ROCs {MPTinR} | R Documentation |
Data of the meta-analysis on recognition memory ROC (receiver operating characteristic) curves reported in Klauer and Kellen (2015). In total there are 850 individual ROCs, 459 6-point ROCs and 391 8-point ROCs. Both data sets first report responses to old items and then to new item. For both item types the response categories are ordered from sure-new to sure-old. Please always cite the original authors when using this data.
data("roc6") data("roc8")
The source of each data set is given in the exp
column of each data set. The id
column gives a unique id for each data set. For the 6-point ROCs the first 12 columns contain the data, for the 8-point ROCs, the first 16 columns.
Whenever using any of the data available here, please make sure to cite the original sources given in the following.
The 6-point ROCs contains data from the following sources:
Dube_2012-P
and Dube_2012-W
:
Dube, C., & Rotello, C. M. (2012). Binary ROCs in perception and recognition memory are curved. Journal of Experimental Psychology: Learning, Memory, and Cognition, 38(1), 130-151. http://doi.org/10.1037/a0024957
heathcote_2006_e1
and heathcote_2006_e2
:
Heathcote, A., Ditton, E., & Mitchell, K. (2006). Word frequency and word likeness mirror effects in episodic recognition memory. Memory & Cognition
, 34(4), 826-838. http://doi.org/10.3758/BF03193430
Jaeger_2013
:
Jaeger, A., Cox, J. C., & Dobbins, I. G. (2012). Recognition confidence under violated and confirmed memory expectations. Journal of Experimental Psychology: General, 141(2), 282-301. http://doi.org/10.1037/a0025687
Koen_2010_pure
:
Koen, J. D., & Yonelinas, A. P. (2010). Memory variability is due to the contribution of recollection and familiarity, not to encoding variability. Journal of Experimental Psychology: Learning, Memory, and Cognition, 36(6), 1536-1542. http://doi.org/10.1037/a0020448
Koen_2011
:
Koen, J. D., & Yonelinas, A. P. (2011). From humans to rats and back again: Bridging the divide between human and animal studies of recognition memory with receiver operating characteristics. Learning & Memory, 18(8), 519-522. http://doi.org/10.1101/lm.2214511
Koen-2013_full
and Koen-2013_immediate
:
Koen, J. D., Aly, M., Wang, W.-C., & Yonelinas, A. P. (2013). Examining the causes of memory strength variability: Recollection, attention failure, or encoding variability? Journal of Experimental Psychology: Learning, Memory, and Cognition, 39(6), 1726-1741. http://doi.org/10.1037/a0033671
Pratte_2010
:
Pratte, M. S., Rouder, J. N., & Morey, R. D. (2010). Separating mnemonic process from participant and item effects in the assessment of ROC asymmetries. Journal of Experimental Psychology: Learning, Memory, and Cognition, 36(1), 224-232. http://doi.org/10.1037/a0017682
Smith_2004
:
Smith, D. G., & Duncan, M. J. J. (2004). Testing Theories of Recognition Memory by Predicting Performance Across Paradigms. Journal of Experimental Psychology. Learning, Memory & Cognition, 30(3), 615-625.
The 8-point ROCs contains data from the following sources:
Benjamin_2013
:
Benjamin, A. S., Tullis, J. G., & Lee, J. H. (2013). Criterion Noise in Ratings-Based Recognition: Evidence From the Effects of Response Scale Length on Recognition Accuracy. Journal of Experimental Psychology: Learning, Memory, and Cognition, 39, 1601-1608. http://doi.org/10.1037/a0031849
Onyper_2010-Pics
and Onyper_2010-Words
:
Onyper, S. V., Zhang, Y. X., & Howard, M. W. (2010). Some-or-none recollection: Evidence from item and source memory. Journal of Experimental Psychology: General, 139(2), 341-364. http://doi.org/10.1037/a0018926
Klauer, K. C., & Kellen, D. (2015). The flexibility of models of recognition memory: The case of confidence ratings. Journal of Mathematical Psychology, 67, 8-25. http://doi.org/10.1016/j.jmp.2015.05.002
## Not run: # This example shows only how to fit the 6-point ROCs data("roc6") # 2HTM (2-high threshold model) htm <- " (1-Do)*(1-g)*(1-gn1)*(1-gn2) (1-Do)*(1-g)*(1-gn1)*gn2 (1-Do)*(1-g)*gn1 Do*(1-do1)*(1-do2) + (1-Do)*g*go1 Do*do1 + (1-Do)*g*(1-go1)*go2 Do*(1-do1)*do2 + (1-Do)*g*(1-go1)*(1-go2) Dn*(1-dn1)*dn2 + (1-Dn)*(1-g)*(1-gn1)*(1-gn2) Dn*dn1 + (1-Dn)*(1-g)*(1-gn1)*gn2 Dn*(1-dn1)*(1-dn2) + (1-Dn)*(1-g)*gn1 (1-Dn)*g*go1 (1-Dn)*g*(1-go1)*go2 (1-Dn)*g*(1-go1)*(1-go2) " # full 2HTM is over-parametereized: check.mpt(textConnection(htm)) # apply some symmetric response mapping restrictions for D and g: check.mpt(textConnection(htm), list("dn2 = do2", "gn2 = go2")) # UVSD (unequal variance signal detection model) uvsd <- " pnorm(cr1, mu, sigma) pnorm(cr1+cr2, mu, sigma) - pnorm(cr1, mu, sigma) pnorm(cr3+cr2+cr1, mu, sigma) - pnorm(cr2+cr1, mu, sigma) pnorm(cr4+cr3+cr2+cr1, mu, sigma) - pnorm(cr3+cr2+cr1, mu, sigma) pnorm(cr5+cr4+cr3+cr2+cr1, mu, sigma) - pnorm(cr4+cr3+cr2+cr1, mu, sigma) 1 - pnorm(cr5+cr4+cr3+cr2+cr1, mu, sigma) pnorm(cr1) pnorm(cr2+cr1) - pnorm(cr1) pnorm(cr3+cr2+cr1) - pnorm(cr2+cr1) pnorm(cr4+cr3+cr2+cr1) - pnorm(cr3+cr2+cr1) pnorm(cr5+cr4+cr3+cr2+cr1) - pnorm(cr4+cr3+cr2+cr1) 1 - pnorm(cr5+cr4+cr3+cr2+cr1) " # confidence criteria are parameterized as increments: check.mpt(textConnection(uvsd)) # cr1 = [-Inf, Inf] # cr2, cr3, cr4, cr5 = [0, Inf] # mu = [-Inf, Inf] # sigma = [0, Inf] # MSD (mixture signal detection model): # NOTE: To follow CRAN rules restricting examples to a width of 100 characters, # the following example is splitted into multiple strings concatenated by paste(). # To view the full model use: cat(msd) msd <- paste(c(" l*(pnorm(cr1-mu)) + (1 - l) * (pnorm(cr1-mu2)) l*(pnorm(cr1+cr2-mu) - pnorm(cr1-mu)) + (1 - l)*(pnorm(cr1+cr2-mu2)-pnorm(cr1-mu2)) l*(pnorm(cr1+cr2+cr3-mu)-pnorm(cr1+cr2-mu)) + (1-l)*(pnorm(cr1+cr2+cr3-mu2)-pnorm(cr1+cr2-mu2)) ", "l*(pnorm(cr1+cr2+cr3+cr4-mu) - pnorm(cr1+cr2+cr3-mu)) + ", "(1 - l)*(pnorm(cr1+cr2+cr3+cr4-mu2)-pnorm(cr1+cr2+cr3-mu2))", " l*(pnorm(cr1+cr2+cr3+cr4+cr5-mu)-pnorm(cr1+cr2+cr3+cr4-mu)) + ", "(1 - l)*(pnorm(cr1+cr2+cr3+cr4+cr5-mu2)-pnorm(cr1+cr2+cr3+cr4-mu2))", " l * (1-pnorm(cr1+cr2+cr3+cr4+cr5-mu)) + (1 - l)*(1-pnorm(cr1+cr2+cr3+cr4+cr5-mu2)) pnorm(cr1) pnorm(cr1+cr2) - pnorm(cr1) pnorm(cr1+cr2+cr3) - pnorm(cr1+cr2) pnorm(cr1+cr2+cr3+cr4) - pnorm(cr1+cr2+cr3) pnorm(cr1+cr2+cr3+cr4+cr5) - pnorm(cr1+cr2+cr3+cr4) 1-pnorm(cr1+cr2+cr3+cr4+cr5) "), collapse = "") cat(msd) # confidence criteria are again parameterized as increments: check.mpt(textConnection(msd)) # cr1 = [-Inf, Inf] # cr2, cr3, cr4, cr5 = [0, Inf] # lambda = [0, 1] # mu, mu2 = [-Inf, Inf] # DPSD (dual-process signal detection model) dpsd <- " (1-R)*pnorm(cr1- mu) (1-R)*(pnorm(cr1 + cr2 - mu) - pnorm(cr1 - mu)) (1-R)*(pnorm(cr1 + cr2 + cr3 - mu) - pnorm(cr1 + cr2 - mu)) (1-R)*(pnorm(cr1 + cr2 + cr3 + cr4 - mu) - pnorm(cr1 + cr2 + cr3 - mu)) (1-R)*(pnorm(cr1 + cr2 + cr3 + cr4 + cr5 - mu) - pnorm(cr1 + cr2 + cr3 + cr4 - mu)) R + (1-R)*(1 - pnorm(cr1 + cr2 + cr3 + cr4 + cr5 - mu)) pnorm(cr1) pnorm(cr1 + cr2) - pnorm(cr1) pnorm(cr1 + cr2 + cr3) - pnorm(cr1 + cr2) pnorm(cr1 + cr2 + cr3 + cr4) - pnorm(cr1 + cr2 + cr3) pnorm(cr1 + cr2 + cr3 + cr4 + cr5) - pnorm(cr1 + cr2 + cr3 + cr4) 1 - pnorm(cr1 + cr2 + cr3 + cr4 + cr5) " uvsd_fit <- fit.model(roc6[,1:12], textConnection(uvsd), lower.bound=c(-Inf, rep(0, 5), 0.001), upper.bound=Inf) msd_fit <- fit.model(roc6[,1:12], textConnection(msd), lower.bound=c(-Inf, rep(0, 7)), upper.bound=c(rep(Inf, 5), 1, Inf, Inf)) dpsd_fit <- fit.model(roc6[,1:12], textConnection(dpsd), lower.bound=c(-Inf, rep(0, 6)), upper.bound=c(rep(Inf, 6), 1)) htm_fit <- fit.mpt(roc6[,1:12], textConnection(htm), list("dn2 = do2", "gn2 = go2")) select.mpt(list(uvsd_fit, dpsd_fit, msd_fit, htm_fit)) # Note that the AIC and BIC results do not adequately take model flexibility into account. ## model n.parameters G.Squared.sum df.sum p.sum p.smaller.05 ## 1 uvsd_fit 7 1820.568 1377 0 50 ## 2 dpsd_fit 7 2074.188 1377 0 64 ## 3 msd_fit 8 1345.595 918 0 51 ## 4 htm_fit 9 1994.217 459 0 138 ## delta.AIC.sum wAIC.sum AIC.best delta.BIC.sum wBIC.sum BIC.best ## 1 0.0000 1 230 0.0000 1 273 ## 2 253.6197 0 161 253.6197 0 183 ## 3 443.0270 0 16 4996.8517 0 3 ## 4 2009.6489 0 56 11117.2982 0 4 ## End(Not run)