Chapter 7 Multiple choice questions

. Download the data file for the multiple-choice questions for Chapter 7, and open it in R. The data frame RTlmm contains reaction time data for the lexical decision experiment described on pages 90 – 95 (though note that the reaction times are different from those given in Table 7.2).

Run a mixed effects model with random intercepts using the lmerTest package. Wordtype and validity are fixed effects, and subject and item are random effects. What is the t-value for the main effect of word type?

18.48

10.90

25.11

-13.17

. For the analysis described in Question 1, what is the F-ratio for the interaction between word type and validity?

118.77

630.36

173.45

-13.17

. For the analysis described in Question 1, is item a significant predictor, and what is its AIC score?

AIC = 16773, is a significant predictor

AIC = 16773, not a significant predictor

AIC = 17833, is a significant predictor

AIC = 17833, not a significant predictor

. For the analysis described in Question 1, what is the conditional R² value for the model?

0.29

16445.31

-0.95

0.76

. Change the model so that item is no longer included as a random effect. Is the fit significantly poorer for this model?

Yes, because the marginal R2 value is lower

Yes, because the conditional R2 value is lower

No, because the main effects and interaction remain significant

Yes, because a chi-square test comparing the models is significant

. The data frame frogdata contains data showing estimates of frogspawn in different ponds across several national parks. The size of each pond is also recorded. Ignoring each pond's location, run a linear regression on these data. Does pond size predict the amount of spawn?

Yes, because the regression is significant (p < 0.05)

No, because the regression is not significant (p > 0.05)

Yes, because the regression slope is positive

No, because the regression slope is negative

. Now run a mixed effects model on the frogspawn data, with parkID as a random effect, allowing random intercepts and slopes. What is the F-ratio and p-value for the effect of pondsize?

F = 135.79, p = 9*10^-16

F = 11.65, p = 9*10^-16

F = 3.53, p = 0.013

F = 49.27, p = 9*10^-16

. Run two further mixed effects models with either random slopes or random intercepts. Compared to the model with random intercepts and slopes, which model changes the outcome most substantially?

The random intercepts model produces a much better fit

The random slopes model produces a much better fit

The random intercepts model produces a much poorer fit

The random slopes model produces a much poorer fit

. Explore the coefficients for the full model with random intercepts and random slopes (from question 7). Which park location has the largest intercept value?

Park 1

Park 3

Park 5

Park 7

. What is the most parsimonious account of the frogspawn data?

Within each individual park there is no relationship between pond size and spawn count, but differences between parks produce an apparent effect

Some parks exhibit a negative relationship between pond size and spawn count, and this disrupts the overall positive relationship

Within each park, pond size predicts the amount of spawn, but there are substantial differences in spawn count between parks

There is no true relationship between pond size and spawn count

Quiz Content