved, replaced with one that satisfies any contradictory finings
how do you report psychological investigations 9 stages
title - clear relevant informative
abstract- summary of the research
intro- why study was conducted and provides context for the rest of the work
hypotheses - experimental and null
method- design, pp, materials, standardised procedures, controls
results- table, descriptive and inferential stats
discussion- explanation, relationship to background research, limitations modifications and implications, suggestions for future research
references
appendices
what is quantitative data positive and negative
measuring behaviour in a numerical way. e.g score on a memory test. easy to analyse, can replicate objective
less meaningful and low in ecological validity
describe qualitative data positive and negative
focuses on pp thoughts and feelings about their life. non numerical data. rich and detailed meaningful with high validity and context
difficult to replicate and analyse low reliability (subjective)
what is primary data positives and negatives
the original data that was collected specifically for the purpose of the investigation done by the researcher. + psychologists can collect exactly what they want and can trust the results
- time consuming and may be difficult to access
describe secondary data and positives and negatives
collected by someone other than the researcher. data already exists before the investigation takes place and comes from an outside source. e,g in journals or websites
+ quicker to collect than primary, can access data that an individual may not e.g crime stats
-might not contain info that relates exactly to the investigation
less trustworthy
describe meta analysis and the positives and negatives
combining the findings from several studies that allows the identification of trends and relationships that might not be possible from smaller studies. especially helpful when smaller studies have found contradictory evidence or weak results.
+ very useful for comparing
- compared studies may not have been carried out in exactly the same way
what are the measures of central tendency
mean- calculated by adding all the numbers together and dividing by the number of values. takes all values into account so is most sensitive but outliers can skew the result
mode- most common. easy to calculate, least sensitive least useful
median - middle score. put in order find middle value
good because not distorted by extreme values. less sensitive than mean because doesn’t take into account every value
what are the measures of dispersion
range and standard deviation
they measure the spread of data
how to find the range
subtract the lowest score from the highest score and add 1. problem= the range can be distorted by an extremely high or low score (has to beat either end) leading us to believe it is spread out when the other data points are close to the mean
what standard deviation does and what it means
takes into account every score so represents the spread as a whole. low sd means the scores aren’t spread out from the mean. its more sensitive as every value taken into account but can be distorted by an extreme value
what do correlation coefficients tell you
the strength of the correlation 0.8 is strong 1 is a perfect correlation
the direction of the correlation positive or negative
what must the title in for the row of a table include
both conditions e.g male and female and the unit of measurement
when are bar charts used
when data is in categories (discrete)
frequencies on the y axis and variables on the x
bars don’t touch
when are histograms used
bars touch each other continuous data
x axis equal sized intervals and y axis the frequency
when are scattergrams used
for correlations, relationship between two co variables
one co variable on each axis
what are the characteristics of a normal distribution curve
specified by the mean and the standard deviation
curve is symmetrical about the midpoint of the horizontal axis
mid point is the mean median and mode
68% is within the standard deviation of the mean
what are the skewed distributions
deviate substantially from the characteristics of the normal distribution
negative skew- most of the values are to the left of the peak frequency, mean is the middle of x axis then median then mode (peak away from y axis)
positive skew - most values to the right of the peak, mode then median then mean. peak close to the y axis
what are the levels of measurement
Nominal- involves frequency data. the number of times something happens in a category e.g number of sunny days
ordinal - data that is ranked with no set interval between the results, e.g 1st 2nd 3rd in a race is ranked but the differences between athletes could be different
interval - most accurate form of measurement, uses equal measurement intervals. standardised measurement units are used like time weight temperature etc.
what does p<0.05 mean
there is a 5% possibility that the observed difference (or relationship) between the two sets of data is not a real difference but occurred due to chance factors.
when do you use a directional hypothesis
if there is evidence form previous research that predicts the direction of the results. if only one result is desired e.g drug efficacy
when do you use a non directional hypothesis
if there is no evidence from previous research or if there is conflicting previous research
whats a type one error
false positive. when we say the difference is significant and its not. null hypothesis wrongly rejected. the significance level too high e.g 10% instead of 5%
whats a type two error
false negative. we reject the experimental hypothesis when we should accept it. happens when the significance level too low e.g 1% instead of 5%
type 1 and 2 errors and how to solve them
the stricter the significance level the less likely a type 1 error is but more likely for type 2 and vice versa. reduce the chance of errors by increasing the sample size
what do you need to include when explaining why you have chosen a particular stats test
difference correlation or association
which measure you’ve used if its a difference e.g independent groups or repeated
which level of measurement it is
write out the table
I r a c
n chi sign chi
o m w wc pearsons
I unr.t rel t spearmans
how to do a sign test
subtract the second column from the first column if answer is positive put a + next to it or a - if its negative
add up total number of + and - whichever is lower that is the calculated value
n = number of participants. if a pp gets a 0 (neither positive or negative) take them out of the sample
write a model conclusion
the calculated value of 7 is more than the critical value of 5 so for a p<0.05 two tailed test we reject the experimental hypothesis and accept the null as there was a greater than 5% probability the difference was due to chance.
