Monday, 22 June 2015

KEY TERMS

Planning and conducting research:
  1. Semantic differential scale - this is a type of rating scale which allows participants to choose between two extremes. For example boring/exciting, friendly/unfriendly, like the scale that was used in the Baron-Cohen eyes task. 
  2. Practical problems – a type of extraneous variable including issues of cooperation of participants, practicalities of equipment and measurement, bad weather. 
  3. Single blind - this is when the participants are unaware of the level of IV in which they are participating in. This helps reduce demand characteristics. 
  4. Double blind - this is when neither the researcher nor the participants are aware of the condition in which they are in. This reduces chances of demand characteristics and researcher bias.
Data recording, analysis and presentation:
  1. Nominal data - is data that are produced as named categories, think of 'nom' meaning 'name' i.e. named categories. These categories can be allocated numbers, but these numbers bare no meaning. For example you may ask someone what their favourite chocolate is and provide them with the nominal categories; 1. milk choc 2. dark choc 3. white choc. As you can see they have been allocated numbers, but the number does not mean anything, white choc is not seen as better than dark choc. Closed questions often produce nominal data, as well as observations which code behaviour.  
  2. Ordinal data - is data which can be ranked in 'order' - ordinal. There needs to be an increase in the value of points along your data (therefore the numbers of your data do have meaning), but the size of each increase does not need to be equal. For example; a rating scale is a classic example of ordinal data, lets say a participant needed to rate how good they think their memory is, and the scale was; 1 very poor, 2 poor, 3 average, 4 good, 5 very good. As you can see, the measure is ranked in order, so that the higher the score, the better the memory and vice versa. However, if someone rated themselves as 'very good' we would know that they are better than someone who rated themselves as 'average' but, we couldn't know for sure if this means they were twice as good, triple, a small fraction? Other examples of ordinal data are; people competing in a race, rating scales, ranking top films etc.  
  3. Interval data - similar to ordinal data, but the divisions between the points on an interval scale are equal. For example, time, volume, speed, height, weight etc. The difference between 1 second and 2 seconds will always be equal. 
  4. < - less than 
  5. > - greater than 
  6. - greater than or equal to
  7. - less than or equal to 
  8. p - probability 
  9. Type 1 error - when you think you've won. i.e. you have accepted your alternate hypothesis when you should have rejected it. 
  10. Type 2 error - when you think their poo i.e. you have accepted your null hypothesis when you should have rejected it. 
  11. Internal reliability - this is the consistency of the items within the measure itself i.e. the questions) This shows that items in a self-report tool are measuring the same phenomenon. 
  12. Split-half reliability - is a measure of internal reliability in which scores from two halves of a test are compared. For example a questionnaire on self-intelligence may contain 20 questions. if scores from the first 10 questions are similar to the second set of 10 questions, the measure is seen to have high internal reliability. In addition, if certain questions do not produce consistent responses, they can be removed in order to improve reliability.  
  13. External reliability - does the measure produce the same results in the same situation with different people.
  14. Test-retest reliability - if a participant responds to the same test in a similar way, the test has high external reliability. 
  15. Mundane realism - the extent to which an experimental task represents a real-world situation. 
  16. Internal validity - are you measuring what you intend to measure; for example, in an experiment, whether changes in the DV are caused by the IV rather than extraneous variables. I will use the example of a self report to consider the many different validity terms:
  17. Face validity - this is whether your measure appears, at face value, to test what it claims to. For example if you took a quick glance at a questionnaire about fears of spiders, does it look like it is actually measuring fear of spiders? If yes, it has high face validity. 
  18. Criterion validity - this is whether a factor measured in one way will relate to, or predict, some other related variable. For example; can your CAT tests in year 7 predict the grades you will get in your GCSE's? This is also known as predictive validity.   
  19. Concurrent validity - whether a measure will produce similar results for a participant as another measure, that measures the same thing. For example; in Baron-Cohen, using the Strange stories task and the eyes task. 
  20. Construct validity - is the foundation of the theory you are testing valid? Does it actually exist? For example does Freud's Oedipus complex have construct validity? The answer would be no in this instant, as he had no valid evidence to support his theory.      
  21. External validity - this relates to the issues beyond the investigation, particularly whether the findings will generalise to other populations, locations, contexts and times than the ones investigated i.e. will the findings still be valid if I carried out a study in the UK and I'm trying to generalise it to America, if yes you have high external validity. 
  22. Population validity - following on from external validity, this is the extent to which findings from one sample can be generalised to the whole population from which the sample was taken and to other populations. If you have a small sample, say 50 participants, you are likely to have low population validity. Many things affect population validity such as; the sampling method used, sample size and narrowness of the sample, in relation to what is being studied. 
  23. Representative - does your sample represent your target population. To achieve representativeness, the sample studied should include a good cross section of the population, so that all categories of people within it are included. 

No comments:

Post a Comment