-IV and DV and hypothesis -Control of Variables -Experimental designs -Types of experiments -Correlations -Presentation of data -Data analysis: Descriptive statistics -Distributions -Statistical testings
What is the IV and DV?
IV: What is manipulated and changed in each condition. Has two levels. An experimental and control condition.
DV: What is measured in each condition, must be operationalised.
Explain what a hypothesis is?
A statement made at the beginning of the study which clearly states a relationship between the two variables.
Directional hypothesis (One-tailed): Makes it clear the kind of differences we will see. Use it when there is past research which will inform us of the likely results
Non-directional hypothesis (Two-tailed): States there will be a difference but doesn’t specify which direction it will go in. Used when there is no past research or contradictory research.
Null Hypothesis: States there will be no difference at all. All researchers must use this
Correlational hypothesis: Starts with there will be a relationship. If directional add positive or negative relationship.
What are the different experimental designs?
Independent groups: Two seperate groups which each experience a different level of the IV. Two groups are then compared. +No order effects -Individual differences so don’t know if IV or participant variables created change, Random allocation could reduce this
Repeated measures: Participants complete all conditions. +Participants variables controlled, need fewer participants - Order effects could be a confounding variable. Could use counter balancing to minimise effect, - Participants guess aim of research lead to demand characteristics
Matched pairs: Participants paired together based on a variable relevant to research. One part of the pair does condition A the other part condition B. Scores then compared. +Reduces order effects and demand characteristics - Still some participant variables - Matching is time consuming and expensive
What is an Extraneous variable?
Extraneous Variables
An unwanted variable which should be identified at the start of the research and steps taken to control it. They are a nuisance but do not impact the DV so much that we cannot tell if it is the IV or the extraneous variable that is making change. There are four main types of an extraneous variable.
What are the 4 main types of an extraneous variable?
Investigator effects
The unconscious behaviour of a researcher which will affect the results eg. body language. When the researcher’s influence, bias or characteristics alter the results. Reduce by using same researcher for all research and randomisation.
Demand Characteristics
Participants guessing the aim. Use deception and/or independent groups design
Participant variables
The individual characteristics of the participants can affect the results. Reduced by using matched pairs/repeated measures
Situational variables
The researchers environment could impact the results ed. Temperature and lighting. Reduce by using standardisation
Confounding variables
They have such an effect we cannot be sure what changed the DV. Therefore we can no longer trust the results
How to reduce the effects of extraneous variables?
Randomisation
Using chance as much as possible to reduce extraneous, confounding and investigator effects. Allocating participants should be done randomly
Standardisation
Making sure every participant gets the same experience.
What is a lab experiment?
A highly controlled environment (doesn’t need to be a lab)
+High control over extraneous variables, +Replication more possible than other types of experiment because of high level of control
-Lab experiments lack generalisability due to low ecological validity
-Demand characteristics as Participants know they are in a lab
-Lacks mundane realism
What is a field experiment?
The IV is manipulated in a natural setting e.g. Every-day life
+Higher mundane realism and ecological validity
-Loss of control of variables. Harder to establish cause and effect. Difficult to replicate
-Ethical issues if participants do not know if they are taking part
What are natural experiments?
A pre-existing independent variable is being studied. The situation would have existed even if the research wasn’t happening (Romanian orphans)
+Opportunity to study things that wouldn’t normally be studied due to practical or ethical reasons
+High external validity
-Happens rarely, not many opportunities for research
-Participants cannot be randomly allocated to conditions so not sure if IV affected DV
What are Quasi-experiments?
The IV is an existing difference between people e.g. Gender, age.
+Often carried out in alb so share strengths
-You cannot allocate participants to conditions so may be a confounding variable
What are correlations?
Correlations show a strength and direction of a relationship between variables. In an experiment a researcher control and manipulates the IV to see the effect on the DV. It is cause and effect. In correlation no manipulation exists. Simply measure two variables and see if any relationship exists between them. Correlations are plotted on scattergrams. One co-variable forms the x-axis and the other the y-axis.
What’s the difference between Positive, Negative and Zero correlation?
Positive Correlation: When one co-variable increases so does the other. (Line goes up from left to right)
Negative correlation: When one co-variable increases the other decreases (Line goes down from left to right)
Zero correlation: No relationship between the co-variables (No pattern)
How to write a hypothesis for a correlation?
Identified whether it is directional or non-directional
Said there will be a relationship
Identified both co variables
What is the evaluation of Correlations?
Strengths
Useful to start with. Suggests areas for future research
Weaknesses
No manipulations we cannot establish a cause and effect
Another untested variable is causing the relationship called the Third variable problem
What are the different presentation and displays of quantitative data?
Table: Summary table
Graphs
Continuous data: Quantitative data that can be measured with an infinite number of possible values within a selected range e.g. temperature
Discrete data: Quantitative data that can be counted and has a finite number of possible values e.g. days of week.
Bar chart: Used with discrete data, shows difference in data
Scatter Graph: Discrete-correlational, Shows association between co-variables
Histogram: Continuous (Grouped data), Shows how grouped data is spread
Line Graph: Continuous (not grouped data), Shows how a variable changes (often over time)
How to measure mean, median and mode?
Mean
Add up all scores and divide by total number of scores.
+ Most sensitive as includes all the scores. More representative
- Distorted by extreme values
Median
Middle score when data is arranged from lowest to highest.
+ Not affected by extreme scores.
- Less sensitive as not all scores represented.
Mode
Most frequently occurring number.
+Easy to calculate. For categorical data it is the only measure that can be used.
- Crude measure
What are the measures of dispersion?
How the scores are spread (how the vary from one another)
Range: Take the lowest value from the highest value and usually adding 1. The 1 accounts for margin of error. + Easy to calculate - Only takes into account 2 most extreme scores so unrepresentative of data set as whole.
Standard deviation: A score that tells you how far all the data deviates (moves away from) the mean. The larger the standard deviation he greater the dispersion within a set of data. If looking at conditions in an experiment, a large standard deviation suggests not all participants are affected by the IV in the same way. A low standard deviation suggests the scores are clustered around the mean so the participants responded in a similar way. + More precise as it uses all data - Can be distorted by a single value.
What are the different distributions?
The distribution of data is often showed in graphical form. The shape shows the range and spread of data.
Normal distribution: A bell curve. It is symmetrical. The mean, mode and median all occupy the same mid-point of the curve which extend outwards and never touches the x axis.
Skewed distribution: Not all distributions form a balanced symmetrical pattern. Some lean to one side or another.
Positive skew: Distribution is focused on the left resulting In a long tail on the right. The mode is the highest peak, median comes next and the mean has been dragged along to the right by extreme scores.
Negative skew: A negative skew has the bulk of the scores concentrated on the right resulting in a long tail to the left. The mode is the highest point. The median is then next to the left. The mean is now furthest left.
What is statistical testing?
Statistical tests show us the results are due to the IV and therefore we can reject the null hypothesis.
Sign Test
To use the sign test we need to be looking for a difference and not a correlation. Use repeated measures. Uses nominal data. In psychology we uses 0.05. Occasionally we use 0.01 if the research is rare and cannot be completed or there is a human cost to our results such as drug trials.
How to do a sign test?
1. Decide if it a one tailed or two tailed hypothesis.
2. Compare condition 1 to condition 2. If the number is increasing give it a +. If decreasing give it a -. If it is the same give it a = and remove that data from your set. Total up the numbers of + and -. The smallest number is your S value.
3. To find N, you count up the number of participants removing anyone who had an = sign.
4. Use the critical value table. If your S value is equal or smaller then it is significant.
5. Check the result goes in the right direction if using a one tailed hypothesis.