stats tests

Question 1

characteristics of non-parametric tests

Answer 1

• 30 data sets or less
• data is not normally distributed
• ordinal or nominal

Question 2

types of non-parametric tests

Answer 2

tests of difference = chi square, Mann-whitney U, Kruskal-Wallis
relationship = spearman’s rank

Question 3

nominal data

Answer 3

names

Question 4

ratio data

Answer 4

%, metres

Question 5

interval data

Answer 5

celcius

Question 6

ordinal data

Answer 6

categories eg. survey

Question 7

characteristics of parametric tests

Answer 7

• more than 30 data sets
• data is normally distributed
• interval/ ratio scale

Question 8

types of parametric tests

Answer 8

difference = t-test, ANOVA
relationship = Pearson's correlation

Question 9

interpreting regression analysis

Answer 9

1. R^2 value to % shows the variability of one can explain the other
2. F ratio- if greater than 1 can say that the explained is greater than the unexplained. Also compare F ratio to p value eg. F ratio is greater than 1 at 95% confidence level so the regression model is significant
3. if p value is lower than 0.05 then at 95% confidence level the regression coefficient is statistically significant
4. overall: comment on proportion of variability between the variables, explained variance, regression coefficient
   overall, is the regression model viable in nature?

Question 10

how to do a normality & null hypothesis question given output and p value

Answer 10

1. state the null and alt hyp
2. work out at 95% and say if significant
3. see if you can say it’s NOT normally distributed
4. 99% level

Question 11

is it good to have an F ratio greater than 1?

Answer 11

yes- means that the explained is greater than the unexplained

Question 12

what do you say if the F ratio is less than 1?

Answer 12

at (confidence level) the regression model is not significant

Question 13

what does p < 0.05 mean?

Answer 13

the p value is less than 0.05

Question 14

discontinuous data

Answer 14

ordinal and nominal (weaker)

Question 15

continuous data

Answer 15

ratio and interval (stronger)

Question 16

what is the t-statistic?

Answer 16

regression coefficient

Question 17

what does Butler argue that participatory research is used for?

Answer 17

• include diverse voices

- understand varied ways of framing energy problems and solutions

Question 18

what can qualitative methods provide?

Answer 18

• potential for important critical forms of engagement with processes of energy system change
• they can provide deeper (and different) understandings of the nature of the problem and what we might need to do to address it (Butler et al. 2018)

Question 19

example of ethnography in research

Answer 19

in Danial Fisher’s research on asylum seekers in the EU, it enables the understanding of asylum journey beyond abstract numbers or narratives of pity

Question 20

spearman’s correlation thresholds

Answer 20

0-0.19 very weak

0. 2-0.39 weak
1. 4-0.59 moderate
2. 6-0.79 strong
3. 8-1.0 very strong