Choosing an inferential statistic and the need for this with potential errors Flashcards
Unrelated data can refer to which experimental design?
Independent groups
Related data can refer to which experimental design?
Repeated measures or matched pairs.
Nominal data -
Data that can be counted into different categories E.G number of red or white roses.
Ordinal data -
Data which is ranked in order (rate of preference for the following foods, Chips, curry, ice-cream) or rated on a scale, these categories labelled can be defined as very confident, confident or not confident at all. Subjective.
Interval data -
A data type which is measured along a scale, in which each point is placed at equal distance from one another
Ratio data -
Like interval data but with a fixed 0 point, as there is a 0 it is possible to make fixed 0 point.
What is the T-test (unrelated) used for? (design and data)
Independent groups and interval/ratio
What is the T-test (related) used for? (design and data)
Repeated measures/matched pairs and interval/ratio.
What is the Pearson’s r test used for? (design and data)
Correlation and interval/ratio data.
What is the Mann Whitney U test used for? (design and data)
Independent groups and ordinal data.
What is the Wilcoxon matched pairs test used for? (design and data)
Matched pairs/repeated measures and ordinal data.
What is the Spearman’s rank test used for? (design and data)
Correlational and ordinal data
What is the Chi squared test used for? (design and data)
Independent groups and nominal data
What is the Sign test used for? (design and data)
Repeated measures/matched pairs and nominal data.
Only correlational tests are?
Spearman’s rank and Pearson’s test.
Why do we need inferential statistics?
Measure probability, to see if anything found was down to chance, so this allows to see if there indeed was a significant difference.
Type 1 error -
When someone is too optimistic about the findings of the research, when someone wrongfully rejects the null hypothesis and accepts the experimental/alternative hypothesis, assuming the findings to be significant but really due to chance.
Type 2 error -
When someone is too pessimistic about the findings where they assume the experimental/alternative hypothesis has not been supported and are due to chance however the results are significant. The researcher wrongfully rejects the alternative/experimental hypothesis and accepts their null hypothesis.
