INVESTIGATION DESIGN, DATA ANALYSIS AND PRESENTATION Flashcards
Hypothesis
A statement that is testable
Null and alternative hypothesis are proposed. Based on findings, the researcher decides whether it is more plausible to accept or reject each hypothesis.
–> if variations were unlikely to be by chance, null is rejected
Alternative hypothesis (H1)
(Also called the experimental hypothesis) predicts that something other than a chance variation alone has played a part in producing the results obtained
-Can be directional/one tailed or non directional/two tailed
Directional/one tailed hypothesis
Predicts the direction in which results are expected to occur
For example: those who eat chocolate will become fatter than those who eat fruit
Non-directional/two tailed hypothesis
Does not predict the expected direction of the outcome, only that there will be a difference caused by a variable
Null hypothesis (H0)
Predicts that the results of an experiment can be explained by chance variation alone rather than by the manipulation of the IV. Predicts that the alternative hypothesis is untrue
Independent groups design
Different participants are used in each condition of the experiment. Experiments using this condition usually have a control/experimental condition OR two+ experimental conditions
Participants are randomly allocated to ensure that participant variables do not differ systematically between the groups, or individual differences relevant to the experiment may confound results. Done with randomizer
Repeated measures design
Exposes every participant to all conditions of the experiment
Order effects can be minimised by counterbalancing or ranomization
(i) counterbalancing: equal number of participants completing tasks in different orders
(ii) Randomly determining the presentation of the experimental conditions
Matched pairs design
Middle way between independent groups/Repeat measures designs. Match participants in one condition as closely as possible with one in the other on all variables considered relative to performance in the study
Evaluation of independent groups design
Advantages: No order effect issues creating positive or negative performance in the task. Through repetition, boredom or fatigue
Disadvantages: Chance that individual differences would contaminate results. Uneconomical as 2* as many participants required
Evaluation of repeated measures design
Advantages: Individual differences removed as a confounding variable. Fewer participants required
Disadvantages: Risk of order effects as both participants exposed to both conditions in sequence. Further limited use, as in some cases it would be inappropriate to expose participants to two conditions
Evaluation of matched pairs design
Advantages: Combines benefits of independent groups/repeated measures designs
Disadvantages: costly, difficult and time consuming. Hard to match people up exactly.
Pilot study
Small scale trial run of a specific research investigation to test the procedure and highlight any potential issues. Identifies problems with design, instructions and measurement of the DV
Examples of random errors (cant be predicted)
- pps state of mind
- pps level of motivation
- incidental noise
- room temperature
- previous experiences of the pps on the day of the study
Random sampling
Every person in the target population has equal chance of being selected for the study. Must be completely random selection (ie random number generator) but cannot guarantee representative sample
Opportunity sampling
Researcher selects anyone available to participate in the study. Convenient but may not be representative
Volunteer sampling
Participants select themselves to take part in the study, walk in or after replying to advert. Hard to get responses and may not represent target population as a certain type of person (ie enthusiastic) will respond more readily than others in the population sample
Examples of demand characteristics (people try to make sense of the research situation they are in and act accordingly)
- Guessing the hypothesis and acting to appease it or derail it
- Act nervously/out of character, ie feeling evaluated and therefore nervous
- Displaying social desirability bias, wanting others to see the person in the most favorable light
Investigator effects
-Effect of the researcher’s own behavior and characteristics on the investigation. Bias observing due to wanting a particular outcome, altering results. Presence of observer in group may cause bias in naturalistic observation, and communication style may have an effect in interviews
Median
middle value of scores arranged in descending or ascending order. Unaffected by outlying values, safe measure of central tendency but affected by alterations to central values in a data set and not effective with small values
Mean
Arithmetic average. Powerful measure of central tendency as uses all data, but affected by anomalies in the data
Mode
Most frequently occurring value. Easy to identify, however when small data sets are compared, even tiny changes massively alter mode
eg: 3356838 vs 3356878. 3 vs 8 (?)
Standard deviation
Measure of the variability of the data from its mean
